Stereo matching based on support points

A stereo matching method based on stable support points was proposed to improve its computing speed and disparity accuracy. With the method, a canny edge operator was used to detect the edge points of the image and act them as 'supporting points. Then those points were processed by a Delaunay triangulation algorithm to divide the whole image into a series of linked triangular facets. These facets were used as matching elements to compose the basic modules to perform a rude estimation of image disparity. Finally, according to the triangular property of shared vertices, the estimated disparity was refined and the disparity map was obtained. The method was tested by Middlebury stereo pairs on the platform, Experimental results show that the time cost of the method is about 1 s and the matching accuracy is 93% as compared with ground truth map. The proposed method has higher positioning accuracy and lower error mathing rate, and improves both the computing speed and computing accuracy. It forms a steady foundation and good application prospect for robot's path planning system with stereo camera devices.