Local stereo matching algorithm for low-texture areas

Wenxian Zeng Sr.; Zhaokun Guo Jr.; Qinglin Meng Jr.

doi:10.1117/12.2524193

Translator Disclaimer

6 May 2019 Local stereo matching algorithm for low-texture areas

Wenxian Zeng Sr.; Zhaokun Guo Jr.; Qinglin Meng Jr.

Author Affiliations +

Proceedings Volume 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018); 110690T (2019) https://doi.org/10.1117/12.2524193
Event: Tenth International Conference on Graphic and Image Processing (ICGIP 2018), 2018, Chengdu, China

Abstract

Aiming at the problem that the matching accuracy for the low-texture areas in the stereo matching is so low that the matching failure is likely to occur. This paper proposes a stereo matching algorithm for low-texture areas. Firstly, using the horizontal and vertical gradient of the pixel and the specificity of the pixel itself to perform the calculation of the matching cost. At the same time, introducing the HSV color space to obtain the adaptive support arm length of each pixel in the image, thereby obtaining the aggregate area of each pixel. Then, in the acquired aggregation area, by using the bilateral filtering method to calculate the aggregation cost. Finally, in the disparity computation and refinement stage, Left-Right Consistency (LRC) check is combined with the iterative guided filtering method to reduce the error matching. At the end of the article, by using the standard images on the Middlebury platform to experiment. The experimental results are compared with the traditional experimental results, which proves the effectiveness of the proposed algorithm in the low-texture areas.

Citation Download Citation

Wenxian Zeng Sr., Zhaokun Guo Jr., and Qinglin Meng Jr. "Local stereo matching algorithm for low-texture areas", Proc. SPIE 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018), 110690T (6 May 2019); https://doi.org/10.1117/12.2524193

ACCESS THE FULL ARTICLE