WO1998008152A2 - Computer vision system modeling apparatus - Google Patents

Abstract

A vision system for use in modeling an environment or scene is presented. A user views a scene through an electronic camera and is presented with CAD type functions for creating graphical representation of the scene. As the apparatus has knowledge of where it is with respect to objects in the scene and with respect to where it has been previously, the model perspective and size can be continuously updated to reflect the view of the scene as the camera addresses from any point in space.

Description

Title. "Computer Vision System Modeling Apparatus"

Specification for a PCT Patent Application

Background of the Invention

This instant invention is generally concerned with computer vision systems and specifically concerned with computer vision systems combined with computer aided design facilities

The following disclosure is presented with the assumption that elements of computer aided design are well known to the reader. Texts explaining the current state of that are available from various sources, but those relating to AutoCad^® may be particularly helpful No attempt is made here to explain the details of implementing computer models with software except as it relates to the new arrangement presented Vision systems which combine electronic cameras with computers to perform image functions are described in various texts including co-pending patent applications of the present inventors having series 08 serial numbers 1 19,360, 270,91 1, 355,710, 335,940, 335,912, 41 1,299, 307,360

It should be understood that all of the herein referenced materials provide considerable definition of elements of the present invention Therefore, those materials are incorporated herein by reference whereby the instant specification can rely upon them for enablement of the particular teachings of each

Computer aided design and modeling techniques are useful in various fields Commonly referred to as CAD or computer aided design, an electronic computer can be used to model devices, objects or environments. Changes to the model are easily made and a great number of variations might be tried before arriving at a final desired result. Examples of uses of CAD include' by architects for buildings proposed to be built, or by automobile designers for cars proposed to be manufactured A designer may start from a conceptual idea, a photograph, an artist drawing or other medium. From that initial idea or representation, a designer may construct on a computer, a model. A model consists of a plurality of individual graphical objects each having some correspondence to some real object. A photograph suggests to the CAD designer how particular features of the subject may look and give hints to how they might be implemented in a graphical representation. With advanced computers, it is even possible for the photograph to be scanned into an electronic image and combined with the CAD drawing, thereby facilitating formulation of a more accurate model. Converting real world objects into a computer model is of great interest to CAD designers as it provides very sophisticated designs in early stages of development.

Novel techniques have been discovered which provide very specialized uses of vision systems, and in particular as they may relate to computer aided modeling and design. While the systems and inventions of the prior art are designed to achieve particular goals and objectives, some of those being no less than remarkable, these inventions have limitations which prevent their use in new ways now possible. These prior art inventions are not used and cannot be used to realize the advantages and objectives of the present invention.

Summary of the Invention It is now first presented here, an invention of a vision system computer modeling apparatus including devices for creating a computer model of a real scene. It is a primary function of this system to provide means and method for computer modeling. It is a contrast to prior methods and devices that present systems do not interact with real scenes with respect to perspective and spatial relationships as measured in real-time. A fundamental difference between the computer modeling apparatus of the instant invention and those of the art can be found when considering its interaction with the scene being modeled.

An electronic imaging system combined with a computer aided design system and in communication with position, attitude, and optionally range measuring devices forms a vision system computer modeling apparatus. This vision system computer modeling apparatus is operable for addressing a real scene from some point-of-view, or perspective. The model presented at a display as a two-dimensional figure has associated therewith a perspective which relates to a perspective of a three-dimensional real scene. The model as presented to a user is continuously responsive to changes in perspective. When the imaging system is moved such that the perspective of the scene changes, the perspective of the model similarly and correspondingly changes as well. This is accomplished via measurements of position and attitude of the apparatus with respect to the scene being addressed.

Tools and principles of computer aided design can be combined with real time continuous imaging systems to result in a sophisticated modeling instrument. As it is readily possible to measure spatial information including position, attitude and range of the imaging system with respect to scenes being modeled, these data can be used to better create advanced models with great accuracy with respect to the real world. For illustration by example, the combination of a computer running CAD software; a camcorder electronic camera; and position and attitude measuring apparatus

Using such a system, a designer may view a scene in real time and construct a model of the scene as it is viewed. Computer graphical objects can be created and associated with attributes of the real scene. In particular, the graphic objects can be indexed to the real position and orientation of objects they represent. As the camera is moved, so can the appearance of the graphical objects in the model in a fashion which corresponds to the way real objects would be imaged from the new camera position. By combining computer aided design with real time electronic cameras and position, attitude and range measuring systems, a novel and powerful modeling system is realized.

