Experiment of Stereo Matching Algorithm Based on Binocular Vision

Abstract

Binocular stereo matching is a hot topic in 3D reconstruction, target detection and target tracking. Through experiments, this paper analyzes and compares several classic binocular stereo matching algorithms, and finally selects SGBM with good comprehensive performance. Moreover, based on this, it gets accurate window size parameters through experiments, which provides a basis for the following 3D modeling.

Review of key technologies for warehouse 3D reconstruction

“…In the binocular stereo vision measurement, when the optical projector is replaced by one camera, the optical projector projects a certain light mode, such as the light plane, cross light plane, network beam, etc., to restrict the position of the scene object in space, unique coordinate value of the point on the scene object can be obtained, thus forming the structural light 3D visual measurement [46] as shown in Figure 3. Surface structured light 3D reconstruction uses a structured light projector (e.g.…”

“…a projector) to project a coded (temporally and spatially coded) pattern or stripe onto the surface of the target to be reconstructed. The camera acquires the scene image containing the projected pattern, decodes the pattern information through image processing to obtain the depth information of the target surface, and then calculates the 3D information of the target to be measured [46][47].…”