DE10362329B4 - Recording method for an image of a recording object and recording device - Google Patents

Abstract

Method for producing an image of a recording object (20), in particular for dental purposes, wherein a plurality of recordings of a stripe pattern (22) projected onto the recording object, each with a different phase position, are generated as raw image (Ri), wherein for the suppression of periodic perturbations from the raw images (R1 , ..., Rm) at least two raw picture groups (R1, R2, ..., Rn; R2, R3, ..., Rn + 1) are formed, from each of which a phase image (Pj) of the recording object (20) is calculated wherein these phase images (P1, P2) are averaged such that a phase image (P) with reduced interference components is produced, and an image of the recording object is calculated from the phase image (P) with reduced interference components. According to the invention, the raw images (R1,..., Rm) are recorded in the field method, wherein a phase image (P1, P2) is calculated from each of the fields and these are averaged such that a phase image (P) with reduced high-frequency Disturbance shares arises. Furthermore, it is provided according to the invention to calculate contrast images instead of the phase images.

Description

Technical area

The invention relates to a method for producing an image of a recording object, in particular for dental purposes, with the method steps: a) projecting a stripe pattern onto the photograph object, b) recording the projected stripe pattern as a raw image with a camera, wherein steps a) and b) at a plurality of different positions of the phase position of the stripe pattern, and c) calculating an image of the photographing object from the plurality of mutually phase-shifted camera images. In particular, height images or contrast images come into question as images.

State of the art

By a direct optical three-dimensional measurement of one or more teeth in the mouth of a patient digital design data for the computer-aided production of dentures can be obtained without impression. Such, in analogy to the physical impression "optical impression" called, three-dimensional representation can be recorded without contact with one or more camera recordings. The camera can be guided freehand like an elbow or supported on the row of teeth with one hand or ambidextrous.

The measuring method uses the principle of active triangulation, in which a single light stroke or a stripe pattern of parallel light strokes is projected onto the object to be recorded by a projection device and the projected image is resumed at a parallax angle with an area camera.

Due to the surface structure of the teeth, the strip pattern no longer appears straight, but curved and shifted with respect to the straight course. From the position of the light stroke lines can be deduced the surface structure of the recording object. After the measuring process, a digital three-dimensional data model of the recording object is present in the memory of the computer, which can be displayed on a monitor as a video still picture or serve as the basis for a computer-controlled production of dental prostheses.

To increase the accuracy of the so-called phase-shift triangulation method (phase-shifting triangulation) can be used. Several raw images with different positions of the phase position of the grating are successively generated and from these a phase image is calculated. From the phase image can then be calculated using calibration data, a height image of the recording object. Different algorithms for calculating the phase image from the individual raw images are known. For example, three, four or five raw images with a relative phase shift of 120 ° or 90 ° can be recorded. From the raw images, a contrast image can also be calculated. For example, a contrast image of four raw images with the intensity values I ₁ ... I _{4 is} calculated according to the formula: K = √ ((I ₁ -I ₃ ) ² + (I ₂ -I ₄ ) ² ).

The phase-shift method leaves typical periodic perturbations due to tolerances or systematic error sources. These periodic disturbances occur at multiples of the grating frequency in the phase image and also in the contrast image or height image.

For example, a deviation from the linear movement of the light grid leads to a disturbance of 2 times the grid frequency, a change in the intensity of the grid illumination to a disturbance of 1 times the grid frequency, and deviations from the linearity between the light signal and the electrical signal to a disturbance 4 times the grid frequency.

In general, the more raw images are taken, the more error-resistant the calculation algorithms are. Nevertheless, the described periodic errors occur in spite of the significantly increased effort in different form.

To increase the quality of the phase image, for example, a filtering can be performed. However, a loss of resolution must be accepted. The quality improvement by averaging several phase images, however, has the disadvantage of a significantly extended recording time.

