Research On Simplification Algorithm Based On Image Matching Point Cloud

With the rapid development of China's "smart city" construction,3d city model,as an important basic data for the construction of "smart city",has become increasingly demanding in many fields such as urban planning,traffic management,land use and social services.The tilt photogrammetry technology has been rapidly developed with the continuous updating of the drone platform.The technology has the characteristics of low cost,high efficiency,real model and rich information,and plays a huge role in promoting the development of "smart city".The core idea of 3D modeling of tilt photogrammetry technology is to construct a triangular grid model based on image matching point cloud.However,the amount of point cloud data acquired by image matching is huge,which increases the load of computer,which is not conducive to the later data processing and 3D model construction.Therefore,it is especially important to realize the simplification of the image matching point cloud and to express the three-dimensional model with the least number of points for the establishment of the three-dimensional city.Based on the theory of tilt photogrammetry,this paper acquired the image matching point cloud data of the building model based on image matching method,and preprocessed the image matching point cloud to obtain the experimental data for the research on the simplified algorithm in this paper.Aiming at the problems of dense point cloud,large amount of data and redundancy of image matching obtained by tilt photogrammetry.Several existing point cloud reduction algorithms are adopted to streamline the experimental data and analyze the advantages and disadvantages of several existing simplification algorithms.On this basis,a combined improved algorithm is proposed.The main research and results are as follows:(1)The point cloud space division method and the existing point cloud simplification algorithm are studied,and the random sampling method,uniform grid method and curvature sampling method are taken as examples to simplify the experimental data,and the advantages and disadvantages of the three algorithms are analyzed through the simplification effect under different simplification rates.(2)Aiming at the shortcomings of the above three existing algorithms,a combined reduction algorithm that preserves the edge features of the building and does not cause data holes in the building plane area is proposed.The basic idea is: Firstly,the experimental data is spatially divided by the three-dimensional grid method to establish a spatial index.On this basis,the curvature reduction algorithm is used to streamline the experimental data,and the points larger than the average value of the curvature are retained.The curvature reduction algorithm is smaller than the average value of the curvature.The cull points are resampled as the data source of the uniform mesh reduction algorithm.Based on the spatial index of the 3D grid,the closest point to the centroid point is calculated and retained.Finally,the merged curvature reduction algorithm and the uniform grid method are retained.The point is the final streamlined result of the experimental data in this paper.(3)The improved algorithm is used to reduce the experimental data,and compared with the simplification effect of the three existing simplification algorithms,the feasibility of the improved algorithm is proved.(4)The surface area evaluation method and volume evaluation method are adopted to evaluate the three existing simplification algorithms and the improved algorithm.It is proved that the improved algorithm is indeed better than the three existing simplification algorithms for point cloud data and has good practical value.