Calibration of central catadioptric cameras based on the projection properties between the right and oblique cones.

Abstract

Central catadioptric cameras with a single effective viewpoint can easily transform the original image into a perspective image. In this study, we proposed a calibration algorithm for central catadioptric cameras. A vanishing line was recovered by the images of two sphere images (or line images) and mirror contour based on the properties of right and oblique cones. Subsequently, the intrinsic parameters were solved by the image of circular points and the orthogonal vanishing points. Finally, the distortion coefficient was estimated based on the geometric properties of the right cone. The simulation and real-time experiments conducted proved the effectiveness and feasibility of the proposed algorithm.