Research On Binocular Stereo Matching Algorithm And Key Technology

Computer vision is the current research focus, and binocular stereo vision, a branch of computer vision, has increasingly gained attention and concern. Binocular stereo vision-based applications are increasingly being wildly used such as drone navigation, industrial control, three-dimensional measurement, virtual reality and etc. Stereo matching technology is one of the key technologies of stereo vision, the basic idea is: By the certain cost function and relevant matching restrictions, it is of essence to get the correspondent points from images which are captured from two or more different viewpoints.The paper has realized and focused on the non-parameter characteristic region matching and the invariant based stereo matching technique on the basis of stereo matching algorithm and their theories. Firstly, the paper systematically deals with the binocular stereo match’s general principle, the key technologies, and the calibration principle which needed in the3D reconstruction. The weight matrix has been introduced to matching technique based on region and it has been proved more reliable and efficient; Secondly, the paper proposes the non-parameter characteristic region matching technique and the invariant stereo match algorithms owing to interference in image acquisition and existing calibration error. The experiment indicates that these two methods are strong in robustness and good in anti-jamming capability. Compared with recently popular SIFT method, the invariant stereo match algorithms is high in speed; finally, on the basis of the above works, the paper discusses the3D reconstruction theory in detail and realizes a3D reconstruction system.