An improved algorithm for disparity estimation of SGM stereo matching based on edge detection

Abstract

In the fields of depth measurement and computer vision, generating sufficiently accurate disparity maps through various stereo matching algorithms is an important foundation. The SGM semi-global matching algorithm is one of the best stereo matching algorithms, But its filling accuracy is poor in the invalid parallax area. Therefore, on the original basis, an improved algorithm of SGM stereo matching disparity filling based on edge detection is proposed. The experiment uses a standard data set to verify the algorithm. The results show that compared with the original SGM stereo matching algorithm, the paper algorithm can achieve better disparity estimation results in the invalid disparity area.