摘要
针对传统迭代最近点算法依赖于初始位置以及传统遗传优化算法搜索效率低的缺陷。提出了一种以遗传算法为基础融合模拟退火算法的GA-SA混合优化算法实现点云粗配准。其中简化了RMSE(作为适应度函数;再基于点到平面ICP实现精配准。实验结果表明:相较传统优化算法,本文GA-SA配准算法精度显著提高,配准效率提升20%以上,保证了ICP算法的精确配准,并对噪声和点云缺失具有一定鲁棒性。
Aiming at the defects that traditional ICP relies on the initial position and the traditional optimization algorithm id inefficient.A GA-SA hybrid optimization algorithm based on GA combined with SA is proposed to achieve rough registration of point clouds.Meanwhile RMSE is simplified as Fitness function.And point-to-plane ICP achieves fine registration.The experimental results show that compared with the traditional optimization algorithm the accuracy of the GA-SA registration algorithm in this paper significantly improved,and the registration efficiency is increased by more than 20%,which guarantees the accurate registration of the ICP algorithm and is robust to noise and point cloud missing.
作者
蓝秦隆
邹进贵
杨丁亮
LAN Qinlong;ZOU Jingui;YANG Dingliang(School of Geodesy and Geomatics,Wuhan university,Wuhan 430079,China)
出处
《测绘科学》
CSCD
北大核心
2022年第7期119-125,共7页
Science of Surveying and Mapping
基金
国家自然科学基金项目(41871373)
关键词
点云配准
遗传算法
模拟退火算法
ICP
point cloud registration
genetic algorithm
simulated annealing algorithm
ICP
