Thinning Algorithm Of LiDAR Bare Earth Surface Point Cloud Under The Restriction Of Precision

Using airborne LiDAR (Light Detection And held) point cloud data to generate Bare-Earth DEMs, has been one of the most elective ways to express ground form in the discipline of space science. Its high precision and high density makes the production’s precision is much higher than the actual demand. However, the huge amount of data directly affects the speed of Bare-Earth DEMs’production, data storage and the facility of data interaction. In other to get an easy operated and processing result based on the precision rules, it’s significant to make data thinning to the point cloud data of bare surface.Many scholars have made many researches in the LiDAR point cloud data thinning area, and there were many remarkable achievements. However, how to better retain the terrain feature points is still the key and difficulty in this problem. In addition, how to make a reasonable points’distribution whilst keeping the terrain feature points as far as possible is stillan ignore problem among the existing researches. According to the above, this paper has carried out the following work:(1) This paper has summarized the techniquesof airborne LiDAR point cloud data filtering, the method of gross error elimination. The experiments containthe filtering of original airborne LiDAR point cloud data, the gross error elimination;(2) This paper has summarized the algorithms of thinning the LiDAR point cloud data both domestic and overseas, and analyzed the limitation of the algorithms above;(3) To study the concept of terrain complexityand use "slope" as an important operator to describe. Put forward the concept of slope-entropy, and the method of quantifying the terrain complexity of local area by slope-entropy. Finally come up with the algorithm of thinning LiDAR point cloud based on the slope-entropy;(4)To compare with the existing data-thinning algorithms and which based on slope-entropy by testing on three types of the terrain--flat ground, hills and mountainland. To find out the relationship between the thinning-percentage and the precision of DEM from the fitting curves graph, and generate the points-distribution graphs from every data-thinning algorithms to make evaluations on three types of terrain;(5) Carrying out experiments using existing data-thinning algorithms and algorithm based on slope-entropy, with precision constraint for different types of terrains under a certain scale according to the DEM specification, data-thinning algorithm based on slope-entropy is proved to be reasonable on the location and spatial distribution of reserved of data points, with the largest dilute ratio.It is shows that the algorithm of thinning LiDAR point cloud based on the slope-entropy not only can keep the local feature well, also can keep the global feature well, which is better than the other algorithms and has the higher compression ratio.under the DEM actual project demand.