基于粒子群算法的摄像机内参数优化方法

Optimization of Camera Internal Parameters Based on Particle Swarm Algorithm

针对MATLAB标定工具箱的标定精度与所拍图像数量成正比的问题,即拍摄照片数量越多标定精度越高,提出了一种基于粒子群算法的摄像机内参数优化方法,从而达到拍摄少量图片也可以有较好精度的效果。首先摄像机从不同角度拍摄4张和20张标定板图片,利用MATLAB标定工具箱分别求取它们的内参数。然后根据标定点的实际坐标和反投影坐标建立目标函数,再由粒子群算法对标定箱求取的内参数进行优化。实验结果对比表明:与MATLAB标定工具箱相比,此方法能够在一定程度上提高少量标定板图片的标定精度。

Aiming at the problem that the calibration accuracy of MATLAB calibration toolbox is proportional to the number of images taken, which means the larger the number of photo frames, the higher the calibration accuracy. A method of internal parameter optimization based on particle swarm algorithm is proposed, and the better effects can be achieved with few pictures. First the camera shoots 4 and 20 calibration plate pictures in different angles, and their internal parameters are obtained with the use of MATLAB calibration toolbox. The objective function is established through the calibration point of the actual coordinates and the back projection coordinates, and then the internal parameters obtained by calibration box are optimized by the particle swarm algorithm. The experimental results show that this method can improve the calibration accuracy of a small number of calibration plate pictures to a certain extent compared with the MATLAB calibration toolbox.