CCD camera optimization calibration technique using genetic algorithm

In order to meet the requirement for high accuracy camera calibration technique in the field of computer vision measurement, a new two-stage calibration method base on linear solution estimation and genetic algorithm optimization was presented in this paper. Firstly, the intrinsic and extrinsic parameters based on ideal pinhole model were estimated by linear algorithm, and that the estimated results were used as the optimization initial values. Then all parameters including the distortions coefficients were optimized by genetic algorithm to get extract solutions. The rotation angels around axes were used as the genes of chromosome vector, the rotation angels were optimized directly by genetic algorithm, consequently the orthonormal constraints of rotation matrix were held. The results of comparison LEDs locations in the calibration template calculated inversely using obtained calibration parameters with LEDs locations calibrated precisely by 3D CMM indicated that the deviations were less than 0.05 mm, so the method could be satisfied with the calibration requirement of camera applied to vision measurement.