CA2058624A1 - Automated high precision fabrication of objects of complex and unique geometry - Google Patents

Abstract

The present invention provides a method and apparatus for manipulating computer-based representations of objects of complex and unique geometry. A computer acquires data describing an object and its surroundings, constructs a computer-based three dimensional model of the object from that data, superimposes an ideal geometry on the computer-based model, and alters the ideal geometry to fit the form and function required of the reproduction. The present invention presents two embodiments for manipulating computer-based representations of three dimensional objects.

Description

- 2(~~
WQ 9O/15376 . PCT'/US90/03209 ~ .

AUTO~ATED FAE~RICATION OF OBJECTS OF
UNIQUE GEOME$~Ri~ `. ;

This in~ention was made with Government support under Grant Numbers 4F-3~-DE 05377 and lR-29-DE 08455 i~
awarded by the National Institute or Dental Research.
The Government has certain rights in this invention.

TECHNICAL FIELD OF THE INVENTION
This invention relates generally to an automated system for directing the operation of machine tools. In particu~ar, it is directed to a computerized system that machines a reproduction of a three dimensional object such that the reproduction substantially duplicates all surfaces of the ob~ect. In the context o~ this specification~ a "reproduction~l is a category of items which comprises replaccment parts satisfying a particular fit and function, facsimiles generated solely from ; `~
computer-based models, as well as exact or close imitations of existing ob~ects.
.. ~ . . . . . ..
BACKGROUND OF THE INVENTION
There are many applications which require the reproduction or fabrication of one-of-a-kind parts of complex geometry. These applications include dental prostheses, in-the-ear hearing aid housings, molds or implants to replace damaged bones, etc. In such applications, the replacement parts are unique~ Most of the application~ rèquire a highly precise fit~of the replacement part. rr~
Dental prostheses require precise replacement parts. Traditionally, the production of dental prostheses or restorations use lost-wax casting techniques. A
dentist first removes all the decay from the tooth and ensures the integrity o~ the patient's remainin~ tooth structure. The dentist then makes an impression of the remaining tooth structure and the other teeth in the same dental arch. Usually, the dentist also makes an , :

.:, . .. .: : . . . ..... .. - . . : . : . .

20~ ~2~
WO90/15376 PcT/usso/o32o9 impression of the teeth in the opposing dental arch. The impressions are made u~ing an elastomeric material~ The elastomeric material is mixed, placed in a mouth-shaped tray for support, place~ in the patient's mouth over the teeth and surrounding tissues, and held in place until it ~sets.~ Setting times vary with the materials used, but generally range from 2 to 12 minutes per impression. When the elastomeric material sets, it has a consistency similar to that of modeling clay, but retains an elastomeric property so that it can slide past the crest of convexity of the teeth without being permanently distorted. When removed from the mouth, the elastomeric material creates a three d-mensional negative of the teeth and their relative posltions to each other. After the impressions are made, dental stone (p}aster) is 'poured into the impre~sions and allowed to harden. The elastomeric material is then removed and discarded. Pins are set into the bottom of'the arch containing the tooth 'i'to be restored. A base of'plas~er is poured over the pins. When the plaster has hardened, the area'o plaster supporting the model of the prepared tooth i3 sectionsd out. ?he pins and the holes in a second pour of the ~plaster -serve~as guides, permitting the cast of the prepared tooth (called a die) to be'removed and reinserted into the plaster cast. Excess plaster around the base of - the die is removed and the position of the margin of the restoration marked. A ~eparating medium is applied to the 'die. A wax pattern i8 then prepared directly on the die.
If ~aw interrelationships"have been recordèd, ~the wax pattern is mounted on an articulator'.'''The articulator is -' used as a de~ice to-represent the motion of the teeth to - each other. ~fter the wax pattern is compléted, it is sprued and invested for casting. "When the investment ''' material has set,' the in~es~ment and pattern are~placed in an oven-where the wax^is-burned àway. The remaining cavity in the investmënt is then filled with'molten metal ''~ or other casting material. The casting is quenched and , ~,, , . :
.
' ~ ' , :.

: . . : . .: .
- . . -2 3 ~ 7 L~
W~90/1~376 PCT/VS90/~320 recovered. The sprue i5 removed and the cas~ing is finished. Finishing operations involve removing all remaining investment material and polishing the surface.
Surface polish is done by hand using an array of burrs and stones of increasingly finer grit. Final finishing is done with rouge on a soft bristle brush or rag wheel.
After final finishing, the restoration is ready for the patient. The restoration is placed on the tooth and the, occlusal, functional, and proximal contacts are checked.
The quality of the margins is also evaluated.
The clinical criteria for acceptance include:
1. Marginal fit: A dental explorer is passed from reqtoration to tooth and from tooth to restoration with the tip of the explorer held perpendicular to the surfaces of both. If the interface between the restoration and tooth can be detected, the restoration should not be permanently seated but instead should be remade. Depending on the location of the margin relative to the gum tissue, a gap of 20 to 100 microns can be detected.
2. Occlusal fit: A mark should be made when the teeth tap together with a piece o~ occlusal marking paper placed between the teeth. The patient should not detect excessive pressure on that tooth or report any discomfort.
All other teeth which contact without the restoration in place must also contact with the restoration in place.
, 3. Interproximal fit: A p1ece of dental floss should ~'snap~' as it passes between the restoration and the ad~acent contact tooth. ' ,, 4. Functional fit: As the teeth move through their normal excursions, the restoration should not contact prematurely.
, If the restoration can be modified to pass the acceptance criteria, it is cemented in place. If the restoration cannot pass these acceptance criteria, a ne~
impres~ion is made and the process begins again.

..
.. , ... . ~ - :

.:
. : :
.: - .. . ~ . : . - :
, WO 90/15376 2 ;~ 2 i PCr/U~g~tO32~9 SU~ARY OF THE INVENTION
To overcome the limitations in the prior art discussed above and to overcome other limitations readily reco~nizable to those skilled in the art, the present invention provides a method and apparatus ~or the automated reproduction of three dimensional objects of complex and unique ~eometry. Thus, a computer acquires data describing an object and its surroundlings, constructs a computer-based three dimensional model o~ the object from that data, superimposes an ideal geometry on the computer-based model, alters the ideal geometry to fit the form and function required of the reproduction, and then guides a milling machine in the fabrication of the reproduction. In contrast to the prior art, the present invention does not require human intervention in the reproduction process.
The present invention provides two pre~erred embodiments for the automated high precision reproduction of objects of complex and unique geometry. Both embodiments can machine reproductions solely from generic forms stor~d in the computer, thus eliminating the need ~for scanning. Alternatively, the reproductions may be machined solely rom scanned data, without resort to generic forms stored in the computer.
First Preferred Embodiment The first preferred embodiment discloses a method for scanning three dimensional objects and recording the resulting coordinates into a computer. The method comprises positioning a scanning head for recording X and Y coordinates nearest the center of the ohject being scanned. Once positioned, the Z coordinate is determined.
The X, Y, and Z coordinates are recorded into the computer. The head is re-positioned, and additional X,Y,Z
coordinates are recorded. This process repeats itself until a complete radial scan line is recorded into the computer. The object is rotated so another radial scan : ~ : . , . .. .
. .. , ~ . . ;. . . . .
:. . . :: :~.
. .: . . .. . . . . .. . . . .
, .: , . . .
::: , .: .

~O90/1s376 ~ 'J ~ PCT/US90~03209 line can be recorded. The process repeats until the surface of the ob~ect is recorded into the computer as a set of coordinates. These coordinates may be filtered in the computer to eliminate noi~e by eliminating any coordinates corresponding to points not substantially proximat2 to the radial scan lines.
The first preferred embod~ment discloses a method of generating parting lines in a computer such that a surface representation of a three di~ensional object stored in a computer is divided into a plurality of regions. These regions prevent undercuts during a machining process. The method comprises finding an X,Y,Z
coordinate on a radial scan line nearest the center of the object. A point is found on the radial scan line with a largest Y distance from the Z axis. Coordinates on the radial scan line below the parting point are clipped from the line.
The first preferred embodiment discloses a method for superimposing coordinates scanned from a three dimensional ob~ect onto a generic ~orm stored in the computer of the ob~ect to create an altered generic set of .c~ordinates. The generic form, which^is comprised of a - generic set of coordinates, is retrieved from database in the computer. The generic form is spatially rotated in the computer, such that it corresponds to a spatial orientation of the scanned coordinates. The generic form is scaled in the computer, such that the generic coordinates are sized substantially the same as tha scanned coordinates.
~ The ~caling step comprises scaling the generic coordinates in the computer afi a function oE measurements made of other objects in proximity to the scanned ob~ect.
These measurements can include `gap measurements between ~ the scanned object and the proximal objects, height measurements for the scanned ob~ect or proximal objects, and width measurements for the scanned ob~ects or th~
proximal ob~ects.

. . . : - . : . . . .
. , . .. . . :

WO90/l5376 2 0 5`~ S ~}. ~ Pcr/us90/03209 Additional coordinate may be ~electively created in the altered generic set o coordinates to amphasize features of either the generic form or the scanned object.
The method of adding feature coordinates comprises first identifying feature coordinates within the coordin~te set.
The feature coordinates are plural}y represented within the coordinate set by computing a plurality of parametric curves through the coordinate. This prevents the feature from being smoothed out or eliminated.
The coordinates may also be modified in the computer using a shaping function.
The coordinates may be further modified in the computer, when other objects exist in proximity of the scanned ob~ect, to ensure that any interference between the object and the proximal objects is avoided.
- ' The first preferred embodiment discloses a method of generating a plurality of contour lines in a computer.
The contour lines provide he tool paths for a machine .
t" tool. The radial scan lines are reparameterized in the computer to ensure that each radial scan line ha3 substantially,,~the same number ~of coordinates. - ~he coordinates are~,connected between adjacent' radial scan lines according to,,the order of the coordinates.' ~hese connections are the contour lines. -' The contour lines may be selectively altered to provide a predetermined offset from the object surface for the tool paths. The contour lines may be selectively altered in regions selected by the computer or in regions selected by an operator. The selectively "dltering step 30 ,,comprises selecting an offset angle and an'offset height from $he contour line stored in the computer.' -An offset distance is determined from the offset angle and the offset height. A data point is projected from the contour line onto a plane to simplify normal vector calculation.
A inormal vector is calculated at the data point. A new offset data point is determined by shifting the data point the offset distance along the normal vector.

- . .. . . . . . . .
: . .. . , .:
~,~ . . . .

: , , . ; : - ~ .
. . . . . . .
, :. ~ . .. .

: , . . : . .

, WO90~15376 2 ~ j ~ 5 ., /4 PCT/US90/03209 7 ~ -A tool path computation may be performed in the computer to prevent the machine tool from erroneously intersecting with the object surface. The computation comprises checking the minimum radii of principal curvature at a plurality of vertices of the tool paths.
Those vertices where the minimum radius is smaller than a radius of the machine tool are identified. A curve of intersection with a plane is found for each the identifiad vertex. The plane passes through the identified vertex and contains one vector normal to the ob~ect surface area and another vector in the direction of the minimum radius.
A plurality of line segments of an intersection curve are generated, which line segments intersect together. The line segments are subdivided by means of a segmentized searching circle until an error value between the line segments and the ob~ect sur~ace area approaches a pre-determined tolerance value. A segmentized searching circ~e is computed for both the current and an ad~acent surface area. An intersection between triangular segments o~ each of the segmentized searching circles is found.
The triangular,segments are bisected and an intersection between them is determined. Additional segmentized searching circles are computed, and the intersection between them determined, until an error value between the intersected edges of the triangular segments and the object surface area approaches the pre-determined ~ `
tolerance. The segmentized searching circles are propagated until they meet or until the bou~daries of the ob~ect surface area are encountered. A I1ew vertex is selected, a curve of intersection is foundt and the process is repeated, until all of the identified vertices are processed.

