论文部分内容阅读
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。
In order to automatically and accurately extract the DBH of a certain range from single station ground lidar (TLS) data, a round-ellipse adaptive DBH method based on point cloud slicing is proposed. The cloud data of the forest sites were sliced in the chest height position, then the cluster points of the chest high-cut slices were clustered. The trunk points were identified by round-ellipse adaptive fitting method and the trunk points of the circular distribution were directly used Calculate the diameter at breast height, the elliptic distribution of trunk points set after the chest height position correction and then calculate the diameter at breast height. The TLS was used to collect cloud data from the artificial willow forests in Dongsheng Country Park in Beijing to verify the circular-elliptical adaptive DBH method and to compare with the simple circular fitting method. The results showed that the mean root mean square error of DBH was 1.1cm and the root mean square error of DBH was 1.99cm within the scope of sampling distance of 56m at the sampling distance of 26m. The average error of the estimation results of DBH distribution decreased by 4.7% than that of the simple circular fitting result. This method can quickly and effectively estimate the DBH.