Automated extraction of building footprints from mobileLIDAR point clouds
车载激光扫描点云中建筑物边界的快速提取

This paper presents a novel method for automated extraction of building footprints from mobile LiDAR point clouds.We first generate the georeferenced feature image of mobile LiDAR point clouds using an interpolation method, and adopt imagesegmentation and contour extraction and tracing to extract building boundaries in the geo-referenced feature image as the coarselevel of building footprints in Two-dimensional imagery space. Then, the coarse level of building footprints is further refined byapplying planar segmentation on the extracted point clouds in the building boundaries. Finally, the triangulated irregular network (TIN)is used to achieve the fine level of building footprints. Dataset of residential areas captured by Optech's LYNX mobile mappingsystem was tested to check the validities of the proposed method. Experimental results show that the proposed method provides apromising and valid solution for automatically extracting building footprints from mobile LIDAR point clouds.