Research On 3D Point Cloud Data Reduction And Fast Browsing Technology

In recent years,with the continuous progress of science and technology,the way people get to know the word not only depends on two-dimensional space of the graphic image,but also the three-dimensional space.The digital earth,digital city,digital community and so on are constantly appearing in our daily life,which all contribute to the rapid development of 3D point cloud technology.The precision of 3D point cloud acquisition equipment is becoming higher and higher,resulting in lots ofpoint cloud data.The massive data makes later processing of point cloud data(such as: browsing,splicing,feature extraction,surface reconstruction,storage and transmission,etc.)complicated,bringing more pressure and challenges on the hardware and software equipment In the ordinary PC end,the data browsing and the man-machine interaction process is easy to make the memory consumption to be serious,the processing to be slow and so on.Therefore,this paper aims at the problem of data reduction and fast browsing of massive scattered point cloud data,the scattered point cloud processing technology are thoroughly analyzed,algorithm of point cloud simplification on the basis of relevant theory is improved,based on the 3D slice theory scattered point cloud simplification algorithm;And in the process of point cloud fast browsing,an improved algorithm is used,improving the point cloud browsing speed.The focus of this paper is as follows:(1)On the basis of the research on the algorithm of scattered point cloud simplification,an improved point cloud simplification algorithm based on slicing theory is proposed.In this paper,the point cloud data section is splitted;then according to the geometric characteristics of each point cloud slices,each layer of point cloud are sorted;finally point cloud simplification is realized using D-P algorithm.The experimental results show that the proposed algorithm can reduce the number of point cloud data.(2)In the point cloud quickly browse,four different methods are applied.They areoctree spatial index method,memory mapped file technology,visual field cutting technology and multi-level LOD technology.In addition,the improved visual field cutting algorithm is used to improve the speed of point cloud.(3)In the windows system,with the point cloud beforehand processing software PointCloudViewer is written,using the C++ programming language under the VS platform based on the open-source cloud library PCL and Visualization Toolkit VTK.The data import and export,point cloud display,data simplification,fast browsing and a series of functions are realized,and the experiment is carried out with some point cloud model data,The result shows that the algorithm can achieve better results,fully preparing for the follow-up of the point cloud modeling.