Plain City Airborne LiDAR Data Filtering And Building Roof Extraction

In recent years, with the development of "Digital Earth" and "Digital City" technology, quickly and accurately obtaining three-dimension spatial data of buildings and terrain information is becoming more and more important. Airborne LiDAR technology is gradually replacing the traditional photogrammetry and becoming an important means of data acquisition for its high efficiency, high accuracy and low cost. As an important part of the "Digital City", three-dimensional building model reconstruction has always been a hotspot.This paper mainly presented the airborne LiDAR data filtering and building feature extraction algorithms. First of all, use the region growth algorithm to filtering point cloud data, Secondly, the orthogonal polynomial filtering algorithm was employed to filter the point clouds, compare the efficiency of these two algorithms, analysis their characteristics. Then take the characteristics of the surveyed area into consideration to determine the optimal filtering threshold based on the compared experiment results. Next, this paper presented a Random Sample Consensus (RANSAC) based building feature extraction algorithm to extract the roof planes and outlines information of the buildings from the non-ground cloud point data.In the end, based on proposed algorithms, the filtering and building feature extraction program was implemented in Qt environment and the three-dimensional graphics library OpenGL was employed to realize visualization. In experiments, the LiDAR data of Changchun, Jilin Province was used to verify the efficiency and accuracy of proposed algorithms.This paper focuses on threshold parameters optimization in the airborne LiDAR data filtering algorithm and proposed the RANSAC model based building feature extraction algorithm. Based on the algorithm research, an airborne LiDAR data processing program was developed. The experiment results proved that this research has a certain application value for the three-dimensional building reconstruction and digital city modeling.