激光点云分类质量检查与解决方案探讨

Quality Inspection and Solution of Laser Point Cloud Classification

点云分类中"人工判别编辑"的好坏直接影响DEM质量。目前,分类成果的质量检查主要要采取人工逐屏浏览检查方式,该方法耗时长,检查不全面。针对该问题,本文对点云分类常见质量错误进行归纳分类,并提出相应的解决方案,三种解决方案采用三种不同算法。从实验结果可以看出,采用本文的检查方法能够快速准确地定位错误位置,在保证检查完整率的同时大幅度数提高检查效率。

The quality of"manual discrimination editing"in point cloud classification directly affects DEM quality.Currently,the quality inspection of classification results mainly adopts manual screen-by-screen browsing inspection method,which takes a long time and is not comprehensive.In order to solve this problem,this paper summarizes and classifies common quality errors in point cloud classification,and proposes corresponding solutions.The three solutions adopt three different algorithms.From the experimental results,it can be seen that the inspection method in this paper can quickly and accurately locate the wrong position and greatly improve the inspection efficiency while ensuring the inspection integrity rate.