文献类型：期刊文献

中文题名：应用改进莱维飞行粒子群算法的相机标定方法

英文题名：Calibration Method with Improved Levy Flight Particle Swarm Algorithm

作者：刘璨[1];李泰星[1];刘焕牢[1];郑重[2]

机构：[1]广东海洋大学机械工程学院,广东湛江524088;[2]广东省湛江市质量计量监督检测所,广东湛江524096

年份：2025

期号：9

起止页码：363

中文期刊名：机械设计与制造

外文期刊名：Machinery Design & Manufacture

收录：北大核心2023、、北大核心

基金：国家自然科学基金项目(52175458)。

语种：中文

中文关键词：相机标定;莱维飞行粒子群算法;初始化种群;自适应变异

外文关键词：Camera Calibration;LFPSO;Initial Population;Adaptive Mutation

中文摘要：针对传统相机标定方法的精度不高、标定结果易陷入局部最优解的问题,提出一种基于改进莱维飞行粒子群算法的相机标定方法。与相机标定模型进行耦合,建立以相机内外参数为优化变量,以焦距、相机光心坐标、扭曲系数及畸变系数等为约束条件,以特征点的理想投影坐标与畸变校正坐标之间的残差为目标的优化函数。通过引入基于混沌搜索和重心反向学习的粒子群初始化机制,基于正弦变化的学习因子调节策略、惯性系数自适应调整以及设计早熟粒子自适应变异机制,实现相机内外参数的快速和全局收敛。实验对比结果验证了提出方法具有更高的标定精度和更强的稳定性。

外文摘要：For the traditional camera calibration method accuracy is not high and the calibration results are easy to fall into the local optimal solution,a novel calibration method based on improved levy flight particle swarm algorithm was proposed.Coupled with the camera calibration model,an optimization function is established with the internal and external camera parameters as the optimization variables;the focal length,camera optical center coordinates,twist coefficient and distortion coefficient as the constraints;and the residual difference between the ideal projection coordinates and the aberration-corrected coordinates of the feature points as the objective.The fast and global convergence is achieved by the particle swarm initialization mechanism based on chaotic search and centroid opposition-based learning,learning factor adjustment strategy based on sinusoidal variation,inertia coefficients adaptive adjustment and an adaptive variation mechanism for premature particles.The Comparison of experimental results verify that the calibration method has higher calibration accuracy and better Consistency.