.
Second Preferred Embodiment In a second preferred embodiment, an object is scanned and the resulting coordinates recorded into a computer. Two methods of scanning may be used.
' ,~ , :,, ` ~ ` ~ . ' , ; . .
', , i, ` ' ` . ~ . ' '':`' ' ''' ' '-' ' ' ' ' ' : ', ~

2 0 ~ 8 ~
W090t15376 PCT/US90/03209 . , In a first scanning method, a plurality of points onto the object. The points are used to generate a ~et of coordinates, which are recorded into the computer. The points are converted to a plurality of profile lines in the computer. These coordinates may be filtered to eliminate noise by eliminating any of the coordinates corresponding to points not substantially proximate to the profile lines. The coordinates may be modified to ensure that each of the profile lines has subst:antially the same number of coordinates as an ad~acent profile line.
In a second method of scanning, a plurality of profile lines are projected onto the object. The profile lines are scanned to generate a set of coordinates, which are recorded into the computer. The coordinates may be filtered in the computer to e}iminate noise. The coordinates may be modified to ensure that each of the pro~ile lines has substantially the same number of coordinates.
~ After the coordinate~ are recorded into the computer, they may be superimposed onto generic forms ; stored in the computer to create an altered generic set of coordinates.- The generic form is retrieved from a database on the computer, which generic form is comprised of a generic set of coordinates. The generic form is spatially rotated in the computer, such that it corresponds to a spatial orientation of the ob~ect. The generic form is scaled, such that the generic set of ,. coordinates is sized substantially the same as the ob~ect.
Scaling comprises altering the generic coordinates as a function of measurements of other objects in proximity to the scanned ob~ect. These methods include measuring gaps between the scanned object and other proximal objects, measuring the height of other proximal objects, and measuring the width of other proximal ohjects^. The generic form, i.e., the generic coordinates, are further modified using the scanned coordinates.

~., ~ . .
: . . . ~ . , . ~ . ;
..

' WO90/15376 2 ~ PcT/usso/o32o Additional coordinates may be selectively generated to emphasize features of either the ob~ect or the generic form. Additional coordinates are generated by identifying the coordinates of the feature and plurally representing the feature coordinates, for ex~mple, by computing a plurality of parametric curves through the feature coordinate. This prevents the feature from being smoothed out or eliminated by contouring or other modifying function~.
Coordinates may also be modified, when there are objects in the proximity of the scanned object, to ensure that any interference between objects i~ a~oided.
When multiple sets of coordinates are used, they are connected at a common edge, when said common edge exists, to creating a master set of coordinates.
A plurality of surface patches are computed from the master set of coordinates. The method of computing comprises deriving a plurality of parametric curves from ~he master set of coordinates. Surface patches are compu~ed from the paramatric curves. The surface patches are combined ~ into a three dimensional surface representation for machining.i : ~
Parting lines are generated~in the computer such that the surface patches are divided into a plurality of regions. The creation o~ these regions prevent undercuts by the machine tool in the machining of the reproduction of the object. ; '' The parting lines are generated by'~inding those surface patches through which the parting }ine'passes, wherein an edgs on the surface patch has'end po~nts with different signs of sur~ace normal component' values parallel to an axis of the machine tool. A see point is found on the edge at which the component of the 's'urface - normal vector parallel to the machine tool axis'is zero.
The seed poi~t on the edge may be found-by bisecting the edge. The parting line is propagated- by means of a segmentized searching circle until an edge of an adjacent ; . , , , . ,; ~ .. . ..... . . . . .
., ' ., , . , . , ,, ~ !

WOso/ls376 2a~62~ PCT/US90/03209''~

surface patch is encountered. An edge point of the adjacent is identified and replaces the seed point. The parting line is propagated in the same manner unti~ all the surface patches have been processed.
The method of propagating parting lines by a segmentized searching circle comprises generating a plurality of enclosing points about the seed point. A
plurality of line segments are derived between the enclosing points, such that the'seed point is enclosed by a substantially polygonal representation of the segmentized searching circle created by the line segments.
A line segment which intersects ~e parting line is found.
The intersecting line segm~nt is subdivided into a plurality of smaller line segments. A new line segment is found and further subdivision occurs. This repeats until an end point of the line segment is located substantially proximate to the parting line. The end point replaces the seed point, and the steps are repeated until an edge of an ad~acent the surface patch is encountered.
The ~urface patches within each region are contoured to create a plurality of contour line~. The contour lines provide the tool paths for the machine tool.
The contouring method comprises identifying the surface patches through which a contour line passes. One edge of the identified surface patches is identified, through which the contour line passes. A seed point is found on the edge,' possibly by bisecting the edge. The contour line is propagated by means of a segmenti~ed searching circle until all of the identified surface patches have been processed in the computer. The propagating'method compriYes generating a plurality of enclosing points about the seed point. A plurality of line segments are derived between the enclosing points, such that the seed point is enc-losed by a substantially polygonal representation of the segmentized searching circle created by the line segments. A line segments which intersects the contour line is identified. The intersecting line segment is ': ~. ' ` , ' :
.. . . .
~ . ., ~ , . ; .

: . . ., , ~ .
.. . . . . . .
.
I' ~, , , . .: ~

W090/15376 ll 2 ~ ~ 8 ~ 2 ~ PcT/us9o/o32o9 ~ ~

subdivided into a plurality of smaller line segments, which replace the original line segments. A smaller line segment which intersects the contour line is identified.
The intersecting line segment is again subdivided. This process repeats itself until an end point of a line segment is located substantially proximate to the contour line. The end point replaces the seed point and the segmentized ~earching circle method repe~ats until an edge of an adjacent surface patch is encountered.
The contour lines may be selectively altered to provide a predetermined offset from the surface patches for the tool paths. The contour lines may be selectively altered in regions selected by the computer or in regions selected by an operator. The selectively altering step comprises selecting an offset angle and an offset height from the contour line stored in the computer. An offset distance i8 determined from the offset angle and the offset height. A data point is pro~ected ~rom the contour line onto a plane to ~implify normal vector calculation.
A normal vector is calculated at the data point. A new offset data point is determined by shifting the data point the offset distance along the normal vector.
A tool,~ path computation may be performed in the computer to prevent the machine tool from erroneou~ly intersecting with the surface patches. The computation comprises checking the minim~m radii of principal curvature at a plurality of vertices of the surfaca patches. Those vertices where the minimum radius is smaller than a radius of the machine tool are identi~ied.
A curve of intersection with a plane is found for each the identified vertex. The plane passes through the identified vertex and contains one vector normal to the surface patch and another vector in the direction of the minLmum radius. A plurality of line segments of an intersection curve are generated, which line segments intersect together. The line segments are subdivided by means of a segmentized searching circle until an error , . , ~ . .

.::. . . ... , , ::

WOso/1s376 2 0 ~ 8 6 2 4 12 PCT/US90/03209 value between the line segments and the surf ace patch approaches a pre-determined tolerance value.i A
segmentized searching circle is computed for both the current and an adjacent surface patch. An intersection between triangular segments of each of the segmentized searching circles is found. The triangular segments are bisected and an intersection between .them is determined.
Additional segmentized searching circri~s are computed, and the intersection between them dete~mined, until an error value between the intersected edges of the triangular segments and the surface patchès approaches the pre-determined tolerance. The segmentized searching circles are propagated until they meet or until the boundaries of the surface patches are encountered. A new vertex is selected, a curve of intersection is found, and the process is repeated, until all of the identified vertices are processed.
, Both Embodiments Nachine tool commands.are generated, directing the machining of a workpiece material into the reproduction within a predetermined tolerance. These commands, for example, can optimize the step length along the tool paths, the scallop height, and other parameters.
The step length optimization comprises checking an error ~alue between the tool path and the surface patch, and dividing the step length until the error value is smaller than a pre-determined tolerance. The tool path can also be optimized by determining a cutter size for the machine tool in the computer based on the characteristics of the machine tool, a feed rate, a depth o~ cut, a spindle speed, and the characteristics of the material.
The machine tool performance can be optimized by calculating the scallop height for the material based on a plurality of machining par~meters. -The parameters used consist of the characteristics of the tool path, the distance between adjacent tool paths, the spindle speed, ..... . . . . . . .. . . . ..

:; . .. , ~: . :
, .,. . . .. , . ::
: ,: . ~. . .. :.:: , . - .:
:; , , - . . ~ .

20~8G~
WO90/15376 PCT/US90/03~09 the feed rate, the characteristics of the material, and the cutter size. The scallop height is compared to a pre-determined boundary value, and the machining parameters may be al~ered when the scallop height exceeds the boundary value.
The material to be shaped into the reproduction must be fixtured so that it may be machined sequentially by regions, usually through the alternate orientation of the material. The fixturing method comprises selecting a workpiece blank or appropriate size. The blanX may be selected from a plurality of pre-formed blanks, such that the need for multiple machining of each region i8 eliminated, thereby speeding the machining process. These pre-formed blanks may be selected automatically by the computer. The blank i5 oriented on the fixture according to a spatial orientation of the region in the computer.
The fixture secures the blank on a bed surface to prevent movement of the blank during machining. The blank is re-orient and re-machined for each region, as needed.
, Once the blank is selected and oriented, the tool paths and commands are transm~tted from the'computer to ,the machine tool. ~he machine tool is directed to machine the material by the region into the reproduction of the object. The tool paths for each region are transmitted separately and the blank machined accordingly.

. .
: .

. . .:

. , . - :- ' . -.

: :
:-. . . . ., .
.

.~: . : , . : . . , .. ~ , . ; `
-...... .:. .. .

20~862~
WO 90/1!;376 PCI/US90/032(19-DESCRIPTION OF THE DRAWINGS
In the drawings, where like numerals ref~r to like elemen~s throughout the several views, Figure 1 illustrates the hardware components o~
the present invention;
Figure 2 is a system description flow chart for the first preferred embodiment;
Figure 3 is a flow chart describing a data acquisition method for the first preferred embodiment;
Figure 4 is a flow chart describing a parting line for the first preferred embodiment;
Figure 5A illustrates the ruled surface connecting the generic form and prepared tooth surface;
Figure SB illustrates the computer-based representation comprised of an altered generic form, a prepared tooth surface, and a ruled surface connecting the generic form and prepared tooth surface;
Figure 6 illustrates a plurality of unwrapped radial qcan lines;
- . Figure 7A illustrates a plurality of wrapped .. ..
. radial scan lines comprising the generic form;
Figure 7B illustrates a plurality of contour - lines generated for the generic form;
. Figure 8 i5 a flow chart describing a superposition method in the first preferred embodiment;
Figure 9 is a flow chart describing a scaling method in the first preferred embodiment;
Figure 10 is a flow chart describing a method ~or emphasizing features in the first preferred embodiment;
Figure 11 is a flow chart describing a method for shaping generic orms in the first preferred embodiment;
Figure 12 describes a dental prosthesis fitted over a prepared ~ooth; . .
Figures 13A and 13B illustrate the dimensions .
that decide the scaling factor;
Figures 14A and 14B illustrate an interference .
check between ad~acent objects;

. . : ~ .. : -.: . .. :
, ~ ; .-- . . .. .. .. - ..
.. , - .
: , . .... . :
~ ~ . : . .
:, ' ~ . . ' ' .. ' ~ '' ;

20581~2~
W~90/15376 PCT/US9o/032 Figure 15 illustrates a digitized impreSQiOn;
Figure 16 i5 a flow chart describing a method for contouring in the first preferred embodiment;
E~igure 17 is a flow chart describing a method for generating offset tool paths in the first preferred embodiment;
Figure 18 is a flow chart describing a method of eliminating tool interferences in the first preferred embodiment;
Figure 19 is a system description flow chart for the second preferred embodiment;
Figure 20 is a flow chart descri.bing a first data acquisition method for the secon~ preferred embodiment;
Figure 21 is a flow chart describing a second data acquisition method for the second preferred embodiment;
Figure 22 illustrates a generic form as a series of profile lines; .
Figure 23 is a flow chart describing a ~uperimposition method in the second preferred embodiment;
Figure 24 is a flow cha~t describing a method for scaling coordinates in the second preferred embodiment;
Figure 25 is a flow chart describing a method of . emphasizing features in the second preferred embodiment;
2gFigure 26 is a flow chart describing a method for generating surface patches in the second preferred - embodiment;
Figure 27 is a flow chart describing a method for generating parting lines in the seco~d preferred embodiment; .~
Figure 28 is a flow chart describing a method for propagating segmented searching circles in the second preferred embodiment; . -: Figure 29 illustrates the segmented searching circle method;
Figure 30 illustrates a surface composed of aight surface patches in a parametric domain;

. . ... ,.. ,........... : . ~ - .

