Analysis of the extraction accuracy of the corner point of the camera using polarization imaging

Abstract

In view of the traditional method of corner extraction, the main idea is to improve the corner extraction algorithm and ignore the imaging process of the calibration image, in this paper, an optimal polarization angle image corner extraction algorithm based on linear polarization feedback is introduced in the process of camera calibration, this method is mainly aimed at the problem of high-light regions which are difficult to detect and eliminate in the multi-position calibration image of space under natural light. The method firstly adopts the linear feedback of Stokes variable through a CCD camera with polarizing plates and obtains the corner image of the optimal angle of the checkerboard lattice in different positions in space, then we use the sub-pixel level detection algorithm and GAUSS's fitting method to precisely locate the corner points in the image and to solve the sub-pixel coordinates of the image corner, at last, the two-dimensional pixel coordinates of the corner points in each checkerboard image are extracted.