Font Size: a A A

Research On Airborne LiDAR Points Cloud Data Building Reconstruction Technology

Posted on:2014-09-18Degree:MasterType:Thesis
Country:ChinaCandidate:Y F ChenFull Text:PDF
GTID:2268330401476817Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
As a new measurement system that can quickly obtain3D information, intensityinformation and optical images of the survey area, airborne LiDAR (Light Detection andRanging) has been widely used in topographic mapping, coastline monitoring,3D modeling ofurban areas, forest resources survey and other fields. Aiming at building reconstruction based onairborne LiDAR data, this paper researches on the detection of outliers, data filtering, buildings’point cloud acquisition and3D reconstruction. The main work and innovations of the thesis arelisted as follows:1. The composition of airborne LiDAR system and characteristics of point cloud data areintroduced. Based on the analysis of common methods of outlier eliminating, a method on thebasis of scan line and pseudo-grid is proposed. The method is approved to be effective inbuilding data preprocessing by experiment.2. On the basis of analysing existing morphological filtering algorithms, for the need ofsetting the parameters of window size and threshold value manually, the adaptive strategy of twothresholds is proposed and an adaptive mathematical morphology algorithm is designed. Themethod is approved to be self-adaptive to extract non-terrestrial point through different area dataexperimentation.3. A method of LiDAR buildings’ point cloud extraction based on spatial clustering andtopological analysis is proposed. Combined with seed-fill and boundary following algorithm,using elevation image created by point cloud data, the separation of buildings’ point cloud dataand vegetation point cloud data, and the extraction of building contour are achieved.4. An airborne LiDAR point cloud building reconstruction algorithm based on the detectionof critical points is proposed. The point cloud of all facets are achieved through clustering ofRANSAC algorithm, the key contour points of all patch are achieved by using Alpha Shapealgorithm,and the construction of the building model is completed. Experiment shows thismethod can automatically rebuilding flat roof,gable roof, hip roof and other roofs with goodeffection.
Keywords/Search Tags:LiDAR, Point Cloud Data, Outlier Eliminating, Adaptive Filtering, SpatialClustering, Building Reconstruction
PDF Full Text Request
Related items