The invention thus stands in contrast to methods and devices known previously. The invention is a computer modeling apparatus with live real-time images and position and attitude measurement devices which relate the computer model to the scene being modeled in proper perspective for all relative positions of the apparatus with respect to the scene. Presently CAD systems are typically computer workstations with the ability to scan photographs and other graphics which may relate to a scene being modeled. It is a primary object of the invention to provide a vision system computer modeling apparatus. It is additionally an object of the invention to provide a computer aided design apparatus whose models are responsive to the position and attitude of the apparatus with respect to the scene being modeled

It is an object of the invention to provide an apparatus comprising the combination of a computer aided design system, an electronic camera, position and attitude measuring apparatus and optionally, ranging apparatus

A better understanding can be had with reference to the detailed description of Preferred Embodiments and with reference to the appended drawings. These embodiments represent particular ways to realize the invention and are not inclusive of all ways possible. Therefore, there may exist embodiments that do not deviate from the spirit and scope of this disclosure as set forth by the claims, but do not appear here as specific examples It will be appreciated that a great plurality of alternate versions are possible.

Brief Description of the Drawings

These and other features, aspects, and advantages of the present invention will become better understood with regard to the following description, appended claims and drawings where:

Figure 1 is line drawing of objects having a certain spatial relationship; Figure 2 is a different view of the same scene,

Figure 3 shows a close-up and normal view at once,

Figure 4 shows a resizing operation,

Figure 5 shows a scene space geometry,

Figure 6 shows a top view of that same geometry, Figures 7 and 8 illustrate line elements generated from various positions,

Figure 9, 10, 11, and 12 show the side of a building being modeled,

Figure 13 is a flow diagram for a modeling method;

Figure 14 is a system block diagram.

Preferred Embodiments of the Invention

In accordance with each of the preferred embodiments of the invention, there is provided an apparatus for creating a computer model of a scene where the model is responsive to the position and attitude of the apparatus. It will be appreciated that the apparatus of one preferred embodiment may be different than the apparatus of another embodiment. Many alternatives and versions may exist for various components of these systems. Adoption of certain variations can be made without deviation from the true spirit of the invention. The scope of the following disclosure should not be limited to the example elements chosen here for illustration. The scope of the invention is set forth by the appended claims. The following examples are considered by the inventors to be the best of many possible modes of executing the invention.

To realize an apparatus of the invention, one may combine: 1) a camera; 2) a computer, and 3) position and attitude measuring devices. Real-time electronic digital video is available by way of common CCD type electronic cameras. Hand held camcorders can be battery powered and highly portable. They allow a user to view some real scene at a display device, for example a liquid crystal display device. In addition, they are particularly well suited for the present subject as the digital signal they produce is compatible with computer processing operations.

Computers use similar liquid crystal displays for user interface. The images computers generate therefore are compatible with presentation on those devices. Therefore the combination of electronic cameras and computers is facilitated by the common display format. Very advanced design software is now available having graphical modeling capabilities. CAD methodology permits graphical objects to be manipulated in a great plurality of ways so that they may represent real objects in a computer model. The present invention therefore can most simply be described as the combination of a real-time electronic camera with a computer running CAD type modeling software. A user may address some real scene by pointing the camera thereat from a fixed position. Where a computer is connected to the camera and is running CAD software, a user may create a model of the real scene as it is viewed.

A careful reader will realize that models created with such a system would necessarily be specific to a certain perspective. For this reason, the system computer is preferably arranged to be in communication with position and attitude determining devices. If the device is moved, then the perspective of the real scene necessarily changes. To account for this change in the model, the computer mathematically determines a new perspective for the objects of the model by applying rotation, scale, translation algorithms which are responsive to the camera's new position such that the perspective of the model continuously updates in accordance with the true perspective of the real scene.

