Building boundary tracing and regularization from LiDAR point cloud

Accurate and automatic extraction of building boundary from LiDAR data has been an active research topic in photogrammetry, computer vision and remote sensing. Boundary detection from LiDAR data consists of four steps: LiDAR data processing, Initial boundary detection, Boundary tracing, and Boundary regularization. Following these steps, the initial boundary is determined from LiDAR data of building. Techniques such as boundary tracing and regularization are also applied to generate the building boundary. Here a novel approach is introduced for automatically obtaining the high-rise building boundaries from LiDAR data accurately, with individual plane detection. Then, the computed boundaries are projected on the corresponding images to check its correctness and have compared the result with the boundary obtained from the images. The evaluation results show that this approach effectively detects the boundary.