Stereoscopic matching method based on double-weight aggregation

Abstract

The invention discloses a stereoscopic matching method based on double-weight aggregation, and relates to the field of stereoscopic matching. The method includes the steps that brightness matching cost and gradient matching cost of reference pixel points in a reference map and a target map are acquired; disparity space is acquired according to the brightness matching cost and the gradient matching cost of the reference pixel points; the minimum value of the disparity space is acquired, and a left initial disparity map with a left map as a reference map and a right initial disparity map with a right map as a reference map are respectively acquired; consistency checking is conducted on the left initial disparity map and the right initial disparity map, and then an accurate left disparity map is acquired; the disparity space is smoothed through a hyperbolic-curve united parabola method, and then a disparity map with higher accuracy and of a sub-pixel level is acquired. The more accurate disparity map can be acquired, robustness of images is enhanced, and the stereoscopic matching method has the advantages of being high in accuracy and expansibility and the like, and has wide application prospect.

Description

technical field [0001] The invention relates to stereo matching in the field of computer stereo vision, and in particular to a stereo matching method based on double weight aggregation, which can be used for three-dimensional reconstruction of images and provide guidance in medical images, media equipment and the like. Background technique [0002] Stereo matching is to obtain the depth information of each object in the scene based on multiple two-dimensional images with different viewing angles in the same scene. This technology is a research hotspot in the field of computer vision and has received extensive attention in recent years. The local stereo matching method uses the local information around the interest point for correlation calculation, which involves less information, and the corresponding calculation complexity is low, and has the advantages of high speed and high efficiency. [0003] The traditional local stereo matching method generally adopts the matching me...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a stereo matching method based on double weight aggregation. The present invention obtains the initial disparity map by improving the matching cost aggregation, and at the same time improves the disparity smoothing strategy in the parallax optimization stage to solve the problem of the traditional local stereo matching method. For the problem of high false matching rate in the matching process, please refer to the following description for details:

Images