For example, a camera is addressing a scene, Fig 1, of a box (11) five meters in front of the camera and a ball (12) five meters further in the same direction. If the camera is moved toward the scene, Figure 2, the objects in the foreground, the box (21), would appear in the display to increase in size at a greater rate than those in the background, the ball (22). Given sufficient views of the objects in question, the apparatus would have sufficient data to recall a model of a cube (33) and a sphere (34) to match to the box (31) and the ball (32). The recalled models would then be scaled and oriented by the user or the apparatus itself, and placed so that the models (43, 44) coincided with the real objects (41, 42) as is shown in Figure 4

For the computer to properly size the two objects it may be required to re-size "stretch" the objects, either manually or automatically via image recognition for example, to fit the new size of the corresponding real box and ball After such manual re-sizing, the computer could estimate via a triangulation routine the relative positions with respect to the measured camera displacement Additional camera displacements could result in automatic graphic object updates The first re-sizing operation would calibrate the device and further adjustments may be calculated and the re-sizing of the model objects performed without further user input. Similarly, other perspective and depth features could be accounted for Mathematical translations from a real world 3- D scene to a 2-D computer display are already very common Many advanced computer routines have been developed to facilitate perspective representations of 3-D objects on a 2-D display It is important to realize here that the invention does not propose any new perspective computation routines The well developed routines known are sufficient. It is, however, a new combination of primary elements which result in the powerful modeling tool.

An alternate apparatus of the invention may comprise: 1) a camera; 2) a computer; 3) position and attitude measuring devices, and 4) a range finding device. The addition of a ranging capability enhances the ability of the apparatus to create three dimensional models of the scene being addressed. To calculate the position of a point in the scene, the user would range on the point of interest and the unit, given position and attitude data, could calculate the coordinates of the point directly It is important to realize that the invention does not propose a new method of position calculation based on unit position and attitude, and range to object. This method has been in use, in radar directed gunnery for example, for many years, though not for developing of computer graphical models. What is important is that range based position calculation enables the user to quickly develop a model based on nodes and connecting lines and curves, much as in CAD type programs. Also, range based position calculation enables the unit to use advanced modeling methods based on image processing such as edge detection within a defined area and coplanar modeling. For example, a user of the apparatus may choose to use a method of coplanar modeling, Figure 13. A user would range to sufficient points, at least three are required to define a plane but the user could define as many points as are needed to define the area of interest, to define the boundaries of a planar area 1302. These boundaries may or may not be part of the model. The user may also define areas within these edge boundaries that the unit is to ignore. In other words, the user may cut pieces out of the defined area that will not be analyzed by the unit, in effect setting up interior boundaries. This would enable the user to define exactly the area that is to be analyzed by the unit. Once the area has been defined the unit then calculated the equation defining the plane 1303. The unit then detects, using edge detection and other methods known, all lines and curves within the defined area 1304 and displays them to the user 1305. The user then deletes all lines that are not to be modeled 1306. By utilizing the knowledge of the location and orientation of the defined plane in relation to the units position, and the ability of the unit to measure angles from its bore-sight, by pixel counting and knowledge of lens zoom state, the unit can calculate the location of all of the nodes needed to define those lines and curves in the defined plane 1308, 1309. Figures 9-12 illustrate the basic steps involved in relation to modeling a wall of an office building. In Figure 9, the user defines the planar area by ranging to vertexes 1-6. In Figure 10 the user confirms the process by using an interactive graphical user interface. In Figure 12 the unit displays the calculated model to the user. A continuation of the coplanar method would be to use a pair of planar models, created from different positions, using the same boundary limits. If any objects projected from, or were inset into, the plane the two images would appear different. The unit would be able to calculate, through triangulation, the correct coordinate for these nodes. Some situations may require a third view from another location for completeness. Figure 5 shows a general situation in which a user defines a planar area 51 that has a line 52 sticking out of it at right angles. The viewer would go through the planar modeling process from viewing position #1 53 and from viewing position #2 54 to generate the coordinates for node A 55 within the plane and to calculate the coordinates of the endpoint, node B 56, without the user having to range to either point specifically. In this simple situation it would seem easier to just range on the two points in question, nodes A & B, and have done with it. But if the area contained many projections or indentations the user, by using this method, would merely have to define a planar area, and view it from several points of view, to model these projections in three dimensions. Figure 6 gives a plan view of the situation and clearly shows that, if node B 62 is assumed to be in the defined planar area 61 , that the coordinates calculated for the intersection of a line, from the two viewing positions 63,64 to node B 62, and the plane 61 will produce different coordinates, nodes Bl 66 and B2 65. In the case of node A 61 the change in viewing position will not produce a change in coordinates because node A 61 is indeed in the defined plane. These coordinates, as calculated from each viewing position, are shown in Figures 7 and 8. These figures clearly shows that node A 71, 81 is within the plane. Figures 7 and 8 also show, through the discrepancies in the positions of modes Bl 72 and B2 82, that the real position of node B is not within the defined plane. To calculate the actual coordinates of node B is quite simple. Given the knowledge of the location of each viewing position and the knowledge of the bearings, both horizontal and vertical, to the node, based on the pixel counting angular offset method, from each viewing position the unit could calculate the actual position of the node by triangulation. The primary elements of the systems are: a camera to acquire a digital image of a real scene; a computer to run CAD type software and combine graphical objects with the real image, and further to compute perspective relationships between a real 3-D scene and a 2-D perspective of that scene, the computing routines being responsive to position, attitude and ranging measurements; and apparatus operable for measuring position, attitude, range. More specifically: the camera may be a camcorder type, a portable battery operated device for producing a digital video signal, in communication with a computer processor; the computer may be the type known as a personal computer: having a microprocessor for example Pentium or Power PC processors which have been employed to execute code developed for computer modeling, further being operable for combining an image from an electronic camcorder with a computer generated graphical model image, further being in communication with and responsive to apparatus for measuring position, attitude and range of the camera.

