A Multi-Measure Semi-global Dense Matching Method

Abstract

The invention discloses a multi-measure semi-global dense matching algorithm. Firstly, matching cost is calculated on the top layer of a pyramid by utilizing a Cenus similarity measure. A disparity map is generated and optimized. Then a disparity map result is transmitted downwards step by step according to a pyramid strategy to serve as disparity map initial values calculated by the mutual information matching cost of the next-level pyramid, and finally a dense matching result of sub-pixel-level precision is obtained. According to the method, for defects of a classic semi-global matching algorithm in the aspects of efficiency, robustness and the like, improvement and expansion are carried out in the aspects of penalty coefficient, similarity measure selection, parallax range adjustment and the like. The method has the characteristics of robustness, reliability, high efficiency, fine matching parallax image and edge protection. Meanwhile, robustness of matching is improved by conducting radiation processing on the image data set, the calculation memory in the matching process is reduced through dynamic adjustment of the self-adaptive disparity range, and the matching efficiency isimproved.

Description

technical field [0001] The invention belongs to the field of image matching, and relates to an improved semi-global pixel-by-pixel image dense matching method, in particular to a multi-measure semi-global dense matching algorithm. Background technique [0002] Image dense matching is a method of obtaining the three-dimensional dense point cloud or pixel-by-pixel disparity map (or depth map) of the photographic object by matching multi-view images with known orientation parameters (including image internal orientation elements and external orientation elements). It is a crucial step in image-based 3D surface reconstruction, a key technology in the automatic generation of digital surface models (Digital Surface Model, DSM) and digital elevation models (Digital Elevation Model, DEM) in photogrammetry, and also a An indispensable part of 3D modeling in the field. [0003] Generally speaking, recovering the three-dimensional shape of a space object from an image is similar to th...

AI Technical Summary

Problems solved by technology

[0008] The present invention mainly solves the problems of the classic semi-global matching algorithm in terms of efficiency and robustness, and proposes a multi-measure semi-global dense matching that uses two similarity measures, Census and mutual information, and uses a pyramid image strategy for matching. method, which can effectively solve the problem of disparity initial value and mismatching that require a priori priori calculation of mutual information. At the same time, this method has the characteristics of small calculation memory usage, high processing efficiency and high matching accuracy, and improves the robustness of matching through radiation processing. sex

Images