Segment-based Stereo Matching Using Edge Dynamic Programming

Most current stereo matching method is based on pixel processing,but object-oriented approach is more conform to the human visual habit.This paper put forward a scale-span multi-scale stereo matching method,achieved by coarse to fine stereo matching.Firstly the reference image is rough segmented by an object-oriented segmentation method,then by employing a refined DP algorithm which is based on the inner edgeline of the segment,disparity are obtained.According to the accuracy demand,determine whether need to do the refined segmentation.Finally,the final disparity images are obtained.The experimental results indicate that the algorithm has high matching precision and fast calculating speed.