. -:: ... . ~ . , ' i ,. .: :: . "
. i, :: : : ' : . :: :
',:: , : , . : ::: : : . : :
: .: . .-, Wog0/15376 2 ~ ~ 8 6 2 ~ PCT/USs0/03209`

Figure 31 illustrates the propagation of a searching circle on a single surface patch in the parametric domain;
Figure 32 describes a method for generating contour lines in the second preferred embodiment;
Figure 33 describes a method for generating offset tool paths in the second preferred embodiment;
Figure 34 describes a method fQr eliminating tool interferences in the second preferred~iembodiment;
Figure 35 describes a method for optimizing step lengths in both preferred embodiments; -Figure ~6 illustrates the optimization of tool path step lengths in both preferred embodiments;
Figure 37 describes a method for optimizing cutter feed rates in both preferred embodiments;
Figure 38 describes a method for optimizing scallop heights in both preferred embodiments;
Figure 39 illustrates the optimization o~ tool path scallop heights in both pre~erred embodiments;
Figure 40 describes a method for ~ixturing workpiece materials in both preferred embodiments;
. Figure 41 describes a:method for machining , . . . . .
workpiece materials in bpth preferred embodiments;
Figures 42A/ 42B, .42C,- and 43 illustrate machining, rough cutting, and finish cutting in both preferred embodiments; and Figure 44 illustrates a three axis machine tool used with both preferred embodiments.
, D~TAILED DESCRIPTION OF THE PREFERRED EMBODINENTS
In the following detailod description of the invention, reference is made to the accompanying drawings which form a part hereof, and in which is shown by way of illustration t~o .preferred embodiments in which the invention may be practiced. It i~ to be understood that other embodiments may be utilized and structural changes may be made without departing from ~he scope of the ..

2 ~g~ 2 ~ ~

present invention.
At the end of the following detailed des'cription of the invention, an appendix of citations can be found.
This appendix discusses in detail the technical background 5 information and specific mathematics involved in both ''' preferred embodiments of the present invention. The citations in the appendix are hereby incorporated by reference for the purpose of providing demonstrative '-evidence of the level of skill in the art.
' ''` ~
HARDWARE COMPONENTS ~ ~
Figure l describes the components of the present '' invention. These components include a computer system l0, ' Computer-Aided-Design/Computer-Aided-Nanufacturing lCAD/CAM) software 12, a database of generic forms l4, a machine tool 16, and a digitizer 18. The computer l0 is a UNIX-based wor~ station built by Silicon Graphics, Incorporated. The CAD/CAM software 12 is a custom package that creates a computer-based model of the ob~ect, manipulates the model, and generates command~ directing the machine tool 16 to reproduce the ob~ect. The databa~e' l4 is comprised of files-storing generic'ob'~ect forms that provide a plurality of idealized versions of the ob~ect.
The machine tool 16 is a three axis milling machine built by Servo Products Company. The digitizer 18 -is an optical, non contact, three dimensional, surface digitizer 18 built by Laser Design Corporation. Vsing these components, the present invention can machine reproductions of objects of complex and precise geometry.

THE FIRST PREFERRED EM~OD MENT
OF THE CAD/CAM SOFTWARE
Figure 2 describes the method steps used-in a first preferred embodiment of the CAD/CAM softwaré 12.
The ob~ect is radially scanned and the resulting coordinates are recorded into the computer (28). Radial ,.,, ........ - ,, . :
:,. ~ ...

.. , . , :

; . ., ...
:: . ~ .
. -, :

Woso/1s376 2 0 5 8 ~ 2 4 PCT/US90/03209 scan lines appea~o,start near the center of the top of the object and p~oceed outwa~d toward the bottom of the tooth. Initially, the coordinates are stored in an "unwrapped" state as successive scan lines. The radial scan lines are "clippedll to determine a parting line (30).
The radial scan lines are then ~wrapped~ about the Z-axis to obtain a three dimensional representation of the object (32). The scanned coordinates may be superimposed onto a generic form of the object stored in the computer, thus creating an altered generic set of coordinates (34).
Contour lines are created by connecting corresponding points on ad~acent radial scan lines ~36). The contour lines may be projected in a normal direction to the surface to create the offset tool paths f or the machine tool (38). Multiple offset tool paths may be generated to provide rough and finish cutting planes, and to achie~e a desired machining tolerance. Tool interference checks are . made to elLminate gouging (40). Additional machine tool commands are generated in the computer as required~(42).
$he material,,to be"machined is.fixtured such that it may , be machined.sequentially by region (44). The offset tool paths and commands are then transmitted from the computer to the,machine tool for the machining step (46)~.
i.. : . . ....... .. - . . :
Data Acquisition .-Figure 3 describes a method of data acquisitionin the first preferred embodiment. Data acquisition is ,the first step in generating a computer-ba~ed .representation of a three dimensional ob~ect. In the first preferred embodiment, the digitizer 18 is a point-by-point triangu}ation system. As described in Figure 1, the digitizer 18 is comprised of a three axis positioning head 20 with a lo~ energy laser source and detector, a . rotational stage 24, and a computer controller 26. The controller 26 positions both the rotational stage 24 and the positioning head 20 (48).. An object is placed on the ,rotational stage 24, and the laser beam reflects from it.

:: ~ ~ , ; . ::
, ; . ~; ,. :.; ` ' ~
., . : :.
.. .. . . .
,. 1 ... . . . :
.:,. .. :
, ~ i, ~' : ,: . . , ` ' ' `

~ i W090/15376 2 ~ 5 ~ ~ 2 4 PCT/US90/03209 The reflected laser beam is used to measure the distance between the object and the laser source (50) ~he ~Z) coordinate on the surface is established by combining the position of the laser source with the determined distance between the object and the laser source. The computer 10 records the head 20 position into the database 14 as (xly~z) coordinates (52). These coordinate~ are stored in an ~unwrapped~' state. The digitizer 18 creates radial scan lines by beginning near the center of the top of the object and moving radially outward from the object. The controller 26 orders either the rotational stage 24 or the positioning head 20 to shift to a new (X,Y) position to obtain a new (Z) value (54). The controller 26 orders the rotational stage 24 to rotate to start a new line (56).
In this manner, the surface of the ob~ect is ~canned and represented by a grid of (X,Y,Z) position coordinates (58). For objects with undercuts or similar features such as vertical walls, the rotational stage 24, or ~ust the ob~ect itself, may be set at any angle relative to the laser source for scanning. If such an ob~ect is not anqled during the scan, data may be lo~t. For example, the digitizer 18 may erroneously sense that an edge of the object has been reached and stop scanning. Data can be obtained at variable step sizes, with a minimum step size of 8 microns. The surface can be scanned at approximately ten to thirty points per second. Naximum digitizing volume is six inches by six inches by three inches. The CAD/CAM software 12 may filter the coordinates to eliminate noise (60). Filtering eliminates any coordinates that correspond to points not substantially proximate to the radial scan line (60).
~ he data acquisition method used in the first preferred embodiment is well suited to the fabrication of dental prostheses. Each patient presents a u~ique set of teeth shapes and sizes. ~hus, each restoration has a unique form, depending upon the history of decay fracture, and previous restorations. The data required to ... .
''.'~. ` , ' ~ ' .

i . ; ,. ~ ' '. ~ ' woso/1s376 2 0 S 8 6 2 4 PCT/US90/03209 ~0 ~' produce the dental prosthesis includes: (1) the configuration of the tooth prepared by the dentist to receive the prosthesis; (2) the gap between, the heights of, and the widths of, the adjacent teeth which provides the scaling factor; (3) the surface configuration of the opposing teeth with which the prosthesis must occlude; and (4) motion of the mandible relative to th~ maxilla during function (in the areas where any of the teeth remain in contact and therefore guide the motion o-f the jaws). With the first preferred embodiment, the process to create a den~al prosthesis begins in a similar fashion as the traditional wax casting technigues discussed hereinbefore.
A dentist makes an impression of the two arches of teeth.
The impression is a three-dimensional negative of the teeth shapes. The dentist creates a model duplicating the teeth by pouring plaster into the impression. When the plaster has set, the impression is discarded. The ~; , digitizer 18 scans the plaster model and records the resulting coordinates into the database 14.
~
Partinq Line Generation Figure 4 describes a method for generating parting lines in the first preferred embodiment.` After recording the coordinates for the radial scan lines, a parting line is generated so that undercuts and their associated interferences are avoided during the machining step. Initially, the coordinates are stored in an l'unwrapped" state as successive scan lines. Starting at a point nearest the center of the ob~ec~ (62)~ a search is made for a point on the radial scan line with the largest Y-axis distance from,the Z-axis (64). The first point with the largest Y-axis distance (i.e., the relative ~, maximum) is a point on the parting line. A parting line is generated by "clipping~ data on each of the radial scan lines below the selected parting linP (66). Each of the successive radial scan lines are processed in a similar manner to form the complete parting line t68).

~- .

:. ., - ~ . ., . ~

:': ' ' . : , '' ;........... . .

2~ s2~
WV90tlS376 PCT/US90/03209 As illus~rated in Figures 5A and 5B, a single parting line 72 in the first preferred embodiment typically defines two ma~or regions for a dental prosthesis 70. Both Figures 5A and 5B are contour line representations, discussed herein later, of the dental prosthesis 70. One region is a generic form 74 representing the occlusal surface and sides of the prosthesis 70 to the crests of convexity (the two maximum horizontal points in any given radial scan line to the parting line). The generic form 74 is altered to fit the space available in the patient~s mouth. The other region is a prepared tooth surface 76 representing the lower part of ,the configuration. The prepared too~h surface 76 of the prosthesis 7D is manipulated to ~it over the prepared tooth. Also included in this region is a connecting surface 78 between the generic form 74 and the prepared tooth surface 76, i.e., the surface below the crest of convexity, typically extending from the parting line 72 of the generic form 74 to ~he margin 80 of the prosthe3is.
,20 Database Of Generic Forms , ~ When data acquisition and parting line generation are complete, the first preferred embodiment of the CAD/CAM software generates a set of data files so a computer-based model of the three dimensional object can be stored on the computer. Figure 6 is an illustration of the radial scan lines stored in an "unwrapped" state.
Figure 7A is an illustration of the ~wrapped~' radial scan ,, lines. . -`
The surface:82 of the ob~ect is represented as a plurality of points 86 and 88 on a plurality of radial scan lines 84. Each radial scan line 84 appears to start near the center of-the top of the object and proceed outward. ~ach of the radial scan lines 84 are at different angular incremen~s around ,the center of the object. Thus, each radial scan line 84 functions as a planar slice or cross-section of the three dimen~ional .... . ,.,, , ~ . .
:' . . ~ . . . .
.. : , ~ , . ~ . -. . : .

. . .

205862'1 W090/15376 PCT/VS90/03209 ~ , surface 82 beginning at the center of the object and extending to the outer surface.
~ sually,,,the`~atabase contain~ a plurality of standardized ob~ect representations, referred to as generic forms. A generic form is comprised of a generic set of (X,Y,Z) coordinates. The generic form pexmits the fabrication of reproductions based on idealized or standardized geometries. -Like the sca~med objects, the surface of each generic form 82 is represented in the computer lO as a plurality of points 86 and 88 on a plurality of radial scan lines 84. Each radial scan line 84 appears to start near the center of the top of the generic form 82 and proceed outward. Each of the .radial scan lines 84 are at different angular increments around the center of the generic form 82. Thus, each radial scan line 84 functions as a planar slice or cross-section of the three dimensional surface beginning near the center of the generic form 82 and extending to the outer surface.
. , When the fixst preferred embodiment is used to produce dental prostheses, the database contains a plurality of standardized generic -tooth forms.., The ,, ,generic tooth forms used ,are ,typically,computer-based representations of standardized plaster-models of teeth.
....
Superposition Of Generic Form Figure 8 describes a method for superimposing the scanned coordinates onto a. generic form stored in the database. Landmarks on the generic form are matched with, and compared to, corresponding landmarks on the scanned 30.. ob~ect, thereby providing a correct orientation and si2e for the generic form. Thé CAD/CAM software retrieves a generic form from the database (90). The generic form is "spatially rotated" and positioned so.that it corresponds .. .to a spstial orientation of the se~ of coordinates.scanned from.the object (92~. This orientation is performed by matching at least three landmarks on the generic form with corresponding landmarks on the scanned object. ~he ' . . - . , ... : . .: . - . . . . .

