基于立体靶标的多摄像机标定方法研究

he accuracy and algorithm stability of camera calibration directly affect the accuracy of measurement results. In this paper, utilizing the characteristic of the anisotropy of the standard sphere ,the stereo target containing the standard sphere was put forward creatively aiming at improve the calibration accuracy of non-parallel multi-cameras system and simplify the calibration process. On the basis of the Zhang Zhengyou's plane template calibration method, the imaging model and the basic principle of camera calibration were introduced. According to the imaging principle of sphere, the extraction method of feature points was introduced in detail. Then the internal and external parameters of the camera can be calibrated by the corresponding algorithm.This method solves the problem of multi-camera calibration partly, to a certain extent, simplify multi-camera calibration steps to improve the calibration accuracy. Finally, the experimental error sources were analyzed, and the theoretical derivation and calculation were carried out. The relevant experiments were done to verify the program and the corresponding conclusions were obtained.