It is also possible to perform the control for recording the raw images very accurately, so that the periodic errors occur to a lesser extent. However, such precise control is very expensive. Also, external conditions such as temperature, humidity or gravity, as well as long-term effects affect the image acquisition and the achievable image quality.

From the DE 40 27 328 A1 are known apparatus and methods for measuring an object by means of triangulation, wherein two different triangulation angles are realized to avoid optical shadows or artifacts resulting from one-sided illumination.

Also from the US 4,837,732 A is known a phase shift triangulation method, which For the suppression of artifacts caused by pixel defects of the sensor, contrast images of raw image groups are calculated with a phase shift of 180 ° to one another.

From "Phase-measuring interferometry: new methods and error analysis" by C. Joenathan, Applied Optics, vol. 33, no. 19, pages 4147-4155, methods are known for analyzing and, for example, by reporting individual groups of raw image calibration errors or non-linear errors to reduce.

This is where the invention starts. The invention, as characterized in the claims, the object underlying the invention to provide a method of the type mentioned for receiving an image of a recording object, which provides a significantly increased accuracy over known techniques with relatively little effort.

Presentation of the invention

This object is achieved by the method according to claim 1 and the receiving device according to claim 13. Advantageous embodiments of the invention are the subject of the dependent claims.

• c₁) aus den Kamerarohbildern zumindest zwei Rohbildgruppen gebildet werden,
• c₂) aus jeder Rohbildgruppe ein Kontrastbild des Aufnahmeobjekts berechnet wird,
• c₃) die berechneten Kontrastbilder mit Wichtungsfaktoren versehen gemittelt werden, um ein Kontrastbild mit reduzierten Störungsanteilen zu erhalten, und
• c₄) aus dem Kontrastbild mit reduzierten Störungsanteilen ein Bild des Aufnahmeobjekts berechnet wird.

The invention is based on the state of the art in that for the purpose of suppressing periodic disturbances in method step c) the calculation of an image of the recording object from the plurality of mutually phase-shifted camera images:

• c ₁ ) at least two raw image groups are formed from the camera image frames,
• c ₂ ) a contrast image of the object to be photographed is calculated from each raw image group,
• c ₃ ) averaging the computed contrast images with weighting factors to obtain a contrast image with reduced perturbation levels, and
• c ₄ ) is calculated from the contrast image with reduced interference components, an image of the recording object.

According to the method c ₁ ) to c ₄ ), noise in a height image or in a contrast image can be suppressed.

The invention is thus based on the idea of dividing the raw images into two or more groups such that the perturbations of the contrast images calculated from the respective raw image groups have a certain phase shift relative to one another. These contrast images are appropriately averaged to thereby suppress or eliminate noise at particular multiples of the grid frequency. As a result, a significant increase in the measuring accuracy can be achieved with only a slight extension of the total recording time by additional raw image recordings.

In the method according to the invention, the raw images are each preferably recorded with a constant displacement of the phase position of the grating.

In a preferred embodiment of the method according to the invention, it is provided that (n + 1) raw images R ₁ , R ₂ ,..., R _{n + 1 are} recorded, successive raw images having a predetermined phase shift. There are two raw image groups R ₁ , R ₂ , ..., R _n ; R ₂ , R ₃ ,..., R _{n + 1 are} formed and there is a first contrast image P ₁ from the first raw image group R ₁ , R ₂ ,..., R _n and a second contrast image P ₂ from the second raw image group R ₂ , R ₃ , ..., R _{n + 1} calculated. The first and second contrast images, P ₁ and P ₂ , are averaged with equal weighting to obtain a contrast image P with reduced perturbation levels. In this case, n represents an integer greater than or equal to 3 in order to produce a contrast image using the known methods.