The apparatus for measuring position, attitude and range may be as follows. the position measuring apparatus may be a large scale system such as the global position system G P S. or may be a local system arranged with a particular environment in mind such as a city block or single room of a building, each employing some radio location measures, depending upon desired accuracy, the choice of appropriate position measuring system will vary; the attitude measuring apparatus may be an electronic compass, flux gate compass, interferrometer ring gyro or other device which is operable for measuring relative pointing direction of the camera imaging axis; and the ranging apparatus may be an ultrasonic or radar device which measures relative position of object by timing a test pulse which is reflected therefrom. Some of these devices are common to simple photographic cameras in auto-focus systems.

Other types of ranging, such as laser or infrared rangefinders, may be equally effective as long as it is possible to determine the distance from the camera to a particular object and relay that measurement to the computer.

In a special version of the invention, the camera is omitted. A viewing path is defined by a reticule system which provides reference or "pointing" axis for the devices. In common CAD methods, "pointing" and "clicking" may be done via a mouse peripheral device. An icon, for example an arrow, designates a position on the model field as displayed on the screen. Objects of the model are manipulated by pointing and clicking, clicking and dropping, etc. The model may relate to some real scene, but is not generally responsive thereto.

In a particular version of the invention, pointing, clicking and dragging operations may be implemented by manipulation of a pointing axis into a real scene. Clicking is accomplished by a simple switch on the device sensitive to tactile manipulation. Again, the system's ability to know its position and pointing attitude as well as certain range data allows it to directly interface with the real world. The model is sensitive to and responsive to the position and attitude of device in the 3-D scene. The displayed model, therefore, corresponds to how the scene would look from any perspective.

Although the present invention has been described in considerable detail with clear and concise language and with reference to certain preferred versions thereof including the best modes anticipated by the inventor, other versions are possible. Therefore, the spirit and scope of the appended claims should not be limited by the description of the preferred versions contained therein.

Claims

What is claimed is:

1.) An apparatus for creating a computer model at a display comprised of imagery from a real-scene and imagery from the model, the apparatus being comprised of: a.) an electronic camera in electrical communication with a; b.) computer operable for running CAD software, acquiring images from said electronic camera, receiving physical measurement information, computing perspective adjustments, combining imagery from said electronic camera with imagery from said CAD software, displaying combined imagery at a display; and c.) position, attitude and range measurement apparatus in electrical communication with said computer.