A Study On DEM Construction Based On Airborne LiDAR Point Cloud

The airborne LiDAR(Light Detection and Ranging)has become the most efficient method for gaining spatial data which has a high resolution.The airborne LiDAR uses an active technology of remote sensing which has less restriction on climate conditions to acquire point cloud data with high temporal and spatial resolutions.And DEM products are on the basis of these data.At present,the main problems in the Study on DEM Construction Based on Airborne LiDAR Point Cloud,are how to acquire accurate point cloud data and how to make an efficient automatic filtering processing to eliminate noise combined with less human-computer interaction and remain the key terrain features,has become the vital aspect.In this paper,those problems are focused to be solved.The coastal area in Yantai is chosen to be the research area.LiDAR acquires raw point cloud data of costal area after air lines,check line and flight quality have been planned or checked.The automatic filtering processing which is on the basis of the moving window method and human-computer interaction are used to gain accurate point cloud data.The DEM is interpolated by point cloud data.Elevation data measured by real time kinematic(RTK)is compared with DEM data which is on the basic of LiDAR point cloud data.The main contents and results of this paper are as follow:Giving an account of each part for airborne LiDAR system and its measuring principle.And analyzing the data errors sources and the methods of improving accuracy.By clarifying the characteristic of photogrammetry and In SAR,we compare LiDAR with them and analyze its advantages and disadvantages.The measurements of LiDAR could acquire more accuracy data because that laser has penetration which could penetrate vegetation.Airborne LiDAR system acquire high precision point cloud data of the coastal area in Yantai for the first time.To acquire the accurate data,the noise should be eliminated firstly.Before eliminating the noise,point cloud data should be given the correct WGS-84 coordinate.Then the obvious noise is preliminarily removed on the basis of elevation.After the pretreatment,the most noise has been eliminated.Then,filtering processing is in place,which uses a moving window method.The scale of the moving window has been changed some times to eliminate the noise.Finally,remained noise is eliminated by Human-computer interaction.High precision LiDAR point cloud data the coastal area in Yantai is used to construct DEM for the first time.To avoid large redundancies and improve precision,Triangulated Irregular Network(TIN)model is used to simulate terrain on the basis of Delaunay triangulation algorithm.DEM is constructed by linear interpolation of TIN,because the spatial density of could point data is so high that reduce the errors and its operating rate is better.The technology of the net real time kinematic(RTK)has been introduced in and to acquire 151 elevation data using this method.Analyzing the logical relation of contours,slope anomaly,root mean square error(RMSE),mean absolute error(MAE)and mean relative error(MRE)to evaluate the accuracy of DEM data.The MRE and MAE are 0.105 m and 0.231 m respectively,and the RMSE is 0.285 m.