In another, likewise preferred embodiment of the method according to the invention, it is provided that (n + 2) raw images R ₁ , R ₂ ,..., R _{n + 2 are} recorded, wherein successive raw images have a predetermined phase shift. There are three raw image groups R ₁ , R ₂ , ..., R _n ; R ₂ , R ₃ , ..., R _{n + 1} ; R ₃ , R ₄ ,..., R _{n + 2 are} formed and a first contrast image P ₁ is selected from the first raw image group R ₁ , R ₂ ,..., R _n , a second contrast image P ₂ is selected from the second raw image group R ₂ , R ₃ , ..., R _{n + 1} and a third contrast image P ₃ is calculated from the third raw image group R ₃ , R ₄ , ..., R _{n + 2} . The first and third contrast images, P ₁ and P ₃ , are averaged with the same weighting to obtain an intermediate image P _z . The second contrast image P ₂ and the intermediate image P _z are averaged to obtain a contrast image P with reduced perturbation levels. Where n is an integer greater than or equal to 3.

In the two described embodiments, it is preferred if n is chosen equal to 4, because here the measurement accuracy and the required measurement time are in a particularly favorable relationship to each other.

According to a development of the method according to the invention, it is provided that the raw images are recorded in the field method, so that when the strip pattern is moved continuously, the two fields have a phase shift relative to one another, which amounts to half the phase shift of consecutive raw images.

In the field method, the image to be recorded is divided into two fields. The first field usually corresponds to the even-numbered lines, the second field corresponds to the odd-numbered lines. The two fields are recorded one after the other and consequently at different times.

Advantageously, in this context, a contrast image is calculated from each of the fields of a raw image and the two contrast images are averaged before further processing in order to obtain a contrast image with reduced high-frequency interference components. If, for example, the raw images are recorded with a continuous movement of the strip pattern causing the grating with a phase shift of 90 °, the result is a phase shift of the two fields of 45 °. Decisive here is the time center of the integration phase in the generation of the fields, wherein the time duration of the integration phase itself, ie the exposure of the receiver plays a minor role. Thus, by averaging the two phase and contrast fields, interference with the 4-times grid frequency can be suppressed. Although a loss of resolution must be taken into account, but in strong disturbances outweighs the positive effect of interference suppression. The averaging is preferably carried out with a moving average.

Suitably, in a method according to the invention, it is provided that an image of a predetermined test object is recorded before step a), and based on an analysis of the image of the test object a suitable scheme for the calculation of the interference-reduced contrast image is selected for the object to be photographed. Since periodic disturbances can occur sporadically as a function of external influences, as well as always be present due to the system, it can be achieved by this measure that an algorithm extending the recording time is only used if the corresponding disturbance is present during the recording.

The recording of the test object can take place during an anyway necessary calibration of the system for calculating contrast or height images from the contrast images without significant additional time. However, it can also be carried out separately so that the most favorable calculation scheme can be selected before each measurement of a recording object.

The inventive method provides in particular in such systems great advantages in which the recording object and the camera for receiving the projected fringe pattern are freely positionable to each other and freehand recordings are performed.

A particularly preferred application example is the recording of a height image of one or more teeth in the oral cavity of a patient, which is performed by a manual measurement in a short measuring time. The inventively achieved increase in measurement accuracy without significant extension of the recording time of particular advantage.

A device according to the invention for carrying out the method described comprises a projection device for projecting a stripe pattern onto the recording object, a camera for recording the projected stripe pattern as a raw image, and means for calculating an image of the recording object from a plurality of mutually phase-shifted camera images, which are combined to form at least two raw image groups ,

Further advantageous embodiments, features and details of the invention will become apparent from the dependent claims, the description of the embodiments and the drawings.

Brief description of the drawing

The invention will be explained in more detail with reference to an embodiment in conjunction with the drawings. In each case only the essential elements for understanding the invention are shown. It shows

1 a schematic representation of a recording device according to the invention according to the principle of phase shift triangulation;

2 a graph representing the phase φ of six raw images R _i with i = 1 .. 6 relative to a reference phase;

3 a graph showing the amplitude A of a fault with 2 times the grid frequency as a function of the phase φ;

4 a graph showing the amplitude A of a disorder with a lattice frequency as a function of the phase φ; and

