利用无人机遥感测定岷江冷杉单木树干生物量

Estimation of stem biomass of individual Abies faxoniana through unmanned aerial vehicle remote sensing

无人机获取的高分辨率遥感影像,已成为单木生物量估算的有效手段。本文以四川省王朗自然保护区岷江冷杉为研究对象,利用本项目组自行研制的无人机获取无阴影遥感影像。设置2块样地,其中plot 1为建模样本,plot 2为验证样本。通过人机交互的方式提取单木树冠面积(CA)数据,并结合野外实测的胸径(DBH),建立DBH遥感估算模型。最后基于CA-DBH模型的有效性,结合已有DBH-SB(树干生物量)经验方程,计算plot 2岷江冷杉单木树干生物量。结果表明:基于无人机遥感影像提取的单木CA与实测DBH存在较好的非线性相关关系,所建立的模型有较好的拟合度,R2达到0.752(P<0.001,n=94)。采用t检验验证CA-DBH模型预测值与观测值的差异,同时计算皮尔森相关系数(Pearson Correlation Coefficient),检验结果表明:该模型估算的DBH与实测值偏差差异不显著(P>0.05),其皮尔森相关系数可达0.879,证明利用无人机获取的遥感影像,通过提取的CA估算DBH是可行的。本次实验表明:利用无人机遥感获取影像,通过提取的单木CA进行树干生物量的估算是有效的。

Fast and accurate quantification of biophysical parameters of trees is essential for forest management, assessment of carbon sequestration and evaluation of regional ecosystem services value. Unmanned aerial vehicle （UAV） is a promising tool to estimate biomass of individual trees due to its extremely high resolution. In this study, we used self-developed UAV to obtain shadow-free remote sensing images, taking Abies faxoniana in Wanglang Nature Reserve of Sichuan Province as an example. There were two plots, one for model training and the other for model validation. Crown area （CA） of individual trees was delineated through man-computer interpretation. Meanwhile, the field inventory was conducted to record the diameter at breast height （DBH） of individual trees, and to establish CA-DBH regression model. Based on the validity of CA-DBH model, the stem biomass （SB） of individual A. faxoniana trees in plot 2 was derived according to the existing empirical DBH-SB equation. There was a strong nonlinear correlation between CA extracted from the UAV remote sensing images and DBH documented in the field visit, with a coefficient of determination R2 = 0. 752 （P 〈0. 001, n = 94）. Then prediction of DBHusing the model in plot 2 was conducted, followed by a T-test. Verification results showed that the difference was not significant between the predicted DBH and observed DBH in the field （P 〉 0.05） , with a Pearson correlation coefficient of 0. 879. This study indicates that it is practically feasible to estimate SB of individual trees through the CA extracted from UAV remote sensing images