Motion stereo vision using a line structural expression of images

A robot vision based on motion stereo vision is presented. The approach uses a method of image registration that uses matching between the two node structures, which are represented by the arcs linked to them in the graphical expression of edge line images. It is necessary, but difficult, to establish matching between feature points in a pair of motion stereo images when a vision system, such as robot vision, extracts three-dimensional positions and shapes of objects. Since the authors´ approach is based on the algebraic expression of the image graph, it is insensitive to scaling, rotation, gray level modification, and small motion between the two images. The matching process makes use of a degree of structural similarity among nodes. The experimental result confirms the reliable matching between a pair of motion stereo images and the applicability of the approach to 3-D robot vision