Font Size: a A A

Research On Fine Boundary Feature Extraction Technology Of Ground 3D Laser Point Cloud Buildings

Posted on:2023-10-22Degree:MasterType:Thesis
Country:ChinaCandidate:Z DaiFull Text:PDF
GTID:2530306788460984Subject:Surveying and mapping engineering
Abstract/Summary:PDF Full Text Request
With the continuous construction of digital city,the demand for building three-dimensional model is increasing.How to fine reconstruct the building facade has become a key problem in the modeling process.The recognition and extraction of building surface boundary features based on3 D laser point cloud data can provide data basis for building fine reconstruction.Therefore,building boundary feature extraction based on 3D laser point cloud has important research value and implementation significance,and extracting the boundary features of non wall elements such as windows in building facade is the difficulty of building fine modeling using point cloud data.To solve this problem,a method for extracting the facade boundary features of building laser point cloud based on template matching is proposed.Firstly,the RANSAC algorithm with adaptive threshold is used to roughly segment the building facade.The wall point cloud is extracted by twoway DBSCAN algorithm,and the wall boundary is extracted according to the wall point cloud;Through the spatial geometric characteristics of the building facade,the window point cloud is separated by using the spatial relative position of the wall point cloud and the window point cloud;Secondly,the European clustering monomer and rectangle fitting are carried out on the window point cloud,and the facade windows are classified according to the feature vector constructed by the fitting rectangle;Then the feature convex hull points of the selected template window point cloud are interpolated to obtain the window boundary template point cloud;Finally,the template matching method is used to obtain the complete building facade window point cloud boundary,and the boundary template point cloud of each window is attached to the original type of window,so as to obtain the overall building facade boundary point cloud data.The experimental results show that this method can accurately extract the point cloud boundary of building wall and supplement the window contour point cloud with missing boundary,so as to meet the needs of building fine modeling.Figure19,Table3,Reference67...
Keywords/Search Tags:point cloud, Building facade, Random sampling consistency(RANSAC), Template matching, Boundary extraction
PDF Full Text Request
Related items