Segment-based stereo matching using edge dynamic programming

Abstract

For the weakness of the pixel-based stereo matching, setting segment as a processing element is more practical. This paper presents a new segment-based stereo matching method using edge dynamic programming (DP). Firstly, from the perspective of human stereo vision, the reference image is segmented by an object-oriented segmentation method, which simulates human perceptual process and fully takes shape, spectral and spatial context into account. Secondly, by employing a refined DP algorithm which is based on the inner edgeline of the segment, the typical "streaking" effect is solved effectively. Finally, the final disparity images are obtained. Consistent with human being's vision, this algorithm combines segmentation and refined DP algorithm speeds up stereo matching process while keeping matching results accurate. The experimental results indicate that the algorithm has high matching precision and fast calculating speed.