摘要
提出了一种激光扫描数据和数字图像配准的方法。该方法基于立体像对匹配点与三维扫描点云的最近邻迭代配准。配准中,采用M-估计的选择权迭代在最邻近点搜索算法中逐步消除立体像对误匹配粗差点的影响,得到正确的收敛结果,并基于地面实验证明此方法是可行的和有效的。
A method for registration between laser scanning data and digital images is disscussed. The registration procedure is based on aligning the stereo matched points in overlapping images and laser scanning point cloud. While taking advantage of robust ICP arithmetic, a 3D similarity transform could be calculated, then converged to the right values. A experiment shows that the algorithm is feasible and effective.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2007年第4期290-292,296,共4页
Geomatics and Information Science of Wuhan University
基金
国家863计划资助项目(SQ2006AA12Z108506)
