基于机载LiDAR数据的建筑物轮廓规则化方法

摘要

在点云数据分类和建筑物轮廓提取的基础上，提出一种基于最小外包矩形的建筑物轮廓规则化方法。以任意两相邻轮廓点构建初始外包矩形，以迭代的方式将内点数最多者作为最终外包矩形，继而以面积比率、距离和均方根作为指标对规则化结果给出评判，最后以实际数据进行实验。实验结果表明，文中提出的方法能够有效对分类后的建筑物轮廓点进行规则化。%Based on the classification of point cloud and extraction of building boundary ,a regularization method for building boundary based on Minimum Bounding Rectangle (MBR) is put forward .First the initial bounding rectangle is constructed by two arbitrary adjacent contour points ,of which the optimal one obtaining the most consistent contour points is determined by an iterative manner ,and then the results are evaluated by two criterions :area ratio and RMS of the distance between contour points and its corresponding rectangle side .Finally a sample data is taken as an example to carry out the experiment .The results show the method proposed can regularize the contour points effectively .