摘要
Abstract
In order to improve the speed and accuracy of point cloud data filtering,this paper proposes an improved progres-sive refinement triangulated irregular network(PTD)point cloud filtering algorithm.In the seed point selection phase,the method draws on the moving surface filtering concept,replacing the fixed grid with a dynamically moving virtual grid.This allows for more accurate extraction of ground seed points,significantly improving the efficiency of subsequent iterations.In the distance threshold determination step,the maximum inter-class variance method(OTSU is introduced.By calculating the distance from each pending point in the initial triangulated irregular network(TIN)to the TIN surface,the distance threshold is determined.Additionally,the slope threshold is computed based on a large number of ground seed points.Experimental results show that for both flat rural and sloped urban point cloud datasets,the average values of Type Ⅰ error,Type Ⅱ error,and total error for the improved algorithm are 8.8％,5.54％,and 7.43％,respectively.These results repre-sent reductions of 4.84％,4.33％,and 4.83％over the classical algorithm,with overall filtering accuracy significantly bet-ter than the classical method.关键词
点云滤波/无人机航测点云/渐进加密不规则三角网算法/移动面滤波Key words
point cloud filtering/unmanned aerial vehicle(UAV)photogrammetry point cloud/progressive refinement triangulated irregular network algorithm/moving surface filtering分类
天文与地球科学
