论文部分内容阅读
建筑物立面点云分割是车载激光扫描数据特征提取与建模的基础。本文将随机抽样一致性算法(Ran-dom Sampling Consensus)方法引入对点云的分割中,并在判断准则中引入了点云的r半径密度,消除了噪声的影响,同时建立角度和距离两个约束条件对平面分割结果进行优化,提取出了最终的建筑物立面特征平面。
Point cloud segmentation of building facade is the basis of vehicle laser scanning data feature extraction and modeling. This paper introduces the method of Ran-dom sampling diversity (Ran-dom Sampling Consensus) into the segmentation of point clouds, and introduces the r-radius density of point clouds into the judgment criteria, eliminating the influence of noise and establishing two angles Constraints on the plane segmentation results are optimized to extract the final building facade feature plane.