To solve the problem of traditional Rank non-parametric matching algorithm

an image matching algorithm based on improved Rank transform is presented.Rank transform is introduced and its limitation is discussed.The mean gray value of all the pixels in the matching window is taken as the gray value of the matching pixel.The gray values of the neighborhood pixels whose relative positions are one unit equal to the matching pixel are kept the same.The gray values of the neighborhood pixels at the other positions are replaced via two-dimensional bilinear interpolation.With the improved rank transform implemented and image matching achieved

the disparity maps are produced.The experiment results indicate that the proposed algorithm can produce dense disparity map.In addition

it does not need iterative calculation and is superior to the tested local stereo matching algorithms in computation efficiency.