Study On The Methods Of Generating Digital Elevation Model In Forest Area Based On LiDAR Data

Light Detection and Ranging system is a system that use laser to scanning and positioning to the ground to obtain the spatial information directly. As a new means of surveying, LiDAR system has advantage of the other methods. Ways of data processing influence the using of use of LiDAR technology directly, surveying for forest area surface information is a content of basis surveying and mapping geographic information, this is also a hotpot in the research of LiDAR. In this paper, the author made great efforts in the extraction of ground points in the LiDAR dataset of forest area.In this paper, the main works and research emphasis is as follows:1. This paper introduces the principle of airborne LiDAR system、two kinds of laser waveform and the generation、structure and features of point cloud data, reviews the classic processing methods of point cloud data. This paper has make a brief review of point cloud data filtering methods, and concluded the problems needed to be solved is these methods.2. This paper concludes multiple echo characteristics of laser, the influence factors of the intensity and the intensity distribution of different features in forest area, puts forward a method that use intensity to classify point cloud into different kind of land feature. The method chooses a certain number of sample data to analysis the intensity distribution and echo information of point in different features,according to statistical results appearance will be classified into vegetation coverage and bare surface area.3. For the distribution of different kinds of echo type in the vertical structure of vegetation are discussed in this paper, improved the existing filtering method which use the last echo of multiple echoes to extract the ground point in forest areas. Firstly, classify the study area into vegetation coverage and bare surface area, points which belong to bare surface area are all ground points; secondly, points which belong to vegetation coverage area needed to be filtered, we have to choose last echo of multiple echoes as the initial dataset filter, the filtering process can be divided into two steps, first step: classify the initial dataset into ground points and non-ground points, second step: use a moving surface fitting method to remove a small amount of remained non-ground points; finally, combine this two parts of ground points into one dataset, and then this dataset can be generate a DEM in this forest area.4. This paper put forward a LiDAR data processing method to extract the ground point of LiDAR data in forest areas, this method has two advantages, one is that it overcome the difficulty that the existing filter methods caused by terrain complexity and the change of feature size; The other one is that it solve the problem caused by mass of point could, the test result meet the requirements of practical application on the precision.