Calibration of Panoramic Cameras Using 3D Scene Information

Abstract

This paper proposes a novel approach for the calibration of a panoramic camera using geometric information available in real scenes. A panoramic camera possesses some flexibility in acquiring different types of panoramas, such as single-center (e.g. as assumed for QTVR), symmetric stereo, concentric or polycentric panoramas. Panoramic camera are based on the use of line sensors rotating around an axis, and are of increasing value for various applications in computer vision, computer graphics or robotics. Previously developed camera calibration methods (for 'standard' camera architectures) are not applicable due to the non-linearity of the panoramic camera, defined by the existence of multiple (nonlinear) optical points and a cylindrical image surface. This article addresses the calibration subject of panoramic cameras for the first time. The paper focuses on the calibration of two dominant parameters that characterize the camera model and provide flexibility in selecting different types of panoramas. It elab rates selected geometric constraints (for increasing numerical stability), experiments with captured image data, an error-sensitivity analysis by simulation, and a discussion why other approaches (designed for 'standard' camera architectures) would fail.