基于DSM深度影像的机载LiDAR建筑物边界提取

Building Boundary Extraction by Airborne LiDAR Based on DSM Deep Image

该文探讨了基于DSM深度影像的机载Li DAR建筑物边界提取。首先利用归一化的DSM数据点生成DSM深度影像,然后对影像进行相关处理,包括中值滤波、形态学处理,以剔除噪声、植被及较小地物等信息,采用阈值分割与二值化,得到明显的建筑物区域,最后使用边界检测的方法将建筑物初始边界提取出来,并检测其角点,利用最小二乘方法进行拟合,得到规则化的建筑物边界。

This paper discusses the building boundary extraction by Airborne LiDAR based on DSM deep image. First using the normalized DSM data point to create the DSM deep image. Then the image is processed,including median filtering and morphological processing to eliminate the noice, vegetation, small surface features and other information. Using the threshold segmentation and binaryzation, the building area is obtained observious. At last, using the boundary detecting method, the initial boundary of building is extracted, and its corner is detected. Using the least square method, the regular boundary of building is obtained.