摘要
针对现有提取单木结构参数的方法精度不高的现状,文章研究利用地面三维激光扫描数据提取单木的高度、胸径和冠幅,特别是将RANSAC算法用于圆拟合,提高了胸径的反演精度。最后利用野外实测数据进行验证,并比较分析了Hough变换和RANSAC算法在提取胸径中的差异。结果表明,本文方法提取的单木参数精度较高;并且无论从相关系数、平均残差还是均方根误差等方面,RANSAC算法提取的胸径精度均高于Hough变换方法。
Aiming at the current status that the methods of extracting individual tree parameters have low accuracy, the paper studied to extract the individual tree height, DBH and crown width by using Terrestrial Laser Scanning （TLS） data, especially, an improved RANSAC algorithm was developed to estimate highly accurate DBH. Finally, experimental results showed that the accuracy of extracting the individual tree parameters by the proposed method would be high; and the comparative analysis demonstrated that RANSAC algorithm would get more accurate DBHs than Hough transform method.
出处
《测绘科学》
CSCD
北大核心
2015年第9期78-81,共4页
Science of Surveying and Mapping
基金
国家科技部国际合作重点项目(2013DFG21640)
国家自然科学基金项目(41271428
41171279)
中国科学院百人计划专项(09ZZ06101B)
