Research on the Effect Optimization of Parallax Map in Robot Vision
The effect of the stereo matching algorithm is very important in robot applications.In order to improve the effect of the stereo matching algorithm.First,use Sobel detection based on first-order difference,Scharr detection and Canny detection to per-form gradient processing on the image to obtain the edge of the image.Secondly,use the edge information of the image as the ini-tial matching cost for stereo matching,which greatly reduces the amount of calculation necessary.Then use the method proposed in this paper to conduct experiments on the images in the Middlebury data set and self-collected images to obtain the mismatch rate and the calculation time of stereo matching.Through the comparison of experimental data,the combination of Sobel detec-tion algorithm and BM(Boyer-Moore)algorithm can get the best disparity map with high speed and accuracy.The experimental results that the proposed method has obvious advantages and can improve the stereo matching effect.