热带作物学报

基于无人机LiDAR的橡胶树单木地上生物量估测

PDF下载

包栎炀 ¹ , 王祥军 ² , 李少达 ¹^,^** , 谭骏祥 ¹ , 黄肖 ² , 宋佳倩 ¹ , 贾文惠 ¹ , 吴满意 ³

热带作物学报 | 农业生态与环境保护 2023,44(6): 1266-1275

收起

热带作物学报 | 农业生态与环境保护 2023, 44(6): 1266-1275

基于无人机LiDAR的橡胶树单木地上生物量估测

全屏

包栎炀¹, 王祥军², 李少达¹^,^**, 谭骏祥¹, 黄肖², 宋佳倩¹, 贾文惠¹, 吴满意³

作者信息

^1.成都理工大学地球科学学院，四川成都　610000

^2.中国热带农业科学院橡胶研究所，海南海口　571101

^3.自然资源部第一地形测量队，陕西西安　710054

包栎炀（1998—），女，硕士研究生，研究方向：激光雷达林业应用。

通讯作者:

^** 李少达（LIShaoda），E-mail: lisd@cdut.edu.cn。

Estimation of Individual Tree Above-ground Biomass of Rubber Tree Based on UAV-LiDAR

Liyang BAO¹, Xiangjun WANG², Shaoda LI¹^,^**, Junxiang TAN¹, Xiao HUANG², Jiaqian SONG¹, Wenhui JIA¹, Manyi WU³

Affiliations

^1.College of Earth Sciences, Chengdu University of Technology, Chengdu, Sichuan 610000, China

^2.Rubber Research Institute, Chinese Academy of Tropical Agricultural Sciences, Haikou, Hainan 571101, China

^3.The First Topographic Survey Team, Ministry of Nature Resources, Xi'an, Shaanxi 710054, China

出版时间: 2023-06-25 doi: 10.3969/j.issn.1000-2561.2023.06.020

文章导航

摘要

收起

橡胶树地上生物量是反映橡胶树生产力、固碳能力和碳储量的重要指标，为提高橡胶树单木地上生物量估测的效率和精度，以橡胶树8年生育种试验林为研究对象，采用无人机LiDAR获取林地点云数据，并实测其地上生物量，基于点云数据提取树高、冠幅、树冠投影面积和树冠体积4个单木结构参数，将此4个参数作为预测因子建立橡胶树单木地上生物量估测模型，比较多元非线性回归和随机森林回归2种模型的估测精度，并分别采用五折交叉验证的方法对2种模型的泛化能力和可靠性进行评估。结果表明：（1）在单木分割的基础上由算法提取的树高和冠幅与直接基于点云人工量测的数值存在很好的一致性，2个参数与实测值的Pearson相关系数分别为0.999和0.951，均方根误差（RMSE）分别为0.109 m和0.452 m；（2）树高、冠幅、树冠投影面积和树冠体积等4个单木结构参数与橡胶树实测单木地上生物量相关性显著，其中树冠体积与地上生物量的Pearson相关系数最高，达0.904，4个参数对橡胶树地上生物量都具有良好的解释性；（3）基于4个单木结构参数建立的2种橡胶树单木地上生物量估测模型均能够取得很好的拟合效果，但是随机森林回归模型的表现更优，其决定系数（R²）为0.85，估测精度较多元非线性回归模型提高了3.64%，相对均方根误差（rRMSE）为30.90%，较多元非线性回归模型减小2.66%。总体而言，随机森林回归模型的拟合优度更高，泛化能力更强，可以更加准确地估测橡胶树单木地上生物量。

关键词

无人机LiDAR / 橡胶树 / 地上生物量 / 单木结构参数

Abstract

收起

The above-ground biomass (AGB) is an important indicator to reflect the productivity, carbon sequestration capacity, and carbon storage of rubber tree. However, the AGB models of individual rubber tree with high estimation efficiency and accuracy are still needed to develop. In the present study, an 8-year-old forest established for rubber tree breeding trial was scanned to obtain the point cloud data by using Unmanned Aerial Vehicle-Light Detecting and Ranging (UAV-LiDAR), meanwhile, the actual AGB of every tree in this forest was measured experimentally. Four individual tree structure parameters, such as tree height, crown width, crown projected are, and crown volume, were extracted from the point cloud data, and then, used as predictors to establish the individual rubber tree AGB model. The multiple nonlinear regression and random forest regression were both applied to establish the model, and the estimation accuracy, generalization ability and reliability were evaluated by five-fold cross-validation. The tree height and crown width values extracted by special algorithms from individual tree point clouds data were highly correlated with the values manually measured on point clouds, the Pearson correlation coefficients of the two parameter wase 0.999 and 0.951 respectively, and the root mean square errors (RMSE) was 0.109 m and 0.452 m respectively. The correlations between the four structure parameters and the AGB of individual rubber tree was significant. Especially, the parameter of crown volume had the highest Pearson correlation coefficient (0.904). All four parameters had good explanations for AGB. Both established AGB models based on the four individual tree structure parameters could achieve good fitting results. However, the method of random forest regression had a better performance compared with multiple nonlinear regression. The coefficient of determination (R²) of the random forest regression model was 3.64% higher than that of the multiple nonlinear regression model, and the relative root mean square error (rRMSE) of the random forest regression model was 2.66% lower than that of the multiple nonlinear regression model. In general, there is higher goodness of fit and stronger generalization ability of the random forest regression model which can more accurately estimate the rubber tree AGB.

Key words

UAV-LiDAR / rubber tree / above-ground biomass / individual tree structure parameters

引用本文

包栎炀, 王祥军, 李少达, 谭骏祥, 黄肖, 宋佳倩, 贾文惠, 吴满意. 基于无人机LiDAR的橡胶树单木地上生物量估测. 热带作物学报, 2023 , 44 (6) : 1266 -1275 . DOI: 10.3969/j.issn.1000-2561.2023.06.020

Liyang BAO, Xiangjun WANG, Shaoda LI, Junxiang TAN, Xiao HUANG, Jiaqian SONG, Wenhui JIA, Manyi WU. Estimation of Individual Tree Above-ground Biomass of Rubber Tree Based on UAV-LiDAR[J]. Chinese Journal of Tropical Crops, 2023 , 44 (6) : 1266 -1275 . DOI: 10.3969/j.issn.1000-2561.2023.06.020

基金

收起

国家重点研发计划项目(2019YFD1000500)
中国热带农业科学院基本科研业务费专项资金(1630032022007)
四川省科技计划项目(2021YJ0369)

参考文献

收起

2023年第44卷第6期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.3969/j.issn.1000-2561.2023.06.020

接收时间：2022-08-22
首发时间：2026-03-05
出版时间：2023-06-25

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2022-08-22
修回日期：2022-09-25

基金

国家重点研发计划项目(2019YFD1000500)

中国热带农业科学院基本科研业务费专项资金(1630032022007)

四川省科技计划项目(2021YJ0369)

作者信息

^1.成都理工大学地球科学学院，四川成都　610000

^2.中国热带农业科学院橡胶研究所，海南海口　571101

^3.自然资源部第一地形测量队，陕西西安　710054

通讯作者:

^** 李少达（LIShaoda），E-mail: lisd@cdut.edu.cn。

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/rdzwxb/CN/10.3969/j.issn.1000-2561.2023.06.020

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT