A Study On Data Filtering And DEM Interpolation Of Airborne Lidar Point Clouds

Airborne light detection and ranging(LiDAR),an airborne lidar technology,can quickly obtain high precision and dense 3D spatial information in large area measurement area.This technique has become an important means to obtain high resolution digital elevation model(DEM).Lidar point cloud data filtering is an extremely important step to generate DEM,which plays a key role in the production and application of subsequent digital products.Therefore,point cloud filtering is an important subject in the research of LiDAR system.At present,the research focus of LiDAR point cloud filtering is how to reduce the manual intervention to improve the filtering accuracy,to achieve automatic highprecision filtering and to preserve terrain to obtain high-precision DEM.In view of this situation,the main work and innovation of this paper include the following three aspects:1?On the basis of summarizing the research status of LiDAR point cloud data filtering algorithms at home and abroad,the advantages and limitations of each filtering algorithm are analyzed,which provides a reference for the improvement and innovation of the existing filtering methods.2?The basic theory system of airborne lidar is expounded systematically.The composition,data storage and characteristics of point cloud data of airborne lidar are introduced,and the difficulties of data processing are analyzed.It provides theoretical basis for practical engineering application and subsequent filtering algorithm design.3?Due to the existence of outliers and large areas of blank area in the process of point cloud filtering,the results of progressive morphological filtering often produce serious deviations.In this paper,the morphological closed operator is introduced to deal with the low outliers to reduce the influence on the filtering quality,and the gradient operator is used to fill the large blank area,which avoids the loss of precision caused by the nearest neighbor interpolation before filtering.The improved algorithm is tested by using ISPRS open test data and sample data.Compared with the original algorithm,the type I error and the total error of the whole filtering results of all samples are reduced.The results show that the improved LiDAR point cloud data filtering algorithm improves the accuracy of the algorithm,and can adapt to different terrain,filter the ground points in various complex environments,and effectively retain the terrain details.4?The main factors that affect the precision of DEM are interpolation method,geomorphologic type,distribution characteristics of sampling data and so on.The interpolation method is the direct factor,the geomorphologic type and the distribution feature of sampling data affect the accuracy of DEM through interpolation algorithm.According to the characteristics of DEM interpolation method,this paper selects terrain slope factor and sampling data distribution as indicators to study the adaptability of interpolation method.It provides a way to study the adaptability of DEM interpolation method under the condition of different geomorphological types and distribution characteristics of sample points.The filtered point cloud data are interpolated by Kriging interpolation method and inverse distance weighting method respectively.The results show that the DEM model constructed by inverse distance weighting method and Kriging interpolation method is suitable for the situation where the number of samples is large and the distribution is uniform.The DEM precision of inverse distance weighting method is better than that of Kriging interpolation method in large slope area.In practical application,it is necessary to select a suitable method to construct DEM model according to the actual terrain situation.