Semi-Global Stereo Matching Algorithm Based on Multi-Scale Information Fusion

Semi-global matching (SGM) has been widely used in binocular vision. In spite of its good efficiency, SGM still difficulties dealing with low-texture regions. this paper, an algorithm based on multi-scale information fusion (MSIF), named SGM-MSIF, is proposed by combining multi-path cost aggregation and cross-scale (CSCA). Firstly, the stereo pairs at different scales are obtained Gaussian pyramid down-sampling. The initial volumes computed census transform color information. Then, introduced into each scale aggregated fused CSCA. Thirdly, disparity map optimized internal left-right consistency check median filter. Finally, experiments conducted Middlebury datasets to evaluate algorithm. Experimental results show that average error rate (EMR) SGM-MSIF reduced 1.96% compared SGM. Compared classical algorithm, EMR 0.92%, while processing efficiency increased 58.7%. terms overall performance, outperforms classic CSCA algorithms. It can achieve high accuracy for vision applications, especially those