Distance Measurement and Camera Calibration based on Binocular Vision Technology

Abstract

Development of distance measure technology will always make contributions to the industry and society. Therefore, distance measure based on binocular vision technology has become a hot research topic. This kind of distance measure technology has relative high measure accuracy as well as lower hardware cost. However, due to the adoption of inexpensive binocular cameras, the tangential and radial distortions exist. Thus, the distortion will in turn bring measure error. In other words, measure accuracy of system will decline. This paper, we propose a stereo calibration method for binocular cameras. Camera calibration method is employed to obtain the intrinsic and extrinsic as well as distortion parameters of cameras. Stereo calibration is conducted to correlate the left and right image plane. Rectification method is to make left and right image plane satisfy epipolar constraint. Because after rectification, we can reduce computation of matching correspond feature points in left and right image as well as improve system response speed. The average reprojection error can be controlled within 9 pixels.