Research On Camera Calibration Method Of Wide Field Based On Improved Genetic Algorithm

The stereo vision technique is one of rapid development in recent years, and it has been widely applied to industrial, agricultural and defense and so on various fields. The camera calibration technique is one of key technology in stereo vision, and it is the first step in the measurement, and the accuracy of the camera calibration can directly affects the measurement accuracy. For nonlinear optimization precision of the traditional camera calibration approach needs to be further improved, and the traditional medium and small scene method cannot be directly used in wide field calibration problem, in this paper, the nonlinear optimization method of camera calibration and wide field calibration method has carried on the system research, and put forward calibration method of camera with wide filed-of-view based on improved genetic algorithm. In this paper, the main research contents and results are as follows:(1) This paper introduced basic knowledge of camera calibration. First, it introduced the concept and significance of the camera calibration; second, it described in detail the camera imaging theory from three aspects of the four coordinate systems,camera optical system and camera imaging model; finally, the theoretical derivation method for camera calibration based on the plane calibration object is presented. These knowledge for subsequent calibration method of camera with wide filed-of-view lay the foundation.(2) This paper complete nonlinear optimization method for camera calibration parameters based on improved genetic algorithm. For the problem of simple genetic algorithm prone to premature convergence and stagnation phenomena in the process of camera calibration parameters’ nonlinear optimization, a new camera calibration parameter optimization method was proposed. Firstly, the genetic algorithm was improved by the use of nonlinear normalized geometric ranking function and stochastic universal sampling mixed as the selection operator. Then, the improved genetic algorithm was adopted to do the nonlinear optimization for the camera calibration parameters. Finally, in the nonlinear optimization comparison experiment, the mean absolute error of the algorithm is lower than simple genetic algorithm. It is concluded that the improved genetic algorithm can alleviate the premature convergence and stagnation phenomena and the calibration accuracy is improved.(3) The realization of calibration method of camera with wide filed-of-view based on virtual calibration board. For the problem of high-precision calibration board is difficult to produce, inconvenient to transportation and maintenance, a virtual calibration plate idea was proposed. Firstly, the two-dimensional image root in shooting small calibration board in wide field. Secondly, image distortion correction based on the invariance of cross ratio. Again, the small calibration plate of the image virtual translate,to build the equivalent of large calibration plate covering the entire viewing area. Finally,the use of a straight line four-point collinear of the cross-ratio invariance solve virtual calibration points in calibration plate. Real image experiments show that calibration precision of the method is higher the small calibration board experiments, but slightly lower than the big calibration board experiments. Above all, this method has the advantages of simple operation, and meet the measurement requirements.