Binocular stereoscopic vision-based stereo matching method

Abstract

The invention relates to a binocular stereoscopic vision-based stereo matching method. The method includes the following six stages: Gaussian pyramid construction; cost calculation matching and cost aggregation; cost fusion matching; disparity computation; disparity map repair and void filling; and disparity refinement. Laplacian pyramid transformation is additionally adopted in the cost aggregation stage. An edge-protection-based interpolation algorithm is used in the disparity map repair and hole filling stage. A weighting and bilateral filtering combination-based disparity refinement method is additionally adopted in the disparity refinement stage, so that a high-accuracy disparity map can be obtained. The calculation amount of the method of the invention is moderate; matching results at different scales are fused, improvement is made in the cost aggregation stage and the disparity refinement stage, and therefore, a better disparity map can be obtained; and the method has certain robustness to illumination, external noises and the like.

Description

technical field [0001] The invention relates to a stereo matching method based on binocular stereo vision, which is a method for calculating depth through stereo matching of image pairs based on binocular stereo vision, and belongs to the field of computer vision. Stereo matching is to estimate the scene by finding the matching pixels between two or more different images of the same scene taken from different perspectives, and then converting the 2D displacement difference (also called parallax disparity) between the matching pixels into 3D depth. A 3D model of . Stereo matching is widely used in virtual reality, robot navigation, 3D scene rendering and reconstruction, large-scale mechanical attitude perception and other fields. Background technique [0002] Professor D.Marr from the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology (MIT) proposed a visual computing theory that was successfully applied to binocular matching. Professor D.Marr ob...

AI Technical Summary

Problems solved by technology

They are caused by objective factors such as cameras and other equipment and the environment, and have little to do with the algorithm itself, but they must be disturbed by imbalances such as brightness, hue, and saturation in the processing process

[0005] (2) Parallax discontinuity problem: the disparity values ​​in the contour area where different objects intersect are discontinuous, and the texture information is particularly rich in the discontinuous disparity area, which undoubtedly increases the difficulty of stereo matching, so when performing disparity estimation, it is necessary to Considering this discontinuous characteristic, at the same time, the discontinuous area of ​​disparity is generally close to the occlusion area, so it is particularly prone to false matching when calculating the disparity here.

[0006] (3) Weak texture and repeated texture problem: For these areas, when looking for matching points, there will be more than one pixel corresponding to them. If you want to find the most suitable parallax in these areas, you need to find enough features to distinguish them. , but how to find these features is not an easy task

Images