A stereo matching method based on kernel density estimation

Stereo image matching is a very important research topic in stereo vision. A new stereo matching method based on kernel density estimation is proposed and tested in this paper. The difference space is selected as the matching feature space. Kernel density estimation of the difference samples is used as the similarity measure. A spatially-smoothed kernel is introduced in order to decrease the influences of perspective distortion. Based on the disparity consistency constraint, a 3D match candidates space model is defined in order to do the filtration. Postprocessings on the disparity map make farther increases to the accuracy of the matching results. Experiments confirm that the proposed method is feasible and effective.