5 a schematic drawing of an intraoral camera; and

6 a schematic drawing of a portion of the projection beam path of the intraoral camera 5 ,

embodiments

1 shows a schematic representation of a generally with 10 designated recording device, which operates on the principle of phase-shift triangulation. The projection device 12 generates a light grid with parallel grid lines on a recording object, a tooth in the embodiment 20 , is projected. Due to the three-dimensional surface structure of the tooth 20 the light grid lines appear curved on the tooth and at uneven intervals. The over the projection beam 13 projected image 22 is at a parallax angle with one in the observation beam path 15 arranged surface camera 14 recorded and for evaluation of an evaluation 16 fed. The projection device 12 and the camera 14 can also be in a structural unit 11 be summarized, as later for an intraoral camera in 5 is shown.

After recording a first raw image R ₁ further raw images are recorded with a defined phase shift, as in 2 illustrated. There, for six raw images R _i with i = 1,..., 6, the respective phase φ is shown relative to the reference phase 0 °. Two successive raw images each have a phase shift Δφ of 90 °. Due to the constant phase shift results in a linear course 30 the phase.

The control of the projection device 12 to generate the phase-shifted light grid is also done by the arithmetic unit 16 ,

In one embodiment of the invention, an algorithm based on the use of four raw images with 90 ° phase shift between successive raw images is assumed. The arithmetic unit calculates 16 in a manner known per se, the phase of a pixel from the intensities of the four raw images. According to the invention, five instead of four raw images R ₁ ,..., R ₅ , which are phase-shifted by 90 ° relative to one another, are recorded relative to the standard algorithm. The recording time is thereby extended by 25%.

A first contrast image P is calculated from the raw images ₁ and R ₂ to R _5, a second contrast image P ₂ then from the raw images R ₁ to R ₄ according to the standard algorithm. Since the individual raw images R ₁ , R ₂ to R ₄ , R ₅ to one another in each case have a phase shift of 90 °, and the two contrast images P ₁ and P ₂ to each other a phase shift of 90 °. Now the two contrast images P ₁ and P _{2 are} averaged. Disturbances with the 2-times grid frequency are strongly suppressed or eliminated.

This is in the graph of 3 illustrating the amplitude A of a perturbation with a 2-fold grid frequency as a function of the phase φ. The curve 32 (open squares) shows the course of the disturbance in the contrast image P ₁ , the curve 34 (filled circles) the shifted by 90 ° course of the disturbance in the contrast image P ₂ . Averaging the contrast images gives the curve 36 (Crosses), where the noise is eliminated at 2 times the grid frequency.

In a further embodiment of the invention, the algorithm based on four raw images with 90 ° phase shift is again assumed. In this exemplary embodiment, six raw images R ₁ ,... R ₆ , which are phase-shifted by 90 ° in each case, are recorded. The recording time is extended by 50% compared to the standard algorithm. Subsequently, three contrast images are calculated according to the standard algorithm, a first contrast image P ₁ from the raw images R ₁ to R ₄ , a second contrast image P ₂ from the raw images R ₂ to R ₅ , and a third contrast image P ₃ from the raw images R ₃ to R ₆ .

Since the individual raw images of the first and third contrast image R ₁ , R ₃ to R ₄ , R ₆ each have a phase shift of 180 ° to each other, also have the two contrast images P ₁ and P ₃ to each other a phase shift of 180 °. By averaging the contrast images P ₁ and P ₃ thus disturbances with the 1-fold grid frequency are strongly suppressed or eliminated. This is in the graph of 4 illustrating the amplitude A of a perturbation with a lattice frequency as a function of the phase φ.

The curve 42 (open squares) shows the course of the disturbance in the contrast image P ₁ , the curve 44 (filled circles) the shifted by 180 ° course of the disturbance in the contrast image P3. Averaging the contrast images gives the curve 46 (Crosses), where the interference is eliminated with a 1-fold grid frequency. Subsequently, the intermediate contrast image P _z obtained from the contrast images P ₁ and P ₃ can be averaged with the contrast image P ₂ as described above, in order to eliminate interference at twice the grid frequency.

