Graph Construction and Matching for 3D Object Recognition

In this paper a stereovision system is proposed, based upon the extraction of low level primitives in both images. These primitives and their attributes are represented in a 2D-graph. Correspondences between primitives in both images are found using triangulation. From the list of correspondences, higher level primitives are constructed (tripods and corners), to be used for 3D-object model matching. The required primitive matching, the 3D transformations involved and an error analysis are given. The system is currently being implemented on a robot work station consisting of an IBM SCARA robot and an IBM PC/AT computer.

[1]  Ray A. Jarvis,et al.  A Perspective on Range Finding Techniques for Computer Vision , 1983, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  R. Haralick,et al.  Morphologic edge detection , 1986, IEEE J. Robotics Autom..

[3]  Frans C. A. Groen,et al.  A fast algorithm for cellular logic operations on sequential machines , 1984, Pattern Recognit. Lett..

[4]  Kendall Preston.  Some Notes on Cellular Logic Operators , 1981, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  A. Oosterlinck,et al.  3-D Perception With A Single Binary Coded Illumination Pattern , 1987, Other Conferences.

[6]  K. S. Arun,et al.  Least-Squares Fitting of Two 3-D Point Sets , 1987, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  J. Y. Yang,et al.  Matching Perspective Views of a Polyhedron Using Circuits , 1987, IEEE Transactions on Pattern Analysis and Machine Intelligence.