基于多适应值全参数自主变异粒子群的立体相机标定

Stereo Camera Calibration Based on Multiple Fitness Full-Parameter Autonomous Mutation Particle Swarm

基于视觉测量的目标参数获取为仿真系统的性能分析与评估提供可信的数据支持,相机参数标定的精度又决定着测量结果的可靠性。提出采用多适应值全参数自主变异粒子群的相机标定方法,利用传统标定法获取相机的初始内参,通过惯性系数收缩调整、给出基于粒距的全局因子学习调节策略、引入多适应值函数以及设计自主变异律,实现基于粒子群优化标定算法的快速和全局化收敛。实验结果显示本文方法在一定程度上提高了相机的标定精度并可以应用于实际工程之中。

The acquisition of target parameters based on visual measurement provides reliable data support for performance analysis and evaluation of simulation system.The precision of measurement results is determined by the accuracy of camera calibration.A calibration method based on full parameter autonomous mutation particle swarm optimization is proposed.Traditional calibration method is utilized to obtain the initial internal parameters.The fast and global calibration algorithm based on particle swarm optimization is achieved by inertial coefficient contraction adjustment,global factor learning adjustment strategy based on particle distance,multi-adaptation function and the independent variation law.The experimental results show that the proposed method can improve the calibration accuracy to a certain extent and can be used in practical engineering.