20~862~
W~9~/15376 PCT/US90/0320 CAD/CAM software also scales the generic coordinates so they are sized substantially the same as the scanned coordinates (~6). Additional coordinate.s may be created to prevent features from being smoothed out or eliminated ~96). The resulting representation may be shaped if desired (98). The representation may all30 be checked for interferences with other objects adjacent to where the reproduction will reside (10~
Fiyure 9 describes the scaling operation performed on ~he generic form. Gap measurements are recorded between the object and other objects adjacent to it (116). Also recorded are height and width measurements of the object and/or adjacent objects (118 and 120). The ratio of these values to the equivalent distances on the generic form yields a scaling factor. This scaling factor may be applied to the set of coordinates to create a linear transformation of the coordinates (122). Thus, the generic form can be sized to fit ~he space available for the reproduction. Note that with this method, a replacement part may be created using the computer-based generic form scaled according to the "gap" left for its - placement, and from measurements of adjacent objects.
Thus, the original object is not required when machining replacement parts, unless it is desired to construct a truej substantially identical -"reproduction" of the objec~.
Figure 10 describes a method Eor creating additional feature coordinate~ in the first preferred embodiment. Additional coordinates can be added to the set of coordinates to emphasize features of the object or the generic form. The additional coordinates ensure that the feature is not smoothed out, or otherwise eliminated, either by system operators or by the CAD/CAM software.
feature coordinate: is identified within` the set of coordinates (124). The- CAD/CAM - sof~ware plurally represents the feature coordinate by computing a plurality of parametric curves extending through the feature :
, . . .

2 0~ 6 2 4 ~?
W090/1s376 PCT/US9OfO3209 coordinate (126). These parametric curves are used to provide additional coordinates, thereby preven~ing the f~ature from being smoothed out or eliminated. ~`
Figure 11 describes a me~hod for shaping the generic form in the first preferred embodiment. Using this method, a point on the surface represented by a coordinate is selected and moved to a n~w position (128).
The set of coordinates can be changed, with local effects, using a free-form deformation techn~gue ~130). The surface geometry is deformed using a shaping function.
The deformations decrease radially about the point of interest. Points very near the mo~ed point will move almost as much. Points further away will move less, as determined by the selected shaping function, decreasing to zero movement at the selected maximum radius. This provides the equivalent of a computer shaping tool for the system operator.
When the firs~ preferred embodiment is us~d to produce dental prostheses, it incorporates information from both the scanned tooth and from the gen~ric tooth form. Figure 12 illustrates this incorporation. A
, . . ..
generic tooth form 136 stored in the database contains a local coordinate system 140, based on maximum height (in the occlusal plane) of the cusp tips. The database also contains the positions of the proximal contacts 142 and 144 relative to that local coordinate system 140. If a landmark 146 is placed on the prepared tooth 148, a local coordinate system 150 relative to the prepared tooth 148 can b~ calculated. Since the two contact points 142 and 30; 144 on the prosthesis must match the two contact points on the proximal teeth, a transformation-from the generic coordinates to the prosthesis coordinates can be made.
With this transformation, a scaled generic form 136 is created for the prepared tooth configuration 148.
The design of the surfaces of the prosthesis is more complicated than simply transforming the coordinate system. The occlusal surface of the prosthesis is . : , . . :, ., . . ;. .: : . :
.
., ~ , . .
. ~: , , ., ~, , . :

., , . . :

2 ~ 2 ~
WO90/1~376 PCT/US90/03209 suggested by the generic tooth form ~tored in tAe database 14. However, as illustrated by Figures 13A and 13B, the prosthesis must contact the adjacent te,eth 152 and 154.
To accomplish this, the CAD/CAN so~tware sca~es the S generic tooth form so that the mesial-distal distance between proximal teeth matches the distance between contact points of the generic fonm. The scaling factor used is simply the ratio of actual mesial-distal length over the mesial-distal length of the generic ~orm. ~hus, the size of the generic form can be altered by checking the gap between the adjacent teeth 152 and 154. The width and height of the generic form can be determined by measuring the thickness and height of the adjacent teeth 152 and 154. Also, the coordinates representing the fossae and cusps of the occlusa} surface may be altered to raise or lower the cusps. This scaling factor is most likely provided by measuring the upper and lower teeth wh~le in occlusion.
Figures 14A and 14B show that the interference between the upper and lower teeth cAn be checked by using - contour lines~ Interferences are solved by moving the -coordinates involved. The occlusal surface of the prosthesis 156 is checked when the maxilla tooth 158 is in contact. First, the relative positions between mandible and maxilla teeth have to be found. Figure 15 illustrates the data collection method. Impressions 160 in a wax plate 162 are digitized to find the relative position between the mandible and maxilla teeth. Lan~marks~164 are plugged into the impression 160. The mandible and maxilla teeth can be digitized from opposite sides of the impression 160. Also, the position of the landmarks 164 are digitized. ~hus, the relati~e position of the mandible and maxilla teeth can be calc~lated based on the position of the landmarks 164.
~s described by Figure 12, the lower part of the prosthesis 134 fits over the prepared tooth 148. This prepared to~th surface 134 must be connected to the .,., , , ~ . . . ..

: : . .
., ~ , . . . . .
.

20~8~2~ ;
WOg~/15376 PCT/~SsO/03209 generic form p~roviding the occlusal surface 136.
Typically, a ruled surface 138 (e.g., straight, ruled lines) is generated from the margin of the prepared tooth surface to the parting line of the prosthesis. The resulting margin of the prosthesis is u~sually sharp. A
small error near the edge could cause severe problems, such as too loose or too tight a fit.

Contourin~
Figure 16 describes a method for contouring the generic object in the first preferred embodiment. Each radial scan line is reparameterized, so that all radial scan lines have the same number of coordinates (166).
Contour lines are generated through these coordinates, matching the order of coordinates in each radial scan line (168). For example, the first coordinate point in the first radial scan line is connected to the first coordinate point in th~ second radial scan line, the cecond coordinate point in the first radial scan line is connected to the second coordinate point in the second radial scan line, etc. Thus, the contour lines are essentially perpendicular to the radial scan lines. These contour lines define the machine tool 16 contact point on . . .
the surface of the prosthesis. Enough contour lines are generated to connect all coordinates ~170). The contour lines may be selectively altered to provide offset tool paths from the reproduction surface ~172 and 174). Offset tool paths are generated in such a way as to prevent the machine too} 16 from erroneously intersecting -with the surface of the prosthesis. ~fsets can be applied to regions automatically selected by the CAD/CAM software.
Alternatively, offsets can be applied to regions manually selected by a systems operator.
Figure 17 describes a method for generating offset tool paths in the first preferred embodiment. The CAD/CAN soft~are selects an offset angle and an offset height from the contour line (176). The offset angle and ' WO90/15376 2 ~ ~ 8 ~ 2 ~ PCT/US90/032V9 offset height can be selected arbitrarily. The offset angle and the offset height determine the off~et distance (178). A data point is projected the offset distance from the contour line ( 180) ~ The CAD/CAM software calculates the normal vectors to the contour lines at each coordinate point (182). The CAD/CAM software detennines a new offset data point by shifting the data point offset distance along the normal vector from the contour line ~184).

Tool Interference Chec~
Figure 18 describes a method for preventing tool interferences in the first preferred embodiment. The CAD/CAN software performs a computation to prevent the offset tool paths from erroneously intersecting with the lS prosthesis surface repr~sented by the original scan lines or the original contour lines. The CAD/CAM software checks the minimum radii of principal curvature at the vertioes or coordinate points of the offset toel paths (186). The vertices where the minimum radius is smaller than the machine tool 16 radius,are identified (188). The CAD/CAN software finds a curve of intersection of each identified vertex with a plane. The plane passes through the identified vertex and contains a first vector normal to the surface and a second vector in the direction of the minLmum radius (190~. A plurality of line segments of the intersection curve are generated, which line segments intersect together (192). The line segments are subdivided using a segmentized searching circle, described hareinbelow, until an error value between the line segments and the surface approaches a predetermined tolerance value (194). A segmentized searching circle is computed for the two adjacent surface areas (196). An intersection between a plurality of triangular segments of the segmentized searching circles is found (198). The triangular segments are bisected (200). An intersection between the triangular segments is formed (202). These steps are repeated (i.e., the segmentized searching circle ~ .

,~ . ,, . , . :
,: : ,. . : . . . : , : , .
: , .. ; . . ~ , . . ... : . : .

W090/l~376 2 0 ~ 8 6 2 ~ PCT/US90/V32~9' ' , computing step (196) through the intersection finding step (202)) until an error value between a plurality of intersected edges of the triangular segments and the surface approaches the predetermined tolerance (204). The segmentized searching circles are propagated until the circles meet or until the c-~rcles reach one of a plurality of boundaries for the surface areas (206). These steps are repeated (i.e., the curve of intersection finding step (190) through the segmentized searching circle propagating step (206)) until all identified vertices are processed (208).

THE SECOND PREFERRED ENBODINENT
OF THE CAD/CAM SOFTWARE
Figure 19 describes the method steps used in a second preferred embodiment of the CAD/CAM software. The object is scanned and coordinates describing the object ,are recorded into the computer (210). More than one set of coordinates may be recorded for'each ob~ect. The set of coordinates may be superimposed onto a generic form of the ob~ect stored on the computer (212). The generic form ,is comprised of a generic set of coordinates. The superposition of the recorded coordinates thus creates an altexed generic set of coordinates. If the object is represented by more than one set of coordinates, the sets are connected to create a master set of coord.inates (214~
A plurality of surface patches are computed from the master set of coordinates (216). Parting lines may be generated to divide the surface patches into'a plurality of regions, thus preventing ~mdercuts by the machine tool 16 (218). Contour lines are generated from the surface patches (220). Offset tool paths are generated from the contour lines that provide a predetermined offset from the surface patches for the machine tool 16 '(222~.' A
computation is performed on the offset tool paths to prevent the machine tool 16 from erroneously intersecting with the surface patches (224). Additional machine tool ,-,, ~ .
:-: . . ,, . ;

. .. . , .. :
.. : .... ~ .
. :. . ,~, .: - , . . .
.: . : , -: . , . .- :, - : . :

2~58~2~
WOgO/15376 PCT/US90/03209 commands are generated in the computer lO (226~. The workpiece material is fixtured such that it may be machined sequentially by region (228). The offset tool paths and commands are transmitted from t:he computer lO to the machine tool 16 (230).

Data Acquisition Figures 20 and 2l describe two alternative methods of data acquisition in the second preferred embodiment. The digitizer is the same used in the first preferred embodiment. However, profile lines rather than radial scan lines are recorded for each ob~ec~. Profile lines appear to be planar slices o~ the three dimensional surface of the scanned object.
In Figure 20, the digitizer projects a plurality of points onto the ob~ect (232). The second preferred embodiment scans the three dimensional ob~ect (234) and records the resulting 3et of coQrdinates of points into the computer,,,~236). More than one set of scanned coordinates can be created to represent each ob~ect. The CADtCAM software converts the set of coordinates into a plurality of profile linas (238) and filters the`'set of coordinates ~,to eliminate ,-noise (240). ' Filtering eliminates any coordinates that correspond to points not substantially proximate to the profile lines. The scanned coordinates representing the profile lines are modified to ensure that each of the profile lines has substantially the same number of coordinates as ad~acent profile lines (242). , , In Figure 21, the digitizer projects a plurality of profile lines onto the ob~ect (244). The three dimensional object is scanned ~246) and the resulting set of coordinates is recorded into ~he computer (248). More than one set of coordinates can be created to~represent each object. The CAD/CAN software filters the set of coordinates to eliminate noise (250). Filtering eliminates any coordinates that correspond to points not ; . . - :................................... . , : .. . . . : .

2~5~24 ~
WOsO/l5376 PCT/US90/032~9 substantially proximate to the profile lines. The scanned coordinates representing the profile lines are modified to ensure that each of the profile lines has substantially the same number of coordinates as adjacent profile lines ~252). The data acquisition method used in the second preferred embodiment, like that of the first preferred embodiment, is well suited to the fabrication of dental prosthese~. Each patient prese~nts a unique set of teeth shapes and sizes. Thus, eac~ restoration has a unique form, depending upon the history of decay, fracture, and previous restorations. The data required to produce the dental prosthesis includes~ he configuration of the tooth prepared by the dentist to receive the prosthesis; (2) the gap bet~een, the heights of, and the widths of, the adjacent teeth which provides the scaling factor; (3) the surface configuration of the opposing teeth with which the prosthesis must occlude; and (4) motion of the mandible relative to the max~lla during function (in the areas where any of the teeth remain in contact and therefore ~uide the motion of the ~aws). In the second preerred embodiment, the process to create a dental prosthesis begins in a similar fas~ion` as the traditional wax casting techniques discussed hereinbefore.
A dentist makes an impression of the two arches of teeth.
The impression is a three-dimensional negative of the teeth shapes. The dentist creates a model duplicating the teeth by pouring plaster into the impression. When the plaster has set, the impres~ion is discarded. The digitizer scans the plaster model and records the resulting coordinates into the database.

