Stereo Vision System Based on Coated Multispectral Camera and Its Calibration and Ranging Method

Abstract

The invention generally relates to a multi-spectral binocular vision system, in particular to a coated multi-spectral camera-based stereoscopic vision system and a calibration ranging method thereof.A mosaic coated video multi-spectral detector is used to build the stereoscopic vision system with a multi-spectral binocular ranging function, and a system calibration and matching algorithm for multi-spectral image design. The method specifically comprises the steps of assembly of a hardware system of the multi-spectral detector, multi-spectral image preprocessing, system calibration, feature extraction, semi-global matching optimization, parallax extraction post-processing and the like. The multi-spectral binocular stereoscopic vision imaging system designed by the invention and corresponding ranging algorithm software can be used for ranging of various complex scenes, for example, environmental perception of unmanned vehicles, automatic analysis of video monitoring data, camouflage coating or camouflage cloth target ranging analysis in the military, and the like.

Description

Technical field [0001] The invention relates to a multi-spectral binocular vision system, in particular to a stereo vision system based on a coated multi-spectral camera and a distance measuring method. Background technique [0002] A stereo vision camera, especially a binocular camera, is a widely used device, which can calculate the three-dimensional position of the photographed object relative to the camera based on the principle of binocular parallax. Such equipment has been widely used in production line inspection, vehicle active safety, unmanned driving, three-dimensional scanning and other fields. [0003] In the process of using binocular cameras to achieve binocular ranging, feature extraction and matching of left and right views are important factors that determine the accuracy of ranging. Most of the existing binocular cameras acquire black and white / color high-definition images based on ordinary CCD or CMOS sensors, and acquire parallax information through various for...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem that the existing binocular vision camera is difficult to effectively extract features when the target and background colors are similar, which leads to large calculation errors of the disparity map and low ranging accuracy, the present invention provides a multi-spectral camera based on coating Stereo vision system, calibration method of the system and method of realizing distance measurement by using the system

The traditional stereo vision algorithm cannot be directly used in the calibration of the internal and external parameters of the camera and the calculation of the disparity map

Images