Design And Development Of 3D Point Cloud Data Processing System

Three-dimensional laser scanning technology is an emerging technology that has developed rapidly in recent years and can directly acquire data such as threedimensional coordinates,color values,and laser intensity.The massive discrete data that obtained,we call it the point cloud.Three-dimensional point cloud data has been widely used in surveying and mapping,medical,artificial intelligence,leisure and entertainment,such as forest resource exploration,digital city construction,auxiliary diagnosis,robot automatic path finding,VR games,and coastal defense security.However,raw point cloud data,due to the accuracy of the instrument itself,mechanical vibration and some accidental environmental factors,will cause many noises and holes,affecting subsequent processing and application.Therefore,it is of great value and significance to independently develop a three-dimensional point cloud data processing system.This paper first studies the basic knowledge of point cloud data processing,including two point cloud data formats,organizational structure,nearest neighbor search algorithm,normal and curvature.Then based on these foundations and characteristics,the related point cloud data processing algorithms are studied.In the aspect of filtering and denoising,various filtering algorithms such as through filtering,radius filtering,SOR filtering and VG down sampling are studied.In the aspect of point cloud segmentation,the random sampling consistency algorithm is studied,and the planar model and the cylindrical model segmentation are realized.At the same time,the European clustering segmentation and region growing segmentation algorithms are studied.In the aspect of point cloud classification,the progressive morphology filtering algorithm is studied and successfully used to classify ground points and non-ground points.This paper proposes a classification strategy based on echo information and elevation texture features,which is successfully used for the extraction of building points and vegetation points.In the surface reconstruction aspect,the moving least squares algorithm is studied and the smoothing of the point cloud is realized.The greedy projection triangulation algorithm is studied and the construction of the surface model is completed.Finally,according to the actual needs,the overall framework and processing flow of the point cloud data processing system is designed.Based on the open source libLAS and PCL,using C++ object-oriented thinking,combined with QT graphical user interface framework,designed a point cloud data processing system for LAS and PCD format point cloud,completed the point cloud The main operations and functions in data processing.In this paper,the relevant point cloud data is selected for various algorithm experiments,and the ideal experimental results are obtained,which verifies the fluency and stability of the system.