Database Of Generic Forms When data acquisition is complete, the second ~preferred embodiment of the CAD/CAM software generates- a set of data files so a computer-based model of the three dimensional ob~ect can be stored on the computer. Figure 22 is an illustration of the profile lines.

. . . -, . . .

:' ~ , ' . ' ' , . . - . .
' ~ ' ' ' . ;

2a5~24 W090/ls376 PCT/VS90/03209 The profile linas represent the surf~ce of the object as a plurality of points 254 on a plurality of profile lines 256. Each of the profile lines 256 are at different increments along the surface of the object.
S Thus, each profile line 256 functions as a planar slice of the three dimensional surface of the scanned ob~ect.
In the second preferred embodiment, as in the first preferred embodiment, the database contains a plurality of standardized object representations, referred to as generic forms. The generic form is comprised of ~
generic set of (x/y/z) coordinates. The generic form permits the fabrication of reproductions based on idealized or standardized geometries. In the second preferred embodiment, the surface of each generic form is represented in the computer as a plurality of points 254 on a plurality of profile lines 256. Each of the profile lines 256 are at different increments along the surface of the generic form. Thus, each profile line 256 functions as a planar slice of the three dimensional surface of the generic form.
When the second preferred embodiment is used to produce dental prostheses, the database contains a plurality of- standardized generic tooth forms. The generic tooth forms used are typically computer-based 25 representations of standardized plaster models of teeth. `-Superposition Of Ideal Generic Form Figure 23 describes a method for superimposing the scanned coordinates onto a generic form stored in the database. Landmarks on the generic form are matched with, and compared to, corresponding landmarks on the scanned object, thereby providing a correct orientation and size for the generic form. The CAD/CAM software retrieYes a generic form from the database (258). The generic form is ~spatially rotated'l and positioned so that it corresponds to a spatial orientation of the set of coordi~ates scanned from the object (260). This orientation is performed by , . . ,,, , . ~ ~ . :

WOgO/15376 2 ~ 5 8 ~ 2 ~ PCT/US90/03209~
~' 32 matching at least three landmarks on the generic form with corresponding landmarks on the scanned object. The CAD/CAM sof~ware also scales the generic coordinates so they are sized substantially the same as the scanned coordinates (262). Additional coordinatles are created to emphasize features of the object or generic form (264).
Interferences among objects are eliminated ~266). ~;
Figure 24 describes the sc:aling operation performed on the generic form. Gap measurements are recorded between the object and other ob~ects adjacent to it (268). Also recorded are height and width measurements of the object and/or adjacent objects (270,272). The ratio of these values to the equivalent distances on the generic form yields a scaling factor. This scaling factor may be applied to the set of coordinates (274). Thus, ~he generic form can be sized to fit the space available for the reproduction. N~te that with this method, a replacement part may be created using the computer-based generic form scaled according to the "gapl' left for its placement, and from measurements o~ ad~acent objects.
Thus, the original ob~ect i8 not required when machining replacement parts, unless it is desired to construct a true, substantially identical "reproduction~` of the object.
Figure 25 describes a method for creating additional feature coordinates in the second preferred embodiment. Additional coordinates can be added to the set o~ coordinates to emphasize features of the object or the generic form. The additional coordinates ensure that the feature is not smoothed out, or otherwise el~minated, either by system operators or by the CAD/CAN software. A
feature coordinate is identified within the set of coordinates (276). The CAD/CAN software plurally -represents the feature coordinate by computing a plurality of parametric curves extending through the feature coordinate (278). These parametric curves are used to provide additional coordinates, thereby preventing the . , ~ . . - . , .. ~ , . .: . ,. ':
, 2o5~24 W090~15376 PCT/~S90/03209 feature from being smoothed out or eliminated.
When the second preferred embodiment is used to produce dental prostheses, it incorporates information from both the scanned tooth and from the generic tooth form, like the first preferred embodiment. Figure 12 illustrates this incorporation. A generic tooth form 136 stored in the data~ase contains a local coordinate system 140, based on maximum height (in the oc:clusal plane) of the cusp tips. The database also contains the positions of the proximal contacts 142 and 144 reIative to that local coordinate system 140. If a landmark 146 is placed on the prepared tooth 148, a local coordinate system 150 relative to .the prepared tooth 148 can be calculated.
Since the two contact points 142 and 144 on the prosthesis must match the two contact points on the proximal teeth, a transformation from the generic coordinates to the prosthesis coordinates can be made. With this transformation, a scaled generic ~orm 136 is created for the prepared tooth configuration 148.
; The design of the.surfaces of the prosthesis is more complicated than simply:transforming the'coordinate system. -- The occlusal surface of the prosthesis is suggested by the. generic. tooth ~form stored``'in the database. However, as illustrated by Figures 13A and 13B, the prosthesis must contact the adjacent teeth 152 and 154. ~o accomplish this, the CADtCAM software scales the generic tooth form so that the mesial-distal distance between proximal teeth matches the distance 'between .
contact points of the generic form. The scaling''factor ~
30 used is simply the ratio o~ actual mesial-distal length ~.
over the mesial-distal length of the generic form. Thus, the size of the generic form can be altered by checking the gap between the adjacent teeth 152 and 154. The width ':~
and height of. the: generic form ~can be determinèd by .. ~.
measuring the thickness and height of the'adjacent-~~teeth 152 and 154. Also, the coordinates representing the '~;
fossae and cusps of the occlusal surface may be altered to :: ':

,,. ;,. ,., ,, ',' ; ', ',` . '' .. ~' "' ~'' ~" , 2 ~ rJ~
WO90~1~376 PCT/US90/0320 raise or lower the cusps. This scaling factor is most likely provided by measuring the upper and lower teeth while in occlusion.
Figures 14A and 14B show that the interference between the upper and lower teeth can be checked ~y using contour lines. Interferences are solved by moving the coordinates involved. The occlusal surface of the prosthesis 156 is checked when the teeth of the mandible and maxilla are in contact. First, the relative positions between mandible and maxilla teeth have to be found.
Figure 15 illustrates the data collection method.
Impressions 160 in a wax plate 162 are digitized to find the relative position between the mandible and maxilla teeth. Landmarks 164 are plugged into the impression 160.
The mandible and maxilla teeth can be digitized from opposite sides of the impression 160. Also, the positions of the landmarks 164 are digitized. Thus, the relative position of the mandible and maxilla teeth can be calculated based on the position of the landmarks 164.
As described by Figure 12, the lower part of the prosthesis 134 ~its over the prepared tooth. This prepared tooth surface--134 must:be connected to the generic form providing the occlusal surface 136.
Typically, a ruled surface 138 ~e.g., straight, ruled lines) is generated from the margin of the prepared ~ooth surface to the parting line of the prosthesis. The resulting margin of the prosthesis is usually sharp. A
small error near the edge could cause se~ere problems, such as too loose or too tight a fit.
30 - j Three Dimensional Representation ~ .
Figure 26 describes a method for computing surface patches from the scanned coordinates in the second preferred e~bodiment. A plurality of parametric;curves are derived from the~coordinates (280). The surface patches are-created from the parametric curves (282). The - ~

~.~.,.. - - - .. . . . :

.: , : , .
.
~" . ~ :`. . ` ; ' ` . .
:, . . . :.
. . : . . :

- 2058~2~
WO90/1~376 ~ PCT/US90/03209 combinsd surface patches provide a three dimensional representation of the object (284).
The surface patches are defined by bi-cubic uniform B-splines with multiple knots fo:r boundaries. B-splines are used for a number of reasons: (1) the surac2can be changed locally by Lmply moving control vertices;
(2) the control vertices approximate the interpolated surface; and (3) the surface can be constructed with sLmple equatio~s. In additionl the B- pline geometry ensures continuity across the surface of the ob~ect without sudden changes to second order derivatives. This property ensures aesthetics and sLmplifies subsequent machining.
When the second preferred embodiment is used to produce dental prostheses, it is important that the prostheses surfaces are smooth and continuous. All external surfaces can be polished with round burrs because they are blended surfaces. ~hus, thore are no sharp points on the sur~ace. These aesthetics are preserved with the use of B-splines. When machining begins, it is important to avoid sudden accelerations. Otherwise, damaged or broken tools result, and the workpiece materials may be-gouged or torn. By ensuring continuity across the surface patches to second order derivatives, these problems can be eliminated.
' Partina Line Generation Figure 27 describes a method for generating parting lines in the second preferred embodiment. Parting line generation occurs in the second preferred embodlment after the surface patches have been constructed. This allows the surface patches to be divided into regivns ~y a p~rting line, so that undercuts and their associated interferences are avoided. The parting line on a surface patch is perpendicular to the cross-product of two vectors. The first vector i~ parallel to a first axis of the machine tool (the vertical Z-axis of a 3-axis milling .. ,, ... : ............. - ;. - - ..................... ,. ~ . - ~
:,: , ... .'" ~ . . , :
, . . . .

3~6 2 ~ 5 8 6 2 ~ PCT/~S90/03209f machine). The second vec~or is a normal ~ector of a point on a surface patch.
The CAD/CAM software finds those surface pa~ches through which the parting line passes ~286). An edge of these surface patches has end points with different signs of surface normal component values parallel to an axis of the machine tool. A seed point is found on the edge (boundary) of a marked surface patch at which the component of the surface normal vector parallel to the machine tool axis equals zero (288). The seed point on the edge of the surface patch can be found, for example, by successively bisecting the edge until the surface normal component is zero. The parting line is propagated using a segmentized searching circle until an edge of an ad3acent surface patch is encountered (290). An edge point of the adjacent surface patch is identified (292) and the seed point is replaced (294). These steps are repeated (i.e., the propagating step (290) through the replacing step (294)) until all found ~ur~ace pa~ches are processed (296). ~
As illustrated in -Figure 12, a single parting line in the second preferred embodiment typically defines two major regions for a dental prosthesis. One part of the configuration is a generic form 136 representing the occlusal surface and sides of the prosthesis to the crests of convexity (the two maximum horizontal points in any given radial scan line to the parting line)~ ~he generic form 136 is altered to fit the space available in the , . . .
patient's mouth. The prepared tooth surface 148 determines the other part of the configuration. ~he lower part of the prosthesis 134 is manipulated to fit over the prepared tooth 148. Also included in this region is a connecting surface 138 between the generic form and the prepared tooth surface, i.e., the surface below the crest of convexity, typically extending from the parting`line to the margin af the prosthesis.

. . .
:-. -~ . .
.. . .

WosO/1s376 PCT/US90/03209 Seqmentized Searchinq Circle Figure 28 describes a searching circle method of propagating lines through surface patches.~ The segmentized searching circle is propagated by first generating a plurality of enclosing points to create a substantially polygonal representation ,about a seed point (298). A plurality of line segments are derived between the enclosing points to enclose the seed point. The points are checked ~300) to find two ad~acent points ~hich have different signs of surface normal components parallel to the machine tool axis (302). A new plurality of points with finer spacing are generated between the two points and the process repeats until the desired accuracy is obtained (304). These steps are repeated (i.e., the checking step (300) through the generating step (304)) until an edge of an ad~acent surface patch is encountered (306).
Figure 29 illustrates the segmented searching .. ..
circle algorithm. A segmentized searching circle is a modification of the bisection algorithm. If any continuum i8 being searched for points having a particular value and a seed point 308 is known- which has that value, an ad~acent point having that value can be located by drawing a circle 310 about the seed point 308 and evaluating points 312, 314, 316, and 318 on the circumference of that circle 310. Instead of evaluating all points on the circumference, only the selected points 312~ 314, 316, and 318 are evaluated to locate a segment that contains the desired value. The segment is successively subdivided .. . . . .
until a point equal to or suf~iciently close to the desired value is located. Subdivision is accomplished by mathematically rotating the search vector.
As shown in Figure 29, when this method is applied to searching, the seed point-308 must be found first. At the first stage of searching, four points 312, 314, 316, and 318, on the circle 310 are evaluated and each segment between two points is checked for the g~ven ','; ` ~ ' , ' ' ' ' " ' '` . '' ' i ' ' ' ' ``
.. .
' ~ ' ' ~, ' ' ` I ' ' WO90/15376 2 ~ 5 8 6 2 ~- PCT/US90/03209 value. If a segment is found, it will be subdivided until the given tolerance or cut-off criterion is reached. In the process, the searching radius can be changed if necessary. Note that for successive steps, the point found at tha given radius becomes the new seed point and the old seed point 308 is used as the first poin~ on the search circle.
Figure 30 shows a surface 320 as defined by 4 x 2 normalized parametric patches A-H. A surface S(x,y,z), in cartesian coordinates, is the mapping of a surface P(u,w), in the parametric domain, wherein P(u,w) is the equation of a uniform bi-c~bic B-spline surface. Each patch A-H has 4 x 4 subpatches. A curve 322 passes through normalized parametric patches B, F and G. Figure 3l is a magnification of region B. Initially, a seed point 324 must be found, using any convenient method such as a bisection algorithm~ If a seed point 324 is found between subpatch 2 and 3, there are -two intersection points between the cur~e 322 and the searching circle 326.
If subpatch 3 i8 selected for processing, the edge between - 2 and 3 is marked. Subpatch 3 is selected by examining normal vectors. The center of the searching:circle 326 is updated by moving it to the point most recently located on the curve 322. If the searching circle 326 meets the boundary of the normalized parametric patch B, subpatch 3 is marked as processed and the center of the searching circle 326 moves to the point in subpatch 2 which was found earlier. Then, subpatch 2 is :processed. When the searching circle 326 passes the edge between subpatch 2 and 6,~!subpatch 2 is marked as processed. If thè
searching circle 326 meets the boundary of normalized parametric patch B, the seed point on the boundary of normalized parametric patch F ~ is computed from calculations based on the normal vector.- This sequence continues until all patches are processed.

