打开文本图片集
摘 要：【目的】背包式激光雷達作为一种新型激光雷达，具有其他激光雷达所没有的优点，但还没有相应的文献参考，因此本文使用背包式激光雷达进行落叶松单木因子提取，为背包式激光雷达的使用提供理论基础。【方法】以黑龙江省桦南县孟家岗林场9块落叶松人工林样地为研究对象，根据树木的形态特征，利用背包式激光雷达扫描样地获取点云数据，使用Lidar360软件对样地内的树木点云数据进行单木识别及胸径、树高的提取，同时与实测数据比较进行精度评价及相关性检验。【结果】①利用背包式激光雷达数据进行单木分割的单木匹配率较高，平均匹配率为80.20%;②单木胸径提取结果决定系数R2最低为0.8，最高可达0.97，均方根误差RMSE最高为1.92 m，最低为0.6 m，胸径提取精度最高为97.62%，最低为92.25%;③单木树高提取的平均精度为80.27%，提取结果与实测相比差距较大。【结论】背包式激光雷达扫描的点云数据可以用于样地内单木的识别;胸径的提取结果可靠性最高，能够满足林下胸径数据的采集;由于遮挡单木点云数据的树冠顶部扫描不完全，导致树高提取的精度相对较低。说明背包式激光雷达还不可以直接用于树高数据的采集和提取，需结合其他种类激光雷达数据进行更高精度的分析。
关键词：背包式激光雷达;点云数据;落叶松人工林;胸径;树高
中图分类号：S771.8 文献标识码：A 文章编号：1006-8023（2019）04-0014-08
Study on Individual Tree Factor Extraction of Larix Olgensis in Backpack Lidar
HUANG Xu1， JIA Weiwei1*， WANG Qiang2， ZHENG Yujie1， LIANG Yuzhao1
（1.College of Forestry， Northeast Forestry University， Harbin 150040;
2.College of Surveying and Mapping Engineering， Heilongjiang Institute of Technology， Harbin 150040）
Abstract：[Objective] As a new type of laser radar， Libackpack has the advantages that other laser radars do not have， but there is no corresponding literature reference. Therefore， this paper uses Libackpack to extract larch factor， which provides a theoretical basis for the use of Libackpack. [Method] Taking 9 plots of Larix Olgensis plantation sample land in Mengjiagang Forest Farm of Huanan County， Heilongjiang Province as the research object， according to the morphological characteristics of trees， the point cloud data was acquired by using the Libackpack scanning plot， and the forest individual tree detection in the sample land and extraction of the diameter at breast height （DBH） and tree height were carried out by Lidar 360 software， then compared with the measured data for accuracy evaluation and correlation test. [Result] （1） The matching rate of individual tree segmentation using Libackpack data was higher， with an average matching rate of 80.20%. （2）The determination coefficient R2 of individual tree DBH extraction results was 0.8 at the lowest and 0.97 at the highest， the root means squared error （RMSE） was 1.92 m at the highest and 0.6 m at the lowest， and the extraction precision of the diameter at breast height （DBH） was 97.62% at the highest and 92.25% at the lowest. （3） The average extraction precision of single tree height was 80.27%， and there was a great difference between the extraction results and the actual measurement. [Conclusion] The point cloud data scanned by Libackpack can be used for the individual tree detection in the sample land. The extraction result of DBH had the highest reliability， which can meet the requirements of collecting DBH data under forests. Due to sheltering， the incomplete scanning of the canopy top of individual tree point cloud data lead to the low extraction precision of the tree height relatively. It illustrated that Libackpack cannot be directly used for tree height data collection and extraction， and it was necessary to combine with other types of laser radar data for more precision analysis.