A Novel Method for the Distortion Modification of Camera Lens

Lens distortion is one of the main factors affecting camera calibration. To improve the accuracy of calibrating camera, a new approach to modify the lens distortion is proposed. Based on Brown model and pinhole imaging principle, the measurement model of the incident angle is built. The sensibilities of different distortion coefficients in the expression of incident angle are analyzed using Monte Carlo method. According to the analytical results, all the coefficients are united into one vector with different proportion, which greatly simplifies the least squares iterative algorithm that is widely applied to the distortion modification. After analyzing the process of iterative optimization, the experiments of measuring the incident angle are carried out and the distortion coefficients are obtained by iterative optimization. The experiments demonstrate the linearity of the incident angle is enhanced greatly through distortion modification with these coefficients. The repetitiveness of the modification effect is also proved well.