点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响.pdf
《点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响.pdf》由会员分享,可在线阅读,更多相关《点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响.pdf(11页珍藏版)》请在文库网上搜索。
1、doi:10.11707/j.1001-7488.LYKX20210831点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响*周相贝1,2李春干1代华兵3余铸1,3李振3苏凯1(1.广西大学林学院南宁 530004;2.广西自然资源职业技术学院扶绥 532199;3.广西壮族自治区林业勘测设计院南宁 530011)摘要:【目的】点云密度是影响机载激光雷达数据获取和预处理成本的关键因素,探明点云密度对森林参数估测精度的影响,为机载激光雷达大区域森林调查监测应用技术方案的优化提供参考依据。【方法】基于我国广西一个亚热带山地丘陵区域获取的机载激光雷达和样地数据,通过系统稀疏方法,将全密度点云(
2、4.35 点m2)分别稀疏至 4.0、3.5、3.0、2.5、2.0、1.5、1.0、0.5、0.2 和 0.1 点m2,得到 11 个样地尺度的点云数据集,包括 1 个全密度和 10 个稀疏密度点云数据集;应用配对样本 t 检验方法,分析 4 种森林类型(杉木林、松树林、桉树林和阔叶林)中稀疏密度点云和全密度点云之间 12 个激光雷达变量的差异;通过变量和结构固定的多元乘幂模型式,分别采用不同密度点云数据集对林分蓄积量(VOL)和断面积(BA)进行估测,比较模型优度统计指标决定系数(R2)、相对均方根误差(rRMSE)和平均预估误差(MPE)的差异,并应用 t 检验方法分析稀疏密度点云 VO
3、L 和 BA 估测值均值和全密度点云相应估测值均值的差异。【结果】1)点云密度较低时,稀疏密度点云分位数高度(ph25、ph50 和 ph75)的均值与全密度点云相应变量的均值存在显著性差异,但不同森林类型、不同变量出现显著性差异时的点云密度不同,各森林类型中稀疏密度点云平均高(Hmean)和点云高变动系数(Hcv)的均值与全密度点云相应变量的均值基本不存在显著性差异,但点云最大高(Hmax)的均值存在显著性差异;2)各森林类型中,稀疏密度点云冠层覆盖度(CC)和中下层分位数密度(dh25)的均值与全密度点云相应变量的均值差异不显著(阔叶林 dh25 除外),但中上层分位数密度(dh50 和
4、dh75)存在显著性差异;3)各森林类型中,稀疏密度点云平均叶面积密度(LADmean)的均值与全密度点云 LADmean的均值存在显著性差异,当点云密度较低时,稀疏密度点云叶面积密度变动系数(LADcv)的均值与全密度点云 LADcv的均值存在显著性差异;4)各森林类型中,不同密度点云 VOL 和 BA 估测值差异很小,且均不存在显著性差异,但随点云密度降低,杉木林、松树林和桉树林 VOL 和 BA 估测模型的 R2缓慢逐渐减小,rRMSE 和 MPE 缓慢逐渐增大,森林参数估测精度逐渐降低,阔叶林 VOL 和 BA 估测模型的 R2、rRMSE 和 MPE 受点云密度变化影响不大。【结论】
5、点云密度降低导致激光雷达变量标准差增大是造成森林参数估测模型精度降低的主要原因,在实际机载激光雷达森林资源调查监测应用中,点云密度以大于 0.5 点m2为宜。关键词:机载激光雷达;LiDAR 变量;林分蓄积量;断面积;模型中图分类号:S757文献标识码:A文章编号:10017488(2023)09002311Effects of Point Cloud Density on the Estimation Accuracy of Large-Area Subtropical ForestInventory Attributes Using Airborne LiDAR DataZhou Xian
6、gbei1,2Li Chungan1Dai Huabing3Yu Zhu1,3Li Zhen3Su Kai1(1.Forestry College of Guangxi UniversityNanning 530004;2.Guangxi Natural Resources Vocational and Technical CollegeFusui 532199;3.Guangxi Forest Inventory and Planning InstituteNanning 530011)Abstract:【Objective】Point cloud density is a critical
7、 factor affecting the cost of airborne LiDAR data acquisition and pre-processing.Therefore,exploring the influence of point cloud density on the estimation accuracy of forest inventory attributes canprovide a reference for optimizing technical schemes for airborne LiDAR-based large-area forest inven
8、tory and monitoring.【Method】In this study,we used airborne LiDAR data and field plot data collected in a subtropical mountainous and hilly regionin Guangxi,China.Firstly,the original point clouds with a density of 4.35 pointsm2 were reduced to 4.0,3.5,3.0,2.5,2.0,1.5,1.0,0.5,0.2,and 0.1 pointsm2 usi
9、ng a systematic thinning method,respectively,resulting in 11 plot-level point cloud datasets,收稿日期:20211116;修回日期:20220110。基金项目:广西林业科技推广示范项目(GL2020KT02);广西壮族自治区林业勘测设计院科研业务费项目(GXLKYKJ201601)。*李春干为通讯作者。本研究得到广西壮族自治区林业局资助并提供机载激光雷达数据,广西壮族自治区林业勘测设计院杨承伶和梁耀领导样地调查工作,众多人员参与样地调查,一并致谢。第 59 卷 第 9 期林业科学 Vol.59,No.9
10、2 0 2 3 年 9 月SCIENTIA SILVAE SINICAESept.,2 0 2 3including one full-density point cloud dataset and ten reduced-density point cloud datasets.Secondly,a paired sample t-test wasused to analyze the differences in 12 LiDAR-derived metrics between reduced-density point clouds and full-density point clou
11、ds infour forest types(Chinese fir,pine,eucalyptus,and broad-leaved).Thirdly,using a multiplicative power model formulation withfixed variables and stable structure,the stand volume(VOL)and basal area(BA)were estimated using various density datasets ofpoint clouds,respectively,and their goodness-of-
12、fit statistics,including coefficient of determination(R2),relative root square error(rRMSE),and mean prediction error(MPE),were compared.Finally,a t-test was used to analyze the differences in the means ofthe estimates between the reduced-density point clouds and full-density point clouds.【Result】1)
