Research On Point Cloud Data Filtering Method Of The Combination Of Progressive Triangulation And Cloth Simulation

In the research field of point cloud data processing,the classification of ground points and non-ground points is an important basic process of point cloud data processing.Analyzing the more commonly used filtering methods,it can be found that there are several types of methods,such as the progressive encryption triangulation filtering method and the cloth simulation filtering method,which have better filtering effects on areas with relatively smooth terrain undulations,which have been experimentally proved and are commonly used.However,actual research has found that in some more complex terrains,such as steep slopes,steep ridges,or low-rise buildings,or target areas of features,the terrain category is generally not unique,but a combination of multiple terrains.According to the above-mentioned traditional methods Filtering point cloud data in such complex terrain areas often results in poor filtering effects.Therefore,research on point cloud data filtering algorithms that can adapt to complex terrain regions still has very important theoretical significance and practical value.The main research contents of this paper are as follows:(1)Aiming at the difficult,time-consuming and inefficient problem of organizing and querying massive point cloud data,this article researches and proposes a point cloud data spatial indexing method based on a virtual grid.Through this algorithm,the point cloud can be better managed.Data,effectively improving the efficiency of querying and filtering massive point cloud data.(2)Using seven sets of standard point cloud data samples provided by ISPRS,they were respectively subjected to progressive encryption triangulation filtering and cloth simulation filtering,and the results of these two types of commonly used filtering methods and standard data were compared and analyzed.(3)Aiming at the problem of low reliability of single filtering method used in point cloud filtering of complex terrain,this paper proposes a filtering method that combines progressive encryption triangulation filtering and cloth simulation filtering.First,progressive encryption triangulation filtering is performed,and then interpolation is performed to obtain coarse DEM;On this basis,the original point cloud data and the coarse DEM are processed in unison to obtain the point cloud data that eliminates the terrain undulations;then the cloth simulation filtering algorithm is introduced for refined filtering processing.This "step-by-step combination algorithm" fully draws on the characteristics of the two traditional filtering algorithms,and can better solve the accuracy of the filtering results of the two types of traditional methods on steep slopes,steep ridges,and complex terrain or surface targets such as low buildings and vegetation.Impact.By comparing the filtering results of seven sets of standard data samples with the standard data,the results of the two traditional filtering methods and the combined filtering methods proposed in this paper are compared and analyzed,and the error model is combined to draw a conclusion.The results of the proposed step-by-step combined filtering method have obvious advantages.When the type II error is basically stable,the type I error is reduced by 3.67%-12.56% of the progressive encryption triangulation filter and 2.02%-20.54% of the cloth simulation filter.When it reaches1.16%-4.94% of the combined filtering algorithm,it is greatly reduced,and the total error is also reduced.