A Thin-method Besed On Local Terrain Complexity Index For LiDAR Bare Earth Surface Point Cloud

Airborne LiDAR(Light Detection And Ranging)point cloud data has the characteristics of high precision,high density and complete preservation of terrain features.The establishment of DEM(Digital Elevation Model)from LiDAR point cloud data has become one of the main methods.DEM directly built from LiDAR point cloud data is slow to process,poor interactive experience and difficult to store due to the large amount of data.Establishing DEM after proper thinning of ground point cloud data is the solution to the problem.How to deal with the relationship between thinning rate and DEM accuracy and retain as much topographic feature points and feature lines as possible is the focus of research.In response to the above problems,this paper carried out the following research work:(1)Around the construction of terrain complexity index,the problem of parameter selection is discussed.Three principles of parameter extraction,relative independence and relative comprehensiveness are introduced;Point cloud data in experimental areas are used to evaluate the relative independence of parameters with correlation coefficient analysis and cluster analysis.So that core terrain factors with relatively good independence are screened out,including elevation standard deviation,slope entropy and terrain Roughness.(2)In order to quantify the local terrain complexity,the CRITIC(Criteria Importance Through Intercriteria Correlation)multi-factor comprehensive weighting method is selected,and the core terrain factors are used as parameters to construct the local terrain complexity measurement index calculation method in experimental area with different terrain.(3)An airborne LiDAR ground point cloud data thinning method is designed.The local terrain complexity index is used as the basis of point cloud thinning,and the checkpoint elevation accuracy is used as the constraint to thin the point cloud.The method of extracting and simplifying terrain feature lines is discussed.The terrain feature lines are added as constraints to the results of point cloud thinning to establish DEM,so as to improve the elevation accuracy of DEM.(4)The method and the other three methods have been compared and analyzed by two aspects: the relationship between the thin rate and the precision,the distribution of thinning point.When the thinning rate is enough high,the spatial distribution problem is considered to improve the thinning method in this paper.The requirement of DEM for elevation accuracy at different scales is taken as the constraint condition to determine the maximum thinning rate.The results show that the airborne LiDAR ground point cloud data thinning method based on local terrain complexity index can retain the sampling points and terrain feature lines with abundant terrain feature information.The high elevation accuracy has also been ensured.When the thinning rate is 90%,the method can keep the uniformity of sampling points in the thinning results of point clouds.The research work in this paper has certain engineering practical value for generating multi-scale and high-precision DEMs by using airborne LiDAR point cloud data.