13、When the point cloud densitywas low,the means of the 25th,50th,and 75th height percentiles(ph25,ph50,and ph75)of the reduced-density point clouds showedstatistically significant differences from those of the corresponding variables of the full-density point clouds.However,whenstatistically significa
14、nt differences were found for different variables in various forest types,the point cloud densities differed.There were no statistically significant differences in the means of mean point cloud height(Hmean)and coefficient of variation ofpoint cloud height distribution(Hcv)between the reduced-densit
15、y point clouds and full-density point clouds in all forest types,butthere were statistically significant differences in the means of maximum height(Hmax)of point clouds between the reduced-densitypoint clouds and full-density point clouds for all forest types.2)The means of canopy cover(CC)and 25th
16、density percentile(dh25)of the reduced-density point clouds were not statistically significantly different from those of the corresponding variables ofthe full-density point clouds for all forest types(except dh25 for broadleaf forests),but statistically significant differences existedfor the 50th a
17、nd 75th density percentiles(dh50 and dh75).3)The means of the mean leaf area density(LADmean)of reduced-densitypoint clouds were statistically significantly different from those of the LADmean of full-density point clouds in all forest types,andwhile the means of the coefficient of variation of leaf
18、 area density(LADcv)of reduced-density point clouds were significantlydifferent from those of the LADcv of full-density point clouds when point cloud density was low.4)The differences in theestimates of VOL and BA for different density point clouds were small among the forest types,and none of the e
19、stimates werestatistically significantly different from each other.However,as the density of point clouds decreased,the R2 of the estimationmodels for VOL and BA for fir,pine,and eucalyptus forests slowly decreased,and the rRMSE and MPE slowly increased,indicating that the estimation accuracy of for
20、est inventory attributes gradually decreased.The R2,rRMSE,and MPE of theestimation models for VOL and BA for the broad-leaved forests were not obviously affected by the change in point cloud density.【Conclusion】The decrease in the density of point clouds leads to an increase in the standard deviatio
21、n of the LiDAR-derivedmetrics,which is the main reason for the decrease in the estimation accuracy of forest inventory attributes.In the operational forestresources investigation and monitoring,the airborne LiDAR point cloud density should be greater than 0.5 pointsm2.Key words:airborne LiDAR;LiDAR-
22、derived metrics;stand volume;basal area;model 机载激光雷达(light detection and ranging,LiDAR)点云数据详细、准确刻画森林冠层三维结构,全面反映林分冠层水平和垂直分布状况,为森林参数估测和制图奠定了坚实的生物物理基础,已逐步成为当前单木和林分尺度森林参数估测的先进遥感技术(李增元等,2016;曹林等,2013;Mascaro et al.,2011;Zolkos etal.,2013;Singh et al.,2015),并成功应用于各类森林,包括温带森林(庞勇等,2012;Ioki et al.,2010;Ahme
23、d etal.,2013)、北 方 森 林(Nsset,2004a;Thomas et al.,2006;Maltamo et al.,2016)、热带林(Drake et al.,2003;Garca et al.,2017)、地中海森林(Garca et al.,2010;Montealegre et al.,2016)、高 度 集 约 经 营 的 桉 树(Eucalyptus)人工林(Grgens et al.,2015)以及城市森林(He et al.,2013;Giannico et al.,2016)、灌 木 林(Estornell et al.,2011)、林下植被(Estorn
24、ell et al.,2011;2012;Li et al.,2017)等,从国家尺度(McRoberts et al.,2010;Watt et al.,2013)和省州尺度(Johnson et al.,2014)的森林资源监测到企业尺度(Straub et al.,2013)的森林资源调查评估均有大量成功应用案例。随着传感器技术的快速发展,机载激光雷达数据精度得到极大改善(Renslow et al.,2000),由离散激光雷达系统产生的点云密度和每个脉冲的回波数量呈指数增加趋势(Singh et al.,2016),一些传感器如 RieglQ680i 和YellowScan Mappe
25、r 等,点云密度达3040 点m2(Latifi et al.,2015),可提取枝条等十分详细的树冠结构信息(Vauhkonen et al.,2013)。在以直升机为平台进行低空飞行和窄扫描角的情况下,激光雷达点云密度高达每平方米上千个甚至数千个(Pearse et al.,2018),能够极为精确刻画林木的枝、干结构,减轻地面调查工作量(Kellner et al.,2019)。点云密度增加有24林业科学59 卷 助于提高冠层垂直剖面描述精度和森林参数估测精度;然而,点云密度与数据获取成本呈正相关关系,点云密度高,意味着飞行高度降低,扫描条带变窄,导致数据获取成本增大,且对于大面积森林监
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- 密度 机载 激光雷达 区域 亚热带 森林 参数 估测 精度 影响