A camera self-calibration method based on parallel QPSO

In the field of machine vision, camera calibration is a key technology. The self-calibration, one of camera calibration methods, is only based on the images to calculate the camera's intrinsic parameters. It has simple calibration process and strong applicability. Traditional self-calibration algorithm needs to calculate the epipole and fundamental matrix by solving the Kruppa equation, but the uncertainty of the epipole always leads to large error and long operation time. To improve the precision of camera calibration and reduce the time consumption, the parallel quantum particle swarm algorithm (QPSO) is introduced to solve the improved Kruppa equation. It can figure out the camera intrinsic parameters and transform the calculation of epipole into the adaptive value of the cost function. Compared with the ordinary particle swarm optimization algorithm (PSO), QPSO has less parameters, better robustness and faster convergence rate. By using a multi-core computer platform, its parallel processing has also combined with the characteristics of parallel computing which improves the calculation efficiency. Experimental results show that the proposed method is more accurate than ordinary PSO, and the program time consuming is significantly reduced.