关键词: 相机检校, 平面畸变, 迭代检校, 切比雪夫多项式

Abstract:

Rigorous camera calibration is fundamental to perform accurate measurement and 3D reconstruction. This paper analyses the calibration models for three types of planimetric distortions namely radial distortion, decentering distortion and the orthogonality and affinity (linear) distortion. These distortions are simulated and introduced to ideal images in order to conduct experimental tests. The results indicate that no matter what kind of distortion the images have, the physical model which takes into account all three type of distortion performs best, however, the experimental results also indicate that the existing physical models cannot completely eliminate the image distortion. Therefore, a mathematical calibration model (Chebyshev polynomial) is introduced, and an iterative calibration approach is adapted to attempt to further improve the calibration results. The hybrid iterative calibration approach chooses the calibration model during iterations alternatively between a physical model and a mathematical model. Calibration results from both simulated images and real images (GoPro camera) demonstrate that the use of the proposed hybrid iterative calibration approach can further reduce the distortion.

Key words: camera calibration, plane distortion, iterative calibration, Chebyshev polynomial