激光雷达建筑物点云处理过程中往往会混入非建筑物点。本文将点云数据与数字线划图融合,进行粗提取建筑物点。然后借助大平面检索、去噪声、连通区域分析等分割技术完成预处理,选出小平面。这些小平面建筑物点、地面点和植被点等。再运用平整度将地面点和植被点小平面剔除。
During the processing of Lidar points cloud of buildings,non-building points usually interfused. This article fused point cloud data with digital line to extract the building points roughly. And then with the help of segmentation, such as big planes detection, connected component analysis, small planes can be achieved which are building points, ground points and vegetation points. At last this article uses flatness to reject non-building points.