Stereo dense matching method and system

Abstract

The invention discloses a stereo dense matching method, which comprises the following steps of extracting feature points of a left image and a right image and determining support points through carrying out feature point matching on the feature points; constructing a Delaunay triangle according to the support points, wherein the Delaunay triangle comprises a prior probability of disparities of all pixel points in the triangle and a minimum support distance between the pixel points and the support points; utilizing a disparity computation method to compute a disparity conditional probability and a disparity confidence level of the pixel points in the left picture; and utilizing a Bayes principle to compute an optimal posterior disparity according to the Delaunay triangle, the disparity conditional probability and the disparity confidence level. According to the method, a high-precision disparity image is obtained through rapid matching, and the method is especially suitable for a mobile platform or the application field with high real-time requirements; the invention also discloses a stereo dense matching system, which has the beneficial effects.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a stereo matching method and system. Background technique [0002] Stereo dense matching refers to the calculation of selected features, the establishment of correspondence between features, and the correspondence of the same spatial physical point in different images. Stereo matching is an important hotspot and difficulty in computer vision research, and one of the key technologies in many applications such as robotics, medicine, and artificial intelligence. In recent years, with the development of mobile platforms, the accuracy and real-time requirements of stereo matching methods have been continuously improved. The stereo matching algorithm is decomposed into four steps: matching cost calculation, matching cost aggregation, disparity calculation, and disparity refinement. According to different constraints, stereo matching algorithms can be divided into local matchi...

AI Technical Summary

Problems solved by technology

The IDR algorithm structure is very conducive to parallel processing, and it can get high computing efficiency after being optimized under the CUDA architecture; however, the IDR algorithm also has two main shortcomings: the running speed is slow under the non-CUDA architecture, and the memory overhead is very large

The main shortcomings of the ELAS algorithm are as follows: the algorithm structure is not conducive to the realization of parallelization, and some pixels cannot calculate the parallax

Images