.. , .. , , ' ~ ;: , ., , .- . .

20~2~
` ` WO90/15376 ~CT/US90/03209 Contourin~
Figure 32 describas a method for contouring surface patch2s in the second preferred embodimen~.
Contouring involves finding an intersect.ion curve be~ween a surface patch and a contour plane. The suxface patches are contoured to provide a plurality of offset tool paths for the machine tool. The CAD/CAM software first identifies all surface patches through which a contour line pas~es (328). An edge of the identified surface patch is found (330). A seed point is selected on the edge (332). The seed point may be found, for example, by bisecting the edge. The contour line i5 propagated ~rom the seed point using a segmentized searching circle until it processes all of the identified surface patches (334).
Figure 33 describes a method for creating offset tool paths in the second preferred embodiment. Offsets can be applied to regions automatically selected by the CAD/CAM software. Alternatlvely, offsets can be applied to regions manually selected by an operator. The CAD/CAM
software selects an of~set angle and an offset heicJht from the contour llne (336). The of~set angle and the offset height determine the offset distance (338). A data point is projected from the ~contour line onto a pIane to simplify normal vector calculation (340). A noxmal vector is calculated at this data point (342). A new offset data point is determined by shifting the data point offset a distance along the normal vector (344).

Tool Interference Check Figure 34 describes a method fox preven~ing tool interference in the ~econd preferred embodiment. The method prevents the machine tool from erroneously intersecting with one of the surface patches. ~he CAD/CAN
software checks the minLmum radii of principal curvature at a plurality of ~ertices-of the ~urface patches (346).
The vertices where the minimum radius is smaller than the tool radius axe idantified (348). The CAD/CAM software .. . , " :

. '.. .. . . . .. .

20~862~ .
W090/15376 PCT/~S9~/0320s'' finds a curve of intersection of each identified vertex with a plane. The plane passes through the identified vertex and contains a first vector normal to the offset patch and a second vector in the direction of the minimum S radius (350). A plurality of line segments of the intersection curve are generated, which line segments intersect together ~352). The line segments are subdivided using a segmentized searching circle, until an error value between the line segments and the surface patch approachPs a predetermined tolerance value (354~.
A segmentized searching is computed circle for the two adjacent surface patches (356). An intersection between a plurality of triangular segments of the segmentized searching circles is found (358). The triangular segments are bisected (360). An intersection between the triangular segments is formed (362). These steps are repeated (i.e.~ the segmentized searching circle computing ; step (356) through the intersect~on finding step (362)) until an error value between a plurality of intersected edges of the triangular segments and the surface patches approaches the predetermined tolerance (364~. The segmentized searching circles are propagated until the circles meet or until the circles reach one of a plurality of boundaries for the surface patches (366). These steps are repeated (i.e., the curve of intersection finding step (350) through the segmentized searching circle propagating step (366)) until all identified ~ertices are processed (368).
, .: -30, NACHINE TOOL OPTINIZATION
Figure 35 describes a method for optimizing ,, offset tool paths in both preferred embodiments. Figure , 36 illustrates offset tool paths 386 and 387. The CAD/CAM
~- software can optimize the step length of the cutter 396 along the tool paths 386 and 387 (370). Since the desired surface 384 is known (372), along with a possible tool path 387 of length 388 (374), the maximum error 390 can be . . ... . . . .

~ - -... . . . . . .
: . . ~ . . .
~ - . .
',; : `: :. ' ` , : ' : . ., : . . . . . . .

205g~24 ~
woso/l5376 - PCT/US90/032~9 obtained (376). If the maximum exror 3gO i~ larger than an allowable value (380), the tool path length 388 is subdivided (382) until the calculated error 394 is less than an allowable value (384). ~he reduction in tool path step length 392 is dete~mined by the known surface 384 and the allowable value for the calculated error 394.
Figure 37 describes a method for optimizing the cutter feed rate in both preferred embodiments. Based on the characteristics of the machine tool ( 98), the cutter (400), a spindle speed (402), a feed rate (404), a depth of cut (406), and the type of material machined (408), the depth of cut is calculated from known surface geometries ~;
(410). An expected cutting force is calculated for the feed rate in the immediately preceding tool path increment }5 (412). This expected force is compared with a limi~ing value of force based on cutter size and material t414).
Feed rate in the current tool movement increment is ad~usted to produce an estimated cutting force of a ~ prede~ined fraction of the limiting force value (416).
Figure 38 describes a method for optimizing scallop heights in both preferred embodiment. Figure 39 graphically illustrates the optimization of the scallop height. The machining parameters include - the characteristics of the tool path (418), the distance between tool paths (420), the spindle speed (422), the feed rate (424), the characteristics of ~he material - (426), and the size of the cutter 446 (428). The scallop height 436 is compared with a predetermined maximum value (430), and if the scallop height 436 exceeds the predetermined maximum value (432), the distance between ad~acent tool paths 438 is reduced and the offset tool paths are adjusted (434). The distance 438 is specific to the cutter geometry and machined surface profile, particularly the cutter radius 440-and surface curvature 442. `- - `` ~

::' WO90/15376 2 0 5 8 6 2 ~ `.!'' '. ~, , FI~TURIN~
.
Figure 40 describes a method for fixturing in both pref~rred emb~iments. A blank of the workpiece material, of appropriate size, is chosen ~448). The blank is oriented according to the spatial orien~ation of the set of coordinates (450). The workpiece material is held on a bed surface of the machine tool to 'prevent movement while the machine tool operates (452).
The blank also may be selected from a plurality of rough cut, preformed blanksO These preformed blanks, which correspond to generic forms stored in the database, eliminate the need for multiple machining passes for each region. The correct blanXs preformed blanks are selected by the CADlCAM software.
A modified vise may be attached to the bed of the machine tool to hold the blank. A special fixture is milled, representing the coniguration of where the reproduction i5 to fit. When uqed to produce dental prostheses, this configuration represents the prepared tooth (similar to a die in the wax casting technique).
The internal surface and the external surface of the reproduction are milled to the heights of contour from another blan~. The partially fabricated reproduction is removed ~rom the vise and placed onto the fixture. When used to produce dental prostheses, this preserves the relative orientation between the internal and external surfaces when the final occlusal surface is mi~chined. The partially fabricated reproduction i8 held in place with either sticky ~ax or epoxy.~ The fixture is held with the vise while the remaining portion of the external surface is cut. The fixture can also ser~e as a holding device and can be used during shipping as protection for the margins. This protection is probably not critical for metal reproductions, but may ver~ practical for ceramic reproductions which are brittle and more susceptible to fracture and cracking.
Alternatively, as described hereinbefore, one or ;. . ... :
- .

': ,, :.`' . . , ' :
' ' ' . - , .
..

WO90/~5376 2 ~ ~ 3 ~ 2 ~ PCT/US90/03209 more parting lines can di~id~ the three dimensional representation of the object into top and bottom portions.
The bottom portion is machined first. The object is then flipped and the bottom portion is secured onto the fixture. Finally, the top portion is machined, completing the reproduction.

MACHINING
Figure 41 describes a method of machining used in ~oth preferred embodiments. ~he machining operation begins with the transmittal of the tool paths and commands from the CAD/CAM software to the machine tool (454). The machine tool machines both rough and finish surfaces as transmitted (456,458). The tool paths are transmitted by region, so that the machine tool, may if desired, machine a blanX by region according to the tool paths received.
Figure~ 42A, 42B, 42C, and 43 illustrate the machining process. The surface of the raw stock or blank 460 and the offset 464 of the part surface 466 are the geometrical constraint3 that determine the tool paths for the rough (i~ required) and finish cutting. The offset value 464 is determined by the tool radius plus the material left for finish cutting. Rough cutting is the process of removin~ the bulk of the material 4S2 from the raw stock 460, using the full capability of the machine tool. The factors related to roughing are the available spindle power, tools, and material properties. The tool path for roughing depends on the shape of ~he raw stock 460 and the of~set surface 464.
The cutting plane 468 moves down according to the maximum cutting depth along the Z axis. For each cutting plane 468, the intersection lines between the raw stock 460 and the offset surface 464 have to be calculated to ., determine the roughing area 4620 The roughing area 462 can be removed by contour cutting. Starting at a tool path that produces the desired contour, the machining continues outward until all rough cutting material is ~ . . . ~ , .
~::: ~, : ~' - ' ,, - .
- : ::: . : ..... : . . . . . . . .. . . .. . ; . . . .

W090/15376 2 0 ~ 8 6 2 ~ PCT/US90/03209 removed. If the machine tool meets the offset surface 464, it is retracted, moved to a position above the next rough cutting tool path and lowered to the cutting plane 468 to continue the rough cutting. The process repeats until all the rough cutting area 462 is removed.
The availability of detailed descriptions of the reproduction to be machined in numerical format enables different cutting paths other than contour cutting paths to be easily generated. Numerical procedures can be used to produce any sequence of points on the surface and hence any desired cutting tool path between points on the surface. Particular applications are cutting tool pa~hs which directly follow the part scanning paths, i.e., radial tool paths that eliminate computation steps, spiral tool paths, and raster tool paths. In spiral cu*ting, tool movement is around the vertical axis of the blank held on the tool bed. If the radial tool position is incremented ~or each tool movement step, an approximately ~piral tool path results. Finish cutting, as shown in Figure 43, removes khe small amount of excess material left during the roughing process, thus bringing the surface of the reproduced object 472 to the required - geometry and surface finish. -Usually, in finish cutting, the feed rate is low and spindle speed high. The tool path for the finish cutting is determined by finding the intersection lines between the offset surface and the X-Z
plane in figure 42A. Polishing is the final process of eliminating all machining marks and shining the surface.
This process can be done by using a rotating brush, grinding tool, or it can be performed manually.
Figure 44 illustrates both one possible ~ixturing method and a three axis machine tool. A blank workpiece of appropriate size is chosen and placed in the adaptable fixture 478. The holder 478 secures the workpiece in place. A region of the workpiece is machined using the cutting tool 474. The workpiece is removed from the .. ~ . . .~.. . . . . .
... . . ~ ,., .: :
: ' . ' . !
;' ' ' "` ' ' ` '~' ' ' ~
" ~ ' ~, . . . .
'.' ' , '' .' .'.
~' . ' ' . ' ~ ,, 2~5g~2-4 `

fixture 478 and re-oriented. The worXpiece may also be ~ :
~ecured in different size fixtures 4B0-490.
Although two embodiments have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that any arrangement which is calculated to achieve the s2~e purpose may be substituted for the specific embodiments shown. This application is intended to cover any adaptations or v2~riations of the present invention. Therefore, it is manifestly intended that this invention be limited only by the claims and equivalence thereof.

, ..
. . .
.. ..
... ..

