Design And Development Of Laser Point Cloud Data Processing System

In recent years,with the development of science and technology and the need of society,3D laser scanning technology has developed rapidly.This technology is an emerging technology that can quickly,directly,all-weather and real-time acquire 3D coordinates,color values,laser intensity and normal data of the measured object.Through 3D scanning the point cloud data have extensive and important applications,such as flood simulation,landslide monitoring,road design,land cover classification and forest management,and other fields.But the original point cloud data due to system error of instrument itself and artificially random error and the influence of the external environment factors-such as temperature,humidity,can produce more error points and empty,affect the subsequent processing and applications.Therefore,it has a certain value and significance to independently research and develop at 3D point cloud data processing system.At present,compared to the rapid development of 3D laser scanning hardware devices,many domestic and overseas point cloud processing software have more or less problems,such as high price and lacking of open source.The intelligence and stability of point cloud processing software,as well as the friendliness of software interface,professionalism and universality of application all need to keep pace with the era.Therefore,our research group develops a set of point cloud data processing and visualization software according to the experimental requirements.The software is mainly based on VS2015 as the development platform,C++ language as the main programming language,open source Liblas and PCL as the basis of point cloud data processing,combined with Qt tools to achieve point cloud data visualization.The main contents and innovations of this paper are as follows:(1)The basic knowledge of point cloud data is introduced in detail,including data format,organization form,nearest neighbor search algorithm,normal and curvature,etc.Then,based on this basis,the relevant point cloud processing algorithm is introduced.In the aspect of point cloud filtering,a variety of filtering algorithms are introduced,such as pass-through filtering,radius filtering,statistical filtering and voxel filtering.In the aspect of point cloud segmentation,the random sampling consensus algorithm is introduced,and the segmentation of plane model and cylinder model is realized.In the aspect of point cloud classification,the progressive morphology filtering algorithm,cloth simulation filtering algorithm,progressive TIN filtering algorithm and power line classification method based on straight-parabola are introduced.In the aspect of surface reconstruction,the moving least squares algorithm is introduced,and the smooth point cloud is realized.The greedy projection triangulation algorithm is introduced,and the construction of surface model is completed.At the same time,the above algorithms are integrated into the developed point cloud data processing system,and various algorithm experiments are carried out using the corresponding point cloud data,and the ideal experimental results are obtained,which verifies the fluency and stability of the system.(2)In order to improve the selection efficiency of the initial seed points in the traditional Progressive Triangular network Densification(PTD)algorithm,this paper proposes a airborne Laser Radar(Li DAR)point cloud filtering algorithm based on Cloth Simulation Filtering(CSF)and PTD.Firstly,the rough error points in Li DAR are removed,and then CSF is used to get the initial ground points after the rough error points is removed.Finally,the improved TIN is used to build a triangulation network for the initial ground points,and then the overall network is constructed for the ground points obtained by continuous iteration to get the final ground points.Three groups of test data provided by International Photogrammetry and Remote Sensing(ISPRS)Association were selected to carry out the experiments.The results show that the algorithm can reduce the error of I in the area with large slope,and control the error of II within a certain range,so as to verify the reliability of the algorithm.