Feature Extraction And 3D Modeling Of Building Point Cloud Data Based On Hough Transform

The airborne lidar system is a high-tech that can quickly acquire real 3D data from the ground.Buildings are one of the landmark landmarks of the city.Its identification,extraction and information updating play an important role in the study of urban dynamics,urban construction and three-dimensional data analysis.Extracting building feature information from airborne lidar is an important part of urban planning,digital city,urban construction,etc.How to efficiently extract building information from airborne LiDAR point cloud data has become a research hotspot.In recent years,most researchers have acquired building feature information in point cloud data in a semi-automated manner.However,the building feature information extraction method has a relatively large limitation and a small application range.This paper focuses on the feature information extraction and 3D modeling technology of point cloud data buildings,optimizes the method of extracting building feature information based on Hough transform algorithm and the process of 3D reconstruction of building roof patch,and verifies the paper through many experiments.The feasibility and robustness of the method.The main work and research results of this paper are as follows:(1)Analyze and compare the classical point cloud data filtering algorithm.By analyzing the four filtering algorithms of mathematical shape,slope,plane/surface and cloth simulation,the robustness of the cloth simulation filtering algorithm is better.(2)Optimized the extraction of building outline algorithm based on Hough transform.DSM depth image is generated by grid resampling method,edge detection algorithm is used for edge detection and Gaussian kernel Hough transform linear fitting method is used to extract buildings from LiDAR point cloud data without external auxiliary data.Contour information.The extracted information reflects the physical structure and geometry of the building,laying the foundation for 3D modeling.(3)Optimized the algorithm for extracting roofs of buildings based on Hough transform.The adaptive octree segmentation is performed on the approximate coplanar point cloud sample set.The principal component analysis method is used to analyze the data and calculate the variance of the point cloud distribution.Finally,the Gaussian kernel function is used to vote and perform peak detection.The robustness of the algorithm is verified by error index,plane detection accuracy and calculation time.(4)Using the obtained contour line and roof patch data,the element library element parameters are set to assist in three-dimensional modeling,and texture mapping technology is used to realize texture mapping of building information.