Building edge extraction based on DSM digital surface model and LIDAR point cloud data

Abstract

LIDAR (Light Detect And Ranging) is a new measurement system that USES laser beam scanning to Detect targets And measure distances, And can directly And quickly obtain high-precision digital surface models (DSM) of different areas, such as urban areas. This can make up for the defect of image data in the third dimension. The emergence of this technology provides a new way for the rapid extraction of surface models, especially for buildings in urban areas. In this paper, based on the airborne LIDAR point cloud data, the irregular discrete LIDAR point cloud data grid resampling rules, get digital surface model (DSM), and then carried out in accordance with the elevation gray quantitative generated DSM depth image, using edge detection operator to extract buildings, using orthogonal constraint to the adjacent segment Angle adjustment, to obtain the edge information in line with the actual situation of buildings. Finally, the accuracy and precision of segmentation results are analysed.