ABSTRACT A hybrid particle swarm optimization algorithm for determining extrinsic parameters of the camera, with the intrinsic parameters known, is presented in this paper. An adaptive inertia weighting strategy is designed, and improved crossover and mutation operation are performed. Taking the result of the particle swarm optimization algorithm as the initial value, the extrinsic parameters of the camera are optimized by Gauss–Newton method. The simulation results demonstrate that, when the standard deviation of image noise is 0.1 pixels, the mean relative error of rotation angle is less than 0.04844% and the mean relative error of translation vector is less than 0.05583%. The experimental results show that the mean reprojection error of the proposed algorithm is no more than 0.06045 pixels. The algorithm can achieve high precision and robustness in practical applications. Besides, the algorithm is computational efficient.

中文翻译：

---

改进粒子群优化算法在求解相机外​​参数中的应用

摘要 本文提出了一种混合粒子群优化算法，用于确定相机的外在参数，其内在参数是已知的。设计了一种自适应惯性加权策略，并进行了改进的交叉和变异操作。以粒子群优化算法的结果为初始值，采用高斯-牛顿法对相机的外参进行优化。仿真结果表明，当图像噪声标准差为0.1个像素时，旋转角度的平均相对误差小于0.04844%，平移向量的平均相对误差小于0.05583%。实验结果表明，该算法的平均重投影误差不超过0.06045个像素。该算法在实际应用中可以达到较高的精度和鲁棒性。此外，该算法计算效率高。