Distortion Correction for the Gun Barrel Bore Panoramic Image

摘要

Single Reflector Panoramic Imaging System (SRPIS) has been widely used because of its advantages such as simple structure, fast imaging, integration and miniaturization. It can observe objects around the reflector mirror, which fits for the quality inspection of gun barrel bore. However, its images often suffer from serious distortions in radial and tangential directions. Therefore, to ensure the accuracy of captured images, distortion must be eliminated. In this paper, a distortion correction method is proposed based on the imaging characteristics of SRPIS. Firstly the relationship between the height of a certain point on the gun barrel bore and the radius of image point is derived. Then the correction model is built based on the relationship. Aiming at the captured annular image, a new chessboard corner detection algorithm is proposed. The correction parameters are obtained by using the algorithm according to the labeled image. The real experiment results demonstrate that the correction effects of radial and tangential distortions are satisfactory. The error is controlled at sub-pixel level.