Stereo matching algorithm for autonomous vehicle navigation using integrated matching cost and non-local aggregation

Stereo matching algorithm plays an important role in autonomous vehicle navigation system to ensure accurate three-dimensional (3D) information is provided. The disparity map produced by the stereo directly impacts quality of 3D provided system. However, accuracy a challenging part be solved since it affected surrounding environment such as different brightness, less texture surface, and image pair exposure. In this paper, new framework that used integration census transform (CT) sum absolute difference (SAD) at cost computation step, non-local aggregation second winner take all strategy third median filter final step minimize error. results show improved using proposed methods after some parameter adjustment. Based on standard Middlebury KITTI benchmarking dataset, shows compared with other established methods.