The statements made with respect to the contrast images apply equally to the phase images.

Analogous to the described procedures, it is also possible to eliminate interference with the 4-times grid frequency by averaging two phase or contrast images, which were calculated from 45 ° out of phase raw images. In this case, a special property of field sensors can be exploited. A phase shift of the raw images by 90 ° results in a phase shift of 45 ° for both fields. This can be exploited by averaging the two phase and contrast fields to eliminate interference at 4 times the grid frequency. The loss of resolution is more than compensated for by strong gains by the gain in measurement accuracy.

Periodic disturbances can occur sporadically in response to external influences or always be present due to the system. Therefore, in a preferred embodiment, the occurrence of 1-time, 2-fold, or 4-times grid frequency perturbations is detected by Fourier analysis of a height image of a test object, for example, a plane.

Without additional time, this can be done during an already required calibration of the system. Then, depending on which multiples of the grating frequency are detected, a suitable calculation scheme can be selected, which only suppresses interference with the detected frequencies and thus can be carried out in the shortest possible time. The disadvantages of the described corrections, a longer recording time and / or a loss of resolution are thus not unnecessarily accepted. A longer recording time is often particularly disadvantageous if the measuring system and the recording object are freely positionable to each other.

An apparatus for carrying out the method described is in 5 in the form of an intraoral camera 1 shown. The intraoral camera can come to an end 2 in an oral cavity 3 a patient 4 introduced and the recorded image data of a recording object 20 can via a cable 6 be read out. About the evaluation unit 16 becomes a height image of the subject 20 from a plurality of mutually phase-shifted Kamerarohbilderern, which are combined to form at least two Rohbildgruppen calculated. The evaluation unit 16 may include components of a commercial PC. The camera 1 includes a projection beam path 13 for generating a stripe pattern on the object to be photographed 20 , About an observation beam 15 that will be on the subject 20 projected stripe patterns as a raw image from a receiver 17 added.

The intraoral camera off 5 contains an in 6 shown projection device 12 for projecting one by one in the projection beam path 13 arranged grid 19 generated stripe pattern on the subject. The grid 19 is about a bending vibrator 23 in one direction 24 , which is substantially perpendicular to the projection direction, moves. This usually takes place continuously, although the properties of the bending oscillator must be taken into account. Since these are small movements, production-related variations in the behavior of the bending vibrator have a not insignificant influence on this movement. In principle, it is also possible to carry out a sudden movement instead of a continuous movement, so that the fringe pattern stands still during the integration time, that is to say during the exposure of the image receiver.

Claims (13)

