A Study Of Algorithm For Generating Raster Grid Based On Massive Airborne LIDAR Data

As a revolutionary new technology in the field of surveying and mapping appearing in recent years, airborne LIDAR technology has been widely used in various fields, for it has lots of advantages, such as high precision, fast speed, little influenced by weather and terrain.However, the massive LIDAR point clouds data has brought unprecedented challenges to the existing algorithms, softwares and applications, which largely restricts airborne LIDAR technology for further development and applications.There are many data processing theories and algorithms of massive airborne LIDAR data,but this paper only study the theory and algorithm for generating raster grid from massive LIDAR point clouds.This paper summaries and analysises the airborne LIDAR data characteristics,the problems and shortcomings of exsting algorithms.To solve the existing shortcomings,firstly,a slicing construction algorithm of Delaunay triangulation is proposed to solve the contradiction between time performance and space performance when Delaunay triangulation is biulded.Based on the linear discriminant mathematics principle of positive and negative region,a new algorithm of generating raster grid from TIN is presented.Taking the advantage of streaming computer dealing with massive data,a theoretical system and processing flow is provided by the author to solve the problem that it takes much time to do I/O operation with TIN in converting massive airborne LIDAR point clouds data to raster data via TIN.The experiments and analysises show that the slicing construction algorithm of Delaunay triangulation has a good time and space performance. The algorithm of generating raster grid from TIN proposed in this paper performs more efficiently in time than existing algorithm and there is a linear relation between the cost time and the number of triangulars, which makes it suitable for massive airborne LIDAR data processing.The streaming algorithm has better performance in time owing to less I/O operation time, requires low computer configuration and fits for processing massive airborne LIDAR point clouds data and parallel processing.