| Airborne LiDAR (Light Detection And Ranging) system provides a bran-new technical instrument for the three dimensional data collection. It obtains 3D information at first hand fleetly and accurately. Therefore, it plays a more and more important role in the applications such as terrain mapping, environmental surveillance, coastal surveying and 3D city modeling. In addition, it indicates the developmental direction of observation technology on the earth to some extent. This dissertation focuses on the theme of airborne LiDAR data classification, and puts keystone on the study of airborne LiDAR data filtering, non-ground point cloud classification and ground point cloud classification, which gets along in theory and arithmetic to some extent. The major works and innovations are included as:1.The components airborne LiDAR system and the format of point cloud are introduced. The characters of the typical objects are analysed. Then the existing methods of airborne LiDAR point cloud filtering are reviewed and concluded, some issues which need to be settled are summarized.2.By analyzing the key problem of airborne LiDAR data filtering methods, a filtering method of airborne LiDAR data based on hierarchical pseudo-grid concept and slope threshold is put forward. The experiment and error analysis validate the efficiency and veracity of the method in this dissertation.3.After summarizing the data types and methods used to classifying, a classification strategy of non-ground point cloud integrating echo information, height texture and spectrum information is presented. And then, the workflow is established, which achieved fusion of LiDAR data and aerial image and supply a gap of spectrum on LiDAR data.4.The character of the echo intensity is expatiated, and the possibility of ground point cloud classification using intensity data is discussed. A method of ground point cloud classification based on clustering on intensity is brought forward. And an experiment using two clustering method had been carried out. |