基于单站地面激光雷达数据的自适应胸径估计方法

发布时间：2018-05-18 07:29

本文选题：遥感 + 地面激光雷达　；参考：《激光与光电子学进展》2016年08期

【摘要】：为了自动化地、准确地从单站地面激光雷达(TLS)数据中提取一定范围内的树木胸径,提出一种基于点云切片的圆形-椭圆自适应胸径(DBH)估计方法。对林地点云数据在胸高位置进行切片,然后对胸高切片点云进行聚类,利用圆形-椭圆自适应拟合方法对聚类结果进行树干点判别,符合圆形分布的树干点集直接用于计算树木胸径,符合椭圆分布的树干点集进行胸高位置校正之后再进行胸径计算。利用TLS在北京市东升郊野公园的人工柳树林进行样地点云数据采集,验证圆形-椭圆自适应胸径估计方法,并与单纯圆形拟合方法对比。结果显示,在扫描距离为26m的样地范围内,树木胸径估计均方根误差为1.1cm,在扫描距离为56m的样地范围内,树木胸径估计均方根误差为1.99cm,判别为椭圆分布的树木胸径估计结果平均误差比单纯圆形拟合结果降低4.7%。该方法可以快速有效地进行自适应胸径估计。
[Abstract]:In order to automatically extract a certain range of tree DBH from single station ground lidar (TLSs) data, a method based on point cloud slice to estimate the DBH of circular ellipse adaptive DBH is proposed. The cloud data of forest location were sliced at the height of the chest, then the point cloud of the section of the height of the chest was clustered, and the tree trunks were identified by the method of circular ellipse adaptive fitting. The tree DBH is directly used to calculate the tree DBH by the tree trunk point set according to the circular distribution, and the DBH is calculated after the chest height position is corrected by the trunk point set consistent with the elliptical distribution. The cloud data of the artificial willow forest in Dongsheng country Park of Beijing were collected by TLS to verify the rotund-ellipse adaptive DBH estimation method and compared with the simple circular fitting method. The results show that the root mean square error of tree DBH estimation is 1.1 cm in the sample plot range of 26 m scanning distance, and in the sample plot range of 56 m scan distance, the root mean square error of tree DBH estimation is 1.1 cm. The root mean square error (RMS) of tree DBH estimation is 1.99 cm. The average error of tree DBH estimation based on elliptical distribution is 4.7cm lower than that of simple circular fitting. This method can be used to estimate the DBH quickly and effectively.
【作者单位】：北京林业大学理学院;
【基金】：北京市自然科学基金青年基金(6164037) 北京林业大学优秀青年教师科技支持专项计划(YX2015-08);北京林业大学青年教师科学研究中长期项目(2015ZCQ-LY-02)
【分类号】：TN958.98

【相似文献】