An Improved Calibration Method for Stereo Fisheye Vision

The fisheye stereo vision system has been widely studied as primary vision sensors in the field of robot vision because of its wide view of scene. However, currently, serious image distortion has delayed its further development. To remedy this, the paper focuses on the calibration and distortion correction for stereo fisheye system. And then an improved, specialized and systematic calibration method named Stereo Fisheye Calibration Method (SFCM) will be proposed to calibrate the parameters and correct the image distortion of the fisheye stereo vision. The SFCM is introduced by the following two aspects. Firstly, a precise mathematical model based on the law of fisheye lens imaging is established, which assumes that the imaging function can be described by a general polynomial approximation. Secondly, a simple five-point calibration algorithm was implemented to find the possible solutions for relative camera pose between two calibrated views. And then a nonlinear refinement based on the maximum likelihood for internal and external parameters was carried out with Levenberg-Marquarat (LM) algorithm. Finally, using real data captured by equipment, we performed experiments covering all the necessary stages to obtain a high-performance stereo fisheye vision system. The correctness and effectiveness of the proposed method are demonstrated with the statistical analyses of the experimental results.