Font Size: a A A

Study On 3D Laser Point Cloud Data Classification And Preprocessing Method For Building Surface

Posted on:2019-05-28Degree:MasterType:Thesis
Country:ChinaCandidate:G C YuFull Text:PDF
GTID:2310330542490471Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
The ground-based 3D laser scanning is a new technology that can quickly obtain high-precision spatial information.Compared with the traditional measurement methods,it has the advantages of fastness,initiative,non-contact measurement,digitalization,automation and high data sampling rate.The laser point cloud data is discrete and irregular in space.Compared with traditional measurement data,the processing and classification of point cloud data is complex.How to extract the characteristic information of scanned buildings from tens of millions of point cloud data is the key problem of 3D model reconstruction.Aiming at the feature extraction of point cloud data,we carry out the research on the processing and classification of surface laser point cloud data,and extract the characteristic points of point cloud data after classification processing to achieve the purpose of building 3D model of buildings.In this paper,we take the point cloud data of buildings as the research object,classify them according to the RGB value and echo intensity of point cloud data,and then extract the feature data to provide the basis for building the 3D building model.Through the statistical analysis of the RGB values and echo intensities of point cloud data,the RGB value distribution and echo intensity correction models of different types of point clouds are explored.Based on the classification criteria,a point cloud classification model is constructed to classify the point cloud data,And extract the characteristic line of the building;complete the three-dimensional model reconstruction based on the characteristic line.The research contents include:(1)Through the statistical analysis of the RGB values of different types of scan objects to determine the RGB value range of different types of scan objects,a certain percentage of point clouds can be extracted,and the classification of point cloud data has Applicability;(2)Correlation analysis can determine the factors that affect the intensity of point cloud echo,including: scanning distance,scanning angle and scanning object,and summed up the echo intensity correction model,the point cloud echo intensity correction to the same scanning distance And the scanning angle under the conditions of intensity values,the use of the classification of intensity values can further improve the accuracy of classification and extraction of different reflection characteristics of the scanned object;(3)The use of classified point cloud data to extract the characteristics of the scanned object information and achieve Three-dimensional model reconstruction of a scanned building.This paper presents a method of point cloud data processing and classification,through the classification model to extract the corresponding percentage of different types of point cloud,according to the classification point cloud to extract the characteristics of the scanned object to achieve the reconstruction of three-dimensional model of the scanned object,the ground laser point cloud data The treatment and application have certain theoretical significance and application value.
Keywords/Search Tags:Laser point cloud, Classification of point cloud, Feature extraction, Building model reconstruction
PDF Full Text Request
Related items