多视点云拼接中的ICP算法优化  

Improved ICP Algorithm in Multi-view Point Clouds Splicing

作　　者：姚晓山[1] 刘健鑫[2] 柯维[2] 

机构地区：[1]空军雷达学院军通系,湖北武汉430019 [2]海军工程大学船舶与动力学院,湖北武汉430033

出　　处：《微电子学与计算机》2012年第8期94-97,共4页Microelectronics & Computer

基　　金：国家自然科学基金资助项目(50979109)

摘　　要：为了提高多视点云拼接算法的稳定性及其拼接的精度和效率,对现有的ICP算法进行了优化,提出了基于最小二乘法的混合定位点云预拼接方法,对两块点云进行了粗略拼接,获取了两块点云中可匹配部分的对应点集,并给出了具体的算法,进而提出了基于总体最小二乘法的点云精确拼接算法,对两块点云进行了精确拼接,实现了现有ICP算法的优化,保证了拼接的全局收敛性和算法的稳定性,减少了算法的迭代次数,提高了点云拼接的精度和效率.How to improve the stability and accuracy of the algorithms for point clouds splicing was studied, and a hybrid positioning point clouds pre-splicing algorithm based on least square method was advanced, and the detailed algorithm was given, and a precise algorithm based on total least squares method was presented, and the optimization of existing ICP algorithms was achieved, and the global convergence and the stability of the algorithm were ensured, and the number of iterations was reduced, and the accuracy and efficiency of the algorithm was improved.

关 键 词：多视点云 点云拼接 定位点 最小二乘法 总体最小二乘法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]