.' ~ ' .. , : . .... ,: : ~ -.. . i . . : ; i : :~ . . , : .
... . ...
': :' . : : ~ . , , : . , . . ,: ., . : .
.:.. - . . . . : , 20~8~2~ .~

APPENDIX
Hereby inco~porated by reference for the purpose of providing demonstrative evidence of the level of skill in the art are the following citations:
S B-Splines and Inverse Alqorithms:
* D.H. Ballard and C.N. Brown, ComPuter Vision, Prentice-Hall, Inc., }g82, p. 239-24:3.
* D.F. Rogers and J.A. Adams, Mathematical Elements for Com~uter Graphics, Prentice-Hall, Inc., 1976, p.
144-155.
* P. Lancaster and K. Salkauskas, Curve and Surface Fittinq, Academic Press, 1986, p. 87-112.
* S. Peterson, X. Zhu, and D. Riley, A New Alqorithm for Constructinq B-Spline Surfaces, Computers in Enginee.ring 1984, Vol. 3, Proceedings of the ASNE Third International Computers in Engineering Conference, Chicago, Illinois, August 1984, p. 137-144 Surface Patches:
* D.F. Rogers and J.A. Adams~ Nathematical Elements 2~ for comPuter GraPhics~ McGraw-Hill, 1976, p. 157-187.
* P. Lancaster and K. Salkauskas, Curve and Surface Fittina, Academic Press, 1986, p. 245-271.
Linear Transformations:
* D.F. Rogers and J.A. Adams, Mathematical Elements for comPuter GraPhics~ McGraw-Hill, 1976, p. 5.
* D.H. Ballard and C.M. Brown, Computer Vision, Prentice-Hall, Inc. 1982, p. 473.

... , . . ~
~.;. : .- , -..
- . , . ,-, ~, . :: ,: .. . .. . - .
". ,. . , . . ,:, : .
. -: . . . .
-:,: . ...... . , - . . . . .
. , - . . . - . . . .

Claims (61)

WHAT IS CLAIMED IS:

1. A method of generating at least one parting line in a computer such that a surface representation of a 3-dimensional first object is divided into a plurality of regions, comprising:
(a) finding an X,Y,Z coordinate on a radial scan line nearest the center of a first object;
(b) finding a point on said radial scan line with a largest Y distance from the Z axis;
(c) clipping coordinates on said radial scan line below said parting point; and (d) repeating the finding step (a) through the clipping step (c) until all said radial scan lines have been processed in the computer.

2. A method for superimposing a first set of coordinates representing a 3-dimensional first object onto at least one generic form of said first object to create at least one altered generic set of coordinates, comprising:
(a) retrieving said generic form from the computer, said generic form. comprised of a generic set of coordinates;
(b) spatially rotating said generic form in the computer, such that said generic form corresponds to a spatial orientation of said first set of coordinates;
(c) scaling said generic form in the computer, such that said generic set of coordinates is sized substantially the same as said first set of coordinates;
(d) selectively creating at least one additional coordinate in said altered generic set of coordinates in the computer to emphasize at least one feature; and (e) selectively modifying said generic set of coordinates in the computer using a shaping function;
(f) modifying at least one coordinate of said altered generic set of coordinates in the computer, when at least one second object exists in proximity of said first object, said modified coordinate ensuring that any interference between said first object and said second object is avoided.

3. The method of claim 2, wherein the scaling step (d) comprises scaling said generic set of coordinates in the computer as a function of at least one measurement of at least one other object in proximity to said first object.

4. The method of claim 3, wherein said measurement is at least one gap measurement between said first object and said other object.

5. The method of claim 3, wherein said measurement is at least one height measurement for said other object.

6. The method of claim 3, wherein said measurement is at least one width measurement for said other object.

7. The method of claim 1, wherein the selectively creating step (e) comprises:
(1) identifying at least one feature coordinate within said altered generic set of coordinates; and (2) plurally representing said feature coordinate within said altered generic set of coordinates.

8. The method of claim 7, wherein the plurally representing step (2) comprises computing a plurality of parametric curves in the computer through said feature coordinate, thereby preventing said feature from being smoothed out or eliminated from said generic form by the modifying step (g) of claim 1.

9. A method of generating a plurality of contour lines in a computer that provide, paths for a machine tool, comprising:
(a) reparameterizing a plurality radial scan lines in a computer, thereby ensuring substantially the same number of coordinates on each of said radial scan lines;

(b) connecting said coordinates between adjacent said radial scan lines according to the order of said coordinates, thereby creating contour lines;
(c) repeating the reparameterizing step (a) through the connecting step (b) until all said coordinates on all adjacent said radial scan lines are connected.

10. An apparatus for generating at least one parting line such that a surface representation of a 3-dimensional first object is divided into a plurality of regions, comprising: (a) at least one computer;
(b) at least one machine tool operatively connected to said computer;
(c) means for finding an X,Y,Z coordinate on said radial scan line nearest the center of said first object;
(d) means for finding a point on said radial scan line with a largest Y distance from the Z axis; and (e) means for clipping coordinates on said radial scan line below said parting point.

11. An apparatus for superimposing a first set of coordinates representing a 3-dimensional first object onto at least one generic form of said first object to create at least one altered generic set of coordinates, comprising:
(a) at least one computer;
(b) means for retrieving said generic form from the computer, said generic form comprised of a generic set of coordinates;
(c) means for spatially rotating said generic form in the computer such that said generic form corresponds to a spatial orientation of said first set of coordinates;
(d) means for scaling said generic form in the computer, such that said generic set of coordinates is sized substantially the same as said first set of coordinates;
(e) means for selectively creating at least one additional coordinate in said altered generic set of coordinates in the computer to emphasize at least one feature; and (f) means for selectively modifying said generic set of coordinates in the computer using a shaping function;
(g) means for modifying at least one coordinate of said altered generic set of coordinates in the computer, when at least one second object exists in proximity of said first object, said modified coordinate ensuring that any interference between said first object and said second object is avoided.

12. The apparatus of claim 11, wherein the means for scaling (d) comprises means for scaling said generic set of coordinates in the computer as a function of at least one measurement of at least one other object in proximity to said first object.

13. The apparatus of claim 12, wherein said measurement is at least one gap measurement between said first object and said other object.

14. The apparatus of claim 12, wherein said measurement is at least one height measurement for said other object.

15. The apparatus of claim 12, wherein said measurement is at least one width measurement for said other object.

16. The apparatus of claim 11, wherein the means for selectively creating (f) comprises:
(1) means for identifying at least one feature coordinate within said altered generic set of coordinates;
and (2) means for plurally representing said feature coordinate within said altered generic set of coordinates.

17. The apparatus of claim 16, wherein the means for plurally representing (2) comprises means for computing a plurality of parametric curves in the computer through said feature coordinate, thereby preventing said feature from being smoothed out or eliminated from said generic form by the means for modifying (g) of claim 11.

18. An apparatus for generating a plurality of contour lines to provide a plurality of tool paths for a machine tool, comprising:
(a) at least one computer;
(b) means for reparameterizing said radial scan lines in the computer, thereby ensuring substantially the same number of coordinates on each of said radial scan lines;
and (c) means for connecting coordinates in the computer between adjacent said radial scan lines according to the order of said coordinates, thereby creating contour lines.

19. A method of scanning at least one 3-dimensional first object and recording into a computer at least one first set of coordinates, comprising:
(a) positioning a scanning head for recording X and Y
coordinates for said first object;
(b) determining a Z coordinate;
(c) recording said X, Y, and Z coordinates into the computer as said first set of coordinates;
(d) repeating steps (a) through (c) until a radial scan line is recorded into the computer;
(e) rotating said first object for scanning another radial scan line;
(f) repeating steps (a) through (e) until the surface of said first object is recorded into the computer as said first set of coordinates; and (g) filtering said first set of coordinates in the computer to eliminate noise by eliminating any of said first set of coordinates corresponding to points not substantially proximate to said radial scan lines.

20. An apparatus for scanning at least one 3-dimensional first object and recording at least one first set of coordinates, comprising:
(a) at least one computer;
(b) means for positioning a scanning head for recording X and Y coordinates for said first object;
(c) means for determining a Z coordinate;
(d) means for recording said X, Y, and Z coordinates into the computer as said first set of coordinates;
(e) means for rotating said first object for scanning another radial scan line; and (f) means for filtering said first set of coordinates in the computer to eliminate noise by eliminating any of said first set of coordinates corresponding to points not substantially proximate to said radial scan lines.

21. A method for positioning a dental prosthesis based on at least one landmark, comprising:
(a) finding at least one relative position between the upper and lower teeth using at least one impression of the upper and lower teeth;
(b) digitizing and recording into the computer said impression;
(c) digitizing and recording into the computer said landmark in said impression;
(d) digitizing and recording into the computer the upper and lower teeth from opposite sides. of said impression;
(e) digitizing and recording into the computer the position of said landmark; and (f) calculating the relative position of the upper and lower teeth in the computer based on the position of said landmark.

22. The method of claim 21, wherein said landmark is pre-existing.

23. The method of claim 21, wherein said landmark is added at the digitizing step (c).

24. An apparatus for positioning a dental prosthesis based on at least one landmark, comprising:
(a) at least one computer;
(b) means for finding at least one relative position between the upper and lower teeth using at least one impression of the upper and lower teeth;
(c) means for digitizing and recording into the computer said impression;
(d) means for digitizing and recording into the computer said landmark in said impression;
(e) means for digitizing and recording into the computer the upper and lower teeth from opposite sides of said impression;
(f) means for digitizing. and recording into the computer the position of said landmark; and (g) means for calculating the relative position of the upper and lower teeth in the computer based on the position of said landmark.

25. The apparatus of claim 24, wherein said landmark is pre-existing.

26. The apparatus of claim 24, wherein said landmark is added by the means for digitizing (c).

27. A method for digitizing the shape of at least one tooth from at least one cast of said teeth mounted in a dental articulator, comprising;
(a) digitizing and recording into a computer at least one shape representation of an occlusal surface of said cast;
(b) digitizing and recording into the computer at least one position representation of a hinge axis of said dental articulator;

(c) digitizing and recording into the computer at least one relative position value of said cast in occlusion; and (d) calculating in the computer an amount of space available in said dental articulator into which a dental restoration can fit using said shape representation, said position representation, and said relative position value.

28. An apparatus for digitizing the shape of at least one tooth from at least one cast of said teeth mounted in a dental articulator, comprising;
(a) at least one computer;
(b) means for digitizing and recording into the computer at least one shape representation of an occlusal surface of said cast;
(c) means for digitizing and recording into the computer at least one position representation of a hinge axis of said dental articulator;
(d) means for digitizing and recording into the computer at least one relative position value of said cast in occlusion; and (e) means for calculating-in the computer an amount of space available in said dental articulator into which a dental restoration can fit using: said shape representation, said position representation, and said relative position value.

29. A method of scanning and manipulating at least one set of coordinates recorded into a computer to represent at least one 3-dimensional first object, comprising:
(a) recording X, Y, and Z coordinates from a point on a radial scan line into the computer;
(b) repeating step (a) until an entire radial scan line is recorded into the computer;
(c) rotating the first object and repeating steps (a) through (b) until the entire surface of the first object is recorded into the computer as the first set of coordinates; (d) superimposing the first set of coordinates onto a generic form of the first object stored in the computer; and (e) contouring the superimposed coordinates such that a plurality of tool paths are created for a machine tool.

30. A method of scanning and manipulating at least one set of coordinates recorded into a computer to represent at least one 3-dimensional first object, comprising:
(a) projecting a plurality of profile lines onto the first object;
(b) scanning each of said profile lines to generate and record into the computer X, Y, and Z coordinates from a point on said profile line;
(c) repeating steps (a) through (b) until the entire surface of the first object is recorded into the computer as the first set of coordinates;
(d) superimposing the first set of coordinates onto a generic form of the first object stored in the computer;
(e) generating surface patches from the superimposed coordinates to represent the surface of the first object in the computer; and (e) contouring said surface patches such that a plurality of tool paths are created for a machine tool.

31. A method of machining a reproduction of a 3-dimensional first object such that the reproduction substantially duplicates all surfaces of said first object, comprising:
(a) radially scanning said first object and recording into a computer at least one first set of coordinates;
(b) generating at least one parting line in the computer such that said coordinates are clipped below said parting line, thereby dividing said object into a plurality of regions;
(c) superimposing said coordinates onto at least one generic form of said first object in the computer to create at least one altered generic set of coordinates;

(d) connecting said altered generic sets of coordinates at a common edge, when said common edge exists, thereby creating a master set of coordinates;
(e) generating contour lines in the computer between said coordinates on adjacent said radial scan lines;
(f) selectively offsetting said contour lines in the computer to provide a plurality of offset tool paths for said machine tool;
(g) preventing said offset tool paths from erroneously intersecting with the surface of the representation;
(h) generating at least one command directing said machine tool to machine a material into the reproduction of said first object;
(i) fixturing said material to be shaped into the reproduction of said first object; and (j) transmitting said tool paths and commands from the computer to said machine tool such.

