A camera self-calibration method based on IOS-PSO

When reconstructing the 3D coordinates of an object in the real world, the necessary step is camera calibration. However, traditional methods cannot achieve online calibration , and active vision methods can not realize specific camera motion. In this paper, self-calibration based on Kruppa equations is adopted for camera calibration and IOS-PSO is presented for the optimization of the cost function. Although it is quick for PSO to converge to the globally optimal solution, heavily numerical oscillation occurs when there are high-dimensional independent variables. For more accurate results, the searching intervals are respectively divided into four small intervals and the searching results follow the defined rules. The experiment results show that the method achieves more accurate camera calibration results and high convergence rate.