CAJ | 학술논문

传统的滤波算法通常是针对具有连续表面的简单区域来进行，因此带有一定的局限性，且不能解决复杂城区地形准确提取的难题。因此，本文提出了一种改进的基于TIN渐次加密的LiDAR点云数据滤波方法，该方法先对原始点云利用多尺度虚拟网格筛选地面种子点；然后，对种子点构建初始TIN表面，在此基础上进行向上加密；最后，得到的TIN三角网则为真实地形表面。实验结果表明，该方法能有效地滤除建筑物、植被和其他地物，并较好地保持地形特征。
전통적려파산법통상시침대구유련속표면적간단구역래진행，인차대유일정적국한성，차불능해결복잡성구지형준학제취적난제。인차，본문제출료일충개진적기우TIN점차가밀적LiDAR점운수거려파방법，해방법선대원시점운이용다척도허의망격사선지면충자점；연후，대충자점구건초시TIN표면，재차기출상진행향상가밀；최후，득도적TIN삼각망칙위진실지형표면。실험결과표명，해방법능유효지려제건축물、식피화기타지물，병교호지보지지형특정。
The traditional filtering algorithm usually aimed at a continuous surface and simple area to carry out , it always with some limitations.It does not solve the problem of accurately extracting the complex urban terrain .This paper presents an improved LiDAR point cloud data filtering method which is based on gradual encryption of TIN .In this paper , firstly, the method of multi -scale virtual grid is used to screen ground seed point from the original point cloud , and then use the seed point to build the initial TIN surface , on this basis, for up encryption , the last TIN terrain was the true surfaces .The results of the experiment show that this method can effec-tively filter out buildings , vegetation and other surface features , topographic features are kept better .