32. The method of claim 31, wherein the radially scanning step (a) comprises:
(1) positioning a scanning head for recording X and Y
coordinates from said first object;
(2) determining a Z coordinate;
(3) recording said X, Y, and Z coordinates into the computer as said first set of coordinates;
(4) repeating steps (1) through (3) until a radial scan line is recorded into the computer;
(5) rotating said first object for scanning another radial scan line;
(6) repeating steps (1) through (5) until the surface of said first object is recorded into the computer as said first set of coordinates; and (5) filtering said first set of coordinates in the computer to eliminate noise by eliminating any of said coordinates corresponding to points not substantially proximate to said radial scan lines.

33. The method of claim 31, wherein the generating step (b) comprises:
(1) finding an X,Y,Z coordinate on said radial scan line nearest the center of said first object;
(2) finding a parting point on said radial scan line with a largest Y distance from the Z axis;
(3) clipping any coordinates on said radial scan line below said parting point; and (4) repeating the finding step (1) through the clipping step (3) until all said radial scan lines have been processed in the computer.

34. The method of claim 31, wherein said generating step (b) comprises generating said parting line in the computer such that said master set of coordinates is divided into said regions to prevent undercuts by said machine tool in the machining of the reproduction of said first object.

35. The method of claim 31, wherein the superimposing step (c) comprises:
(1) retrieving said generic form from the computer, said generic form comprised of a generic set of coordinates;
(2) spatially rotating said generic form in the computer, such that said generic form corresponds to a spatial orientation of said first set of coordinates;
(3) scaling said generic form in the computer, such that said generic set of coordinates is sized substantially the same as said first set of coordinates;
(4) selectively creating at least one additional coordinate in said generic set of coordinates in the computer to emphasize at least one feature;
(5) selectively modifying said generic set of coordinates in the computer using a shaping function until said generic form takes a desired shape;

(6) modifying at least one coordinate of said altered generic set of coordinates in the computer, when at least one second object exists in proximity to said first object, said modified coordinate ensuring that any interference between said first object and said second object is avoided.

36. The method of claim 35, wherein the scaling step (3) comprises altering said generic set of coordinates in the computer as a function of at least one measurement of at least one other object in proximity to said first object.

37. The method of claim 36, wherein said measurement is at least one gap measurement between said first object and said other object.

38. The method of claim 36, wherein said measurement is at least one height measurement of said other object.

39. The method of claim 36, wherein said measurement is at least one width measurement of said other object.

40. The method of claim 35, wherein the creating step (5) comprises:
(a) identifying at least one feature coordinate within said altered generic set of coordinates; and (b) plurally representing said feature coordinate within said altered generic set of coordinates.

41. The method of claim 40, wherein the plurally representing step (b) comprises computing a plurality of parametric curves in the computer through said feature coordinate, thereby preventing said feature from being smoothed out or eliminated from said generic form by the contouring step (e) of claim 31 or the modifying step (6) of claim 35.

42. The method of claim 35, wherein the selectively modifying step (5) comprises:
(a) identifying at least one coordinate to move within said generic set of coordinates; and (b) changing said generic set of coordinates with local effects using a free-form deformation technique.

43. The method of claim 31, wherein the generating step (e) comprises:
(1) reparameterizing said radial scan lines in the computer, thereby ensuring substantially the same number of coordinates on each of said radial scan lines;
(2) connecting coordinates in the computer between adjacent said radial scan lines according to the order of said coordinates, thereby creating contour lines;
(3) repeating the reparameterizing step (1) through the connecting step (2) until all said coordinates on all adjacent said radial scan lines are connected.

44. The method of claim 31, wherein the selectively offsetting step (f) comprises:
(1) selecting an offset angle and an offset height from said contour line stored in the computer;
(2) determining an offset distance from said offset angle and said offset height from said contour line in the computer;
(3) projecting at least one data point from said contour line onto a plane in the computer to simplify normal vector calculation;
(4) calculating at least one normal vector at said data point in the computer; and (5) determining at least one new offset data point by shifting said data points said offset distance along said normal vector in the computer.

45. The method of claim 31, wherein the selectively offsetting step (f) further comprises selectively offsetting said contour lines in said regions selected by the computer.

46. The method of claim 31, wherein the selectively offsetting step (f) further comprises selectively offsetting said contour lines in said regions selected by an operator.

47. The method of claim 31, wherein the preventing step (g) comprises:
(1) checking a minimum radii of principal curvature at a plurality of vertices of said offset tool paths in the computer;
(2) identifying said vertices in the computer when said minimum radius is smaller than a radius of said machine tool;
(3) finding a curve of intersection of each said identified vertex with a plane in the computer, aid plane passing through said identified vertex and containing a first vector normal to a surface of the representation and a second vector in the direction of said minimum radius;
(4) generating a plurality of line segments of said intersection curve in the computer, which line segments intersect together;
(5) subdividing said line segments in the computer by means of at least one segmentized searching circle until an error value between said line segments and said surface approaches a pre-determined tolerance value;
(6) computing at least one segmentized searching circle in the computer on a first and an adjacent second said surface patch;
(7) finding an intersection in the computer between a plurality of triangular segments of each of said segmentized searching circles;
(8) bisecting said triangular segments in the computer;
(9) finding an intersection between said triangular segments in the computer;
(10) repeating the computing step (6) through the finding step (9) until an error value between a plurality of intersected edges of said triangular segments and said surface patches approaches said pre-determined tolerance;
(11) propagating said segmentized searching circles in the computer until said segmentized searching circles meet or until said segmentized searching circles reach one of a plurality of boundaries for said surface patches; and (12) repeating the finding step (3) through the propagating step (11) until all of said identified vertices are processed in the computer.

48. A method of preventing a plurality of tool paths generated in a computer from erroneously intersecting with one of a plurality of surface patches representing a first object to be machined, comprising:
(a) checking a minimum radii of principal curvature at a plurality of vertices of said surface patches in the computer;
(b) identifying said vertices in the computer when said minimum radius is smaller than a radius of said machine tool;
(c) finding a curve of intersection of each said identified vertex with a plane in the computer, said plane passing through-said identified vertex and containing a first vector normal to said surface patch and a second vector in the direction of said minimum radius;
(d) generating a plurality of line segments of said intersection curve in the computer, which line segments intersect together;
(e) subdividing said line-segments in the computer by means of at least one segmentized searching circle until an error value between said line segments and said surface patch approaches a pre-determined tolerance value;
(f) computing at least one segmentized searching circle in the computer on a first and an adjacent second said surface patch;
(g) finding an intersection in the computer between a plurality of triangular segments of each of said segmentized searching circles;

(h) bisecting said triangular segments in the computer;
(i) finding an intersection between said triangular segments in the computer;
(j) repeating the computing step (f) through the finding step (i) until an error value between a plurality of intersected edges of said triangular segments and said surface patches approaches said pre-determined tolerance;
(k) propagating said segmentized searching circles in the computer until said segmentized searching circles meet or until said segmentized searching circles reach one of a plurality of boundaries for said surface patches; and (l) repeating the finding step (c) through the propagating step (k) until all of said identified vertices are processed in the computer.

49. A method of optimizing a step length along a tool path in the computer, comprising:
(a) checking an error value in the computer between said tool path and at least one surface patch representing a material to be machined; and (b) dividing said step length in the computer until said error value is smaller than a pre-determined tolerance.

50. A method of optimizing a command for a machine tool in a computer comprising optimizing along at least one tool path by determining a cutter size for said machine tool in the computer based on the characteristics of said machine tool, a feed rate, a depth of cut, a spindle speed, and the characteristics of a material being machined.

51. A method of optimizing a plurality of tool paths for a machine tool in a computer, comprising:
(a) calculating a scallop height for a material to be machined based on a plurality of machining parameters, said machining parameters comprising the characteristics of said tool path, a distance measurement between said tool path and an adjacent said tool path, a spindle speed, a feed rate, the characteristics of said material, and a cutter size;
(b) comparing said scallop height with a pre-determined boundary value; and (c) altering at least one of said machining parameters when said scallop height exceeds said boundary value.

52. A method of fixturing a material to be shaped into a reproduction of a first object such that said material may be machined sequentially by a plurality of regions through alternate orientation of said material, comprising:
(a) selecting a blank of said material of appropriate size;
(b) orienting said blank according to a spatial orientation of said region in the computer; and (c) holding said blank in a fixture on a bed surface of said machine tool to prevent movement of said blank while said machine tool operates; and (d) repeating the orienting step (b) through the holding step (c) as needed.

53. The method of claim 52, wherein the selecting step (a) further comprises selecting from a plurality of preformed blanks, such that the need for multiple machining of said region is eliminated, thereby speeding the machining of the reproduction.

54. The method of claim 52, wherein said preformed blanks are represented in the computer and selected by the computer.

55. An apparatus for preventing a plurality of tool paths generated in a computer from erroneously intersecting with one of a plurality of surface patches representing a first object, comprising:

(a) at least one computer;
(b) means for checking a minimum radii of principal curvature at a plurality of vertices of said surface patches in he computer; (c) means for identifying said vertices in the computer when said minimum radius is smaller than a radius of said machine tool;
(d) means for finding a curve of intersection of each said identified vertex with a plane in the computer, said plane passing through said identified vertex and containing a first vector normal to said surface patch and a second vector in the direction of said minimum radius;
(e) means for generating a plurality of line segments of said intersection curve in the computer, which line segments intersect together;
(f) means for subdividing said line segments in the computer by means of at least one segmentized searching circle until an error value between said line segments and said surface patch approaches a pre-determined tolerance value;
(g) means for computing at least one segmentized searching circle in the computer on a first and an adjacent second said surface patch;
(h) means for finding an intersection in the computer between a plurality of triangular segments of each of said segmentized searching circles;
(i) means for bisecting said triangular segments in the computer;
(j) means for finding an intersection between said triangular segments in the computer;
(k) means for repeating the means for computing (g) through the means for finding (;) until an error value between a plurality of intersected edges of said triangular segments and said surface patches approaches said pre-determined tolerance;
(l) means for propagating said segmentized searching circles in the computer until said segmentized searching circles meet or until said segmentized searching circles reach one of a plurality of boundaries for said surface patches; and (m) means for repeating the means for finding (d) through the means for propagating (1) until all of said identified vertices are processed in the computer.

56. An apparatus for optimizing a step length along a tool path, comprising:
(a) at least one computer;
(b) means for checking an error value in the computer between said tool path and at least one surface patch representing a material to be machined; and (c) means for dividing said step length in the computer until said error value is smaller than a pre-determined tolerance.

57. An apparatus for optimizing a command for a machine tool in a computer comprising optimizing along at least one tool path by determining a cutter size for said machine tool in the computer based on the characteristics of said machine tool, a feed rate, a depth of cut, a spindle speed value, and the characteristics of a material being machined.

58. An apparatus for optimizing a plurality of tool paths for a machine tool, comprising:
(a) at least one computer;
(b) means for calculating a scallop height for said material based on a plurality of machining parameters, said machining parameters comprising the characteristics of said tool path, a distance measurement between said tool path and an adjacent said tool path, a spindle speed, a feed rate, the characteristics of said material, and a cutter size;
(c) means for comparing said scallop height with a pre-determined boundary value; and (d) means for altering at least one of said machining parameters when said scallop height exceeds said boundary value.

59. An apparatus for fixturing a material to be shaped into a reproduction of a first object such that said material may be machined sequentially by a plurality of regions through alternate orientation of said material, comprising:
(a) at least one computer;
(b) at least one machine tool operatively connected to said computer;
(c) means for selecting a blank of said material of appropriate size;
(d) means for orienting said blank according to a spatial orientation of said region in the computer; and (e) means for holding said blank in a fixture on a bed surface of said machine tool to prevent movement of said blank while said machine tool operates; and (f) means for repeating the means for orienting (d) and the means for holding (e) as needed.

60. The apparatus of claim 59, wherein the means for selecting (c) further comprises means for selecting from a plurality of preformed blanks, such that the need for multiple machining of said region is eliminated, thereby speeding the machining of the reproduction.

61. The apparatus of claim 60, wherein said preformed blanks are represented in the computer and selected by the computer.