Sparse disparity acquisition method based on target edge features and gray similarity

Abstract

The invention relates to a sparse parallax obtaining method based on target edge features and gray scale similarity and belongs to the technical field of stereoscopic vision. According to the method, the image edge features are extracted by means of the Canny image edge detection method, a strong edge feature point of a target in an image is chosen as a point to be matched, feature similarity and the gray scale similarity of the image are combined, a similarity detecting method of fusion of gray scale and gray scale Census change is adopted, error matching points are eliminated by means of cross detection, and accuracy of matching is further improved. The sparse parallax obtaining method based on the target edge features and the gray scale similarity provides necessary sparse parallax information for recovery of the three-dimensional position and attitude information of an object, has the advantages of being high in accuracy, high in matching speed and stable in matching, and can effectively provides the necessary sparse parallax information for binocular vision system measurement especially active measurement of the attitude information of the object.

Description

technical field [0001] The invention relates to a sparse parallax acquisition method based on target edge features and gray-scale similarity, and belongs to the technical field of stereo vision. Background technique [0002] Stereo matching is a key step in recovering scene 3D information in binocular stereo vision. The three-dimensional space scene forms a two-dimensional plane image through the perspective projection of the camera, and a large amount of information is lost in the process of projecting from the high-dimensional space to the low-dimensional space, resulting in the ambiguity of restoring the three-dimensional scene information from the plane image. Affected by factors such as camera parameters, uneven illumination of the scene, and reflectivity of object surfaces in different directions, corresponding points in the two images cannot have exactly the same grayscale and color characteristics, which brings additional difficulties to restore 3D scene information ...

AI Technical Summary

Problems solved by technology

The three-dimensional space scene forms a two-dimensional plane image through the perspective projection of the camera, and a large amount of information is lost in the process of projection from the high-dimensional space to the low-dimensional space, resulting in the ambiguity of restoring the three-dimensional scene information from the plane image

[0005] The binocular stereo matching methods for the purpose of restoring the attitude information of the target object in the scene all use the sparse matching method. The common problem of the existing sparse matching algorithms is that the correct matching rate is not high.

Images