Abstract: Polarization detection is becoming increasingly widespread in many related fields, the calibration and correction of the polarization detection system is a premise to extract exploration targets accurately. In this paper, the principle of camera calibration algorithm based on 2D planar is introduced, and the linear model as well as nonlinear distortion model of the camera is set up. Then we get accurate internal and external camera parameters by linear and nonlinear algorithm. We obtain the projection coordinates before distortion by solving binary higher-order nonlinear equations based on the lens distortion model. To achieve effective correction of distorted images, we set up the relationship between amount of distortion of each pixel and distance to the distortion center. In addition, we carry on a test to the distortion image before and after correction, the result shows that the actual amount of distortion is consistent with theory of distortion model; the polarization image after correction can effectively raise the accuracy of object extraction, and the correction of polarization image has the most positive impact for the accuracy of object extraction in complex background.