Method for creating an image of a recording object ( 20 ), in particular for dental purposes, comprising the steps of: a) projecting a stripe pattern onto the object to be photographed ( 20 b) picking up the projected fringe pattern ( 22 ) as a raw image (R _i ) with an image receiver ( 14 ; 17 ) at a different angle than the projection, wherein steps a) and b) are performed at a plurality of different positions of the phase position of the stripe pattern, and c) calculating an image of the recording object from the plurality of mutually phase-shifted camera images (R ₁ , ..., R _n ), characterized in that for the suppression of periodic disturbances in step c) c ₁ ) from the Kamerarohbilderern (R ₁ , ..., R _m ) at least two raw image groups (R ₁ , R ₂ , ..., R _n ; R ₂ , R ₃ , ..., R _{n + 1} ), c ₂ ) from each raw image group (R ₁ , R ₂ , ... R ₂ , R ₃ , ..., R _{n + 1} ) Contrast image (P _j ) of the subject ( 20 ), c ₃ ) the calculated contrast images (P ₁ , P ₂ ) are averaged such that a contrast image (P) with reduced perturbation components is produced, and that c ₄ ) calculates an image of the object to be photographed from the contrast image (P) with reduced perturbation components becomes. A method according to claim 1, characterized in that the calculated contrast images (P1, P2) provided with weighting factors are averaged. A method according to claim 1 or 2, characterized in that the raw images (R ₁ , R _m ) each with a constant displacement of the phase position of the grating ( 19 ). Method according to one of claims 1 to 3, characterized in that - (n + 1) raw images (R ₁ , R ₂ , ..., R _{n + 1} ) are recorded, wherein successive raw images have a phase shift, - two raw image groups (R ₁ , R ₂ , ..., R _n ; R ₂ , R ₃ , ..., R _{n + 1} ) are formed, - A first contrast image (P ₁ ) from the first raw image group (R ₁ , R ₂ , ..., R _n ) and a second contrast image (P ₂ ) from the second raw image group (R ₂ , R ₃ , ..., R _{n + 1} ), and - the first (P ₁ ) and second (P ₂ ) contrast images are averaged to obtain a contrast image (P) with reduced perturbation levels, where n represents an integer greater than or equal to 3. Method according to one of claims 1 to 3, characterized in that - (n + 2) raw images (R ₁ , R ₂ , ..., R _{n + 2} ) are recorded, wherein successive raw images have a phase shift, - three raw image groups (R ₁ , R ₂ , ..., R _n , R ₂ , R ₃ , ... R ₃ , R ₄ , ..., R _{n + 2} ) are formed, - a first contrast image (P ₁ ) the first raw image group (R ₁ , R ₂ , ..., R _n ), a second contrast image (P2) from the second raw image group (R ₂ , R ₃ , ..., R _{n + 1} ) and a third contrast image (P ₃ ) is calculated from the third raw image group (R ₃ , R ₄ ,..., R _{n + 2} ), and - the first (P ₁ ) and third (P ₃ ) contrast image are averaged to form an intermediate image (P _z ) to obtain the second contrast image (P ₂ ) and the intermediate image (P _z ) are averaged to obtain a contrast image (P) with reduced perturbation proportions, where n represents an integer greater than or equal to 3. A method according to claim 4 or 5, characterized in that n is chosen equal to 4. Method according to one of claims 1 to 6, characterized in that the raw images (R ₁ , ..., R _m ) are recorded in the field method, so that the two fields to each other have a phase shift. A method according to claim 7, characterized in that the two fields to each other have a phase shift, which is half the phase shift of consecutive raw images (R ₁ , ..., R _m ). A method according to claim 7, characterized in that from each of the fields of a raw image (R ₁ , ..., R _m ) a contrast image (P ₁ , P ₂ ) is calculated and the two contrast images (P ₁ , P ₂ ) before Further processing are averaged such that a contrast image (P) with reduced high-frequency interference components arises. Method according to one of claims 1 to 9, characterized in that before step a) an image of a predetermined test object is taken and based on an analysis of the image of the test object, a suitable scheme for the calculation of the interference-reduced contrast image is selected for the recording object. Method according to one of claims 1 to 10, characterized in that the receiving object ( 20 ) and the camera ( 1 ) for recording the projected fringe pattern ( 22 ) are freely positionable to each other. Method according to one of claims 1 to 11, characterized in that an image of one or more teeth ( 20 ) in the mouth ( 3 ) of a patient ( 4 ) is recorded by a manual measurement in a short measuring time. Apparatus for carrying out the method according to one of claims 1 to 12, comprising - a projection device ( 12 ) for projecting a stripe pattern on the object to be photographed ( 20 ), - a camera ( 14 ) for recording the projected fringe pattern ( 22 ) as raw image (R ₁ , ... R _m ), - means ( 16 ) for calculating an image of the photographing object ( 20 ) of a plurality of mutually phase-shifted Kamerarohbildern (R1, ... Rm) to form at least two Rohbildgruppen (R1, R2, ..., Rn; R2, R3, ..., Rn + 1) and - means for calculating a contrast image from the at least two raw image groups (R ₁ , R ₂ ,..., R _n ; R ₂ , R ₃ ,..., R _m1 ) in order to obtain a contrast image (P) with reduced perturbation proportions.

Images