基于Otsu方法点云粗分类的渐进三角网滤波算法研究  

Study on adaptive TIN filtering point clouds coarse classification based on otsu algorithm

作　　者：李成仁[1] 岳东杰[1] 于双[1] 

机构地区：[1]河海大学地球科学与工程学院,江苏南京210098

出　　处：《测绘工程》2014年第7期34-37,共4页Engineering of Surveying and Mapping

摘　　要：针对传统渐进三角网滤波方法需要针对不同的地形条件频繁调整滤波参数,并且对低矮地物滤波效果较差等问题,结合图像分割中的Otsu方法,提出一种基于Otsu方法点云粗分类的渐进三角网滤波算法。在对原始点云数据粗分类的基础上,以点云类别属性引导滤波过程。实验结果表明,方法简单可行,可以有效地控制低矮点被误分类成地面点的可能性,提高滤波处理结果的准确性。The traditional method of adaptive TIN filtering for different terrain conditions requires frequent adjustment of filtering parameters and the filtering effect on low surface features is poor. Combined with the image segmentation Otsu algorithm, an adaptive TIN filtering of point clouds coarse classification tmsed on Otsu algorithm is proposed. This algorithm is based on coarse classification of the original point clouds. Point clouds filtering process is guided by class attribute. The results show this algorithm is simple, which can effectively control the possibility of low points misclassified into the ground points and improve the accuracy of filtering process.

关 键 词：机载LIDAR 滤波 TIN OTSU方法 粗分类 深度图像 

分 类 号：P237[天文地球—摄影测量与遥感]