Research On Building Edge Feature Extraction And DEM Construction Method Based On LiDAR Point Cloud Data

Airborne LiDAR,as an emerging three-dimensional spatial data acquisition technology,has the characteristics of fast and effective,high precision and massive data,which provides important support for obtaining high-resolution spatial information.At present,airborne LiDAR point cloud data has become an important three-dimensional spatial data source and is used in many industries.Especially in the large-scale digital surface model data collection,it has obvious advantages.It can accurately describe building outlines,vegetation,the complex structure of roads,bridges and other features.The method establishment of quickly and accurately extracting the edges of buildings and constructing DEM by using airborne LiDAR point cloud data is the fundamental research content of airborne LiDAR point cloud data processing and 3D modeling.It has important theoretical significance and application value.In this paper,the method of extracting building edges using LiDAR data based on the combination of Canny algorithm and GVF-Snake model was explored,and the method of LiDAR data thinning algorithm and DEM construction was studied.The main research contents are as follows:(1)The characteristics of airborne LiDAR point cloud data collection and the causes of errors,and the current methods of airborne LiDAR point cloud data processing were studied.In this paper,based on the CSF cloth filtering algorithm,the experiment of the separation of ground points and ground points is carried out.On this basis,a filtering idea based on the fusion of scan lines and progressive triangle networks is proposed and explored.(2)Building edge extraction was the key research in this paper.The main contents include the acquisition of DSM distance images and DSM depth images,the filtering of image data and the extraction of building edges.Firstly,based on the elevation value and reflection intensity value of the airborne LiDAR point cloud data,the spatial data interpolation method was used to obtain the DSM range image and the DSM depth image;the DSM range image and the DSM depth image are filtered based on the adaptive filtering method,which is effectively eliminated the noise point.On this basis,the building edge extraction method based on the combination of Canny algorithm and GVF-Snake model was studied.Through the comparison experiments of straight-line edge detection methods with the Sobel operator and the Laplace operator it was proved that building edges can be expressed more precisely.(3)In this paper,a thinning method was established for airborne LiDAR point cloud data by improving the thinning algorithm based on the distance and height difference data and the one based on slope,which balances the relationship between the accuracy of DEM and the thinning rate of point cloud.The experiments of point cloud data thinning and DEM construction were carried out.The experiment showed that,under the condition of maintaining a certain DEM accuracy,appropriate point cloud data thinning had an acceptable impact on the terrain,and it could meet the needs of practical applications.Nevertheless,the speed of generating DEM could be increased.