基于LiDAR点云的建筑物边界提取及规则化

Extraction and Regularization of Building Boundaries Based on LiDAR Point Clouds

在无需定义主方向及建筑物形状的前提下,提出了一种适合于各种类型建筑的边界提取算法。首先利用凸包算法进行建筑物边界点的提取及排序,接着利用分组的Douglas-Peucker(D-P)算法进行边界特征点的提取,最后实现建筑物边界弧段部分的拟合。

Without defining the dominant orientation and the shape of the building,a new building boundaries extraction algorithm from LiDAR point clouds was proposed in this paper.The first step was obtaining and sorting the boundary points of building.In this step,the convex hull algorithm was used.The second step was using the grouped Douglas-Peucker(D-P) algorithm to extract the characteristic points from the boundary points.Finally,curved building boundaries were fitted by adjacent straight line segments.