Image analysis – Applications – 3-d or stereo imaging analysis
Reexamination Certificate
1997-11-26
2001-06-19
Couso, Jose L. (Department: 2621)
Image analysis
Applications
3-d or stereo imaging analysis
C382S285000, C382S318000, C345S420000
Reexamination Certificate
active
06249600
ABSTRACT:
FIELD OF INVENTION
The present invention relates to the efficient generation of three-dimensional representations of an object or scene which can be used for many purposes.
BACKGROUND OF THE INVENTION
Automatic three-dimensional construction of a computer representation of an object or a scene performed in an efficient manner is becoming increasingly important for many different types of applications utilizing computer modeled objects. An object which is scanned and three-dimensionally modeled by a computer can be used for such purposes as rapid prototyping, advanced structural analysis, manufacturing design and process planning for the object or scene being imaged. For example, using three-dimensional modeling a doctor could scan a hip replacement prototype located in his office, store the image of the prototype as a three-dimensional representation in a computer and transmit over communication lines or by satellite the representation to other doctors across the country and anywhere in the world in real time. The doctors receiving the transmission can then view and analyze the representation on a computer and can even produce a solid object with the same physical characteristics based on the received representation using known techniques. The three-dimensional image can also be used for complicated computer analysis of the structure and can be stored for further analysis at a later time. The stored three-dimensional image can be also sent to doctors, hospitals or research facilities at a later time for further opinions.
Conventional techniques which construct three-dimensional images require a great number of scans from many angles around the object in order to achieve a representation of sufficient resolution which is useful for a particular application. Conventional techniques move the relative position of scanner so that every surface of the object is scanned with overlapping views. Conventional techniques require as much as seventy scans to sufficiently scan an object. In most cases, if an occlusion occurs (e.g., a surface such as the inside of a cavity hidden from the sensor), the conventional imaging technique will ignore the portions of the occlusion and treat them as already scanned even if the occlusion may be able to be better imaged from a different angle. This results in a model that is not substantially accurate for the actual object scanned. When the scans required for the imaging process are large in number, the processing time for the model is extended such that real time imaging cannot be accomplished. Therefore, an efficient technique for creating the three-dimensional representation which limits the number of required scans would be very desirable.
If an efficient three-dimensional modeling technique can be accomplished, many new applications can be realized. For example, a three-dimensional facsimile can be sent electronically of a modeled object where the object at the sending end is scanned and three-dimensionally modeled. The imaged object could then be reproduced at the destination using rapid prototyping at a site anywhere in the world or could even be transmitted through space to a space shuttle or space station. A high quality three-dimensional representation which is efficiently created can also be used to manufacture multiple items simultaneously using copies of the same three-dimensional representation in order to increase manufacturing speed.
SUMMARY OF THE INVENTION
The invention is directed to a system and method for generating a three-dimensional representation of an object or scene which minimizes the number of required scans needed to create a high-quality representation at a desired resolution. The method includes the steps of scanning an object from a number of initial positions to create an initial representation which includes occlusion tags which show if an associated surface has been properly imaged or if the associated surface requires further imaging from a different perspective. The representation is in the form of a solid model which ensures that the entire object is captured in the representation. The object to be imaged may contain occlusions or jagged edges which would not be represented in conventional imaging.
After an initial representation is created, a planning stage determines which next view will include the most surfaces tagged “occluded” by using a volume visibility technique. A target area is then acquired and the scanning equipment position is adjusted to set up for a subsequent scan. The results of the recent scan are combined with the current representation to form a more accurate composite representation. The surfaces which were tagged as “occluded” and which had required further scanning now have their tags changed to “imaged.” Thus the scanned surfaces do not require further scanning.
The modeling process can also include the step of shifting the vertices in the mesh model to expand the mesh for the particular scan iteration. This ensures that the entire object is included in the model and outside edges are not cut off due to the resolution of the scanner. The technique eliminates portions of the representation which do not correspond to the actual image by intersecting each model created for each scanning iteration.
REFERENCES:
patent: 4912664 (1990-03-01), Weiss et al.
patent: 4982438 (1991-01-01), Usami et al.
patent: 5027292 (1991-06-01), Rossignac et al.
patent: 5249035 (1993-09-01), Yamanaka
patent: 5375156 (1994-12-01), Kuo-Petravic et al.
patent: 5621648 (1997-04-01), Crump
patent: 5675377 (1997-10-01), Gibas
patent: 5761332 (1998-06-01), Wischmann et al.
patent: 5831621 (1998-11-01), Pito
patent: 5887083 (1999-03-01), Sumi et al.
patent: 5914721 (1999-06-01), Lim
Curless et. al., “A Volumetric Method for Building Complex Models from Range Images,”Proceedings of 23rdInternational Conference on Computer Graphics and Interactive Techniques, New Orleans, LA, Aug. 4, 1996, pp. 303-312.
Laurentini, “Inferring the Shape of the Real Object from the Object Reconstructed by Volume Intersection”,Proc. of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1993, pp. 280-285.
Martin et. al., “Volumetric Descriptions of Objects from Multiple Views,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 5, 1983, pp. 150-158.
Strenstrom et al., “Constructing Object Models from Multiple Images”, Int. Journal of Computer Vision, 9:3, 1992, pp. 185-212.
Turk et. al., “Zippered Polygon Meshes from Range Images”,Proceedings of 21stInternational SIGGRAPH Conference, Orlando, FL, 1994, pp. 451-458.
Wang et. al., “Construction of Surface Representation from 3-D Volumetric Scene Description”Proceedings CVPR '85, San Francisco, CA, 1985, pp. 130-135.
Allen Peter K.
Reed Michael K.
Baker & Botts LLP
Couso Jose L.
Patel Kanji
The Trustees of Columbia University in the City of New York
LandOfFree
System and method for generation of a three-dimensional... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for generation of a three-dimensional..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for generation of a three-dimensional... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2486850