A Stereo Matching Method Based on Adaptive Gaussian Weighting

Abstract

The invention relates to a stereo matching method based on self-adaptation Gaussian weighting. The stereo matching method based on the self-adaptation Gaussian weighting includes the following steps: respectively selecting two center points along the same horizontal direction in a target view and a reference view and so as to build a correlation window; calculating a gray difference value of the central points of the target view and the reference view, which are located in the correlation window; calculating color weight of a preset number of pixels around the center points of the target view and the reference view and the center points; respectively calculating Gaussian weight of the preset number of the pixels around the center points of the target view and the reference view and the center points; calculating aggregation cost by using the above steps in combination mode, and calculating the aggregation cost in the maximum range by repeating the above steps after moving the center point of the reference view in the maximum range of the reference view; using a WTA method to perform parallax error optimization. The stereo matching method based on the self-adaptation Gaussian weighting is a stereo matching method based on Gaussian distribution improvement, clearly reserves edge detail information, solves the problems caused by confirming the size of the correlation window from experience, and obtains an accurate parallax error map.

Description

technical field [0001] The invention relates to obtaining a disparity map through stereo matching, in particular to a stereo matching method based on adaptive Gaussian weighting. Background technique [0002] The basic principle of stereo matching technology is to observe the same object from multiple viewpoints to obtain images under different perspectives; calculate the corresponding disparity map through algorithm processing to obtain the three-dimensional information of the object. In recent years, stereo matching technology has become the focus of research in the field of vision, and has been widely used in aspects such as: drawing virtual viewpoints, reconstructing 3D images, and virtual realization. [0003] Stereo matching methods are generally divided into two categories: local stereo matching methods and global stereo matching methods. The global stereo matching method has high precision, but the processing speed is very slow. After years of development, the local...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a stereo matching method based on adaptive Gaussian weighting, to solve the problems of traditional adaptive weighting method that the edge details are not prominent, and the window size is set by experience.

Images