WorldView-2纹理的森林地上生物量反演

发布时间：2018-04-15 08:34

  本文选题：地上生物量 + 纹理因子　； 参考：《遥感学报》2017年05期

【摘要】：使用高空间分辨率卫星WorldView-2的多光谱遥感影像,构建植被指数和纹理因子等遥感因子与森林地上生物量的关系方程,并计算模型估测精度和均方根误差,探索高分辨率数据的光谱与纹理信息在温带森林地上生物量估测应用中的潜力。以黑龙江省凉水自然保护区温带天然林及天然次生林为研究对象,通过灰度共生矩阵(GLCM)、灰度差分向量(GLDV)及和差直方图(SADH)对高分辨率遥感影像进行纹理信息提取,并利用外业调查的74个样地地上生物量与遥感因子建立参数估计模型。提取的遥感因子包括6种植被指数(比值植被指数RVI、差值植被指数DVI、规一化植被指数NDVI、增强植被指数EVI、土壤调节植被指数SAVI和修正的土壤调节植被指数MSAVI)以及3类纹理因子(GLCM、GLDV和SADH)。为避免特征变量个数较多对估测模型造成过拟合,利用随机森林算法对提取的遥感因子进行特征选择,将最优的特征变量输入模型参与建模估测。采用支持向量回归(SVR)进行生物量建模及验证,结果显示选入模型的和差直方图均值(sadh_mean)、灰度共生矩阵方差(glcm_var)和差值植被指数(DVI)等遥感因子对森林地上生物量有较好的解释效果;植被指数+纹理因子组合的模型获得较精确的AGB估算结果(R2=0.85,RMSE=42.30 t/ha),单独使用植被指数的模型精度则较低(R~2=0.69,RMSE=61.13 t/ha)。
[Abstract]:Using the multispectral remote sensing image of high spatial resolution satellite WorldView-2, the relationship equation between vegetation index and texture factor and forest aboveground biomass was established, and the estimation accuracy and root mean square error of the model were calculated.To explore the potential of spectral and texture information of high-resolution data in the estimation of aboveground biomass of temperate forest.Taking temperate natural forest and natural secondary forest in Liangshui Nature Reserve of Heilongjiang Province as the research object, the texture information of high resolution remote sensing image was extracted by GLCM, GLDV and SADH.A parameter estimation model was established by using the aboveground biomass and remote sensing factors of 74 plots investigated in the field.The extracted remote sensing factors include 6 cropping cover indices (ratio vegetation index RVI, differential vegetation index DVI, normalized vegetation index NDVI, enhanced vegetation index SAVI and modified soil regulated vegetation index MSAVI) and 3GLDV and SADH.In order to avoid the overfitting of the estimation model caused by the large number of feature variables, the extracted remote sensing factors were selected by the stochastic forest algorithm, and the optimal feature variables were input into the model to participate in the modeling and estimation.Support vector regression (SVR) was used to model and verify the biomass. The results showed that the sum-difference histogram mean value of the selected model, grey co-occurrence matrix variance glcmvar.) and differential vegetation index (DVI) had better interpretation effect on forest aboveground biomass.The model with texture factor combination of vegetation index obtained more accurate AGB estimation results than that using vegetation index alone, the accuracy of the model was lower than that of the model using vegetation index alone at 61.13 t 路ha ~ (-1) ~ (-1) ~ (-1) ~ (-1) ~ (-1) ~ (-1) 路ha ~ (-1) ~ (-1) ~ (-1) ~ (-1) ~ (-1) ~ (-1) ~ (-1).
【作者单位】： 北京师范大学信息科学与技术学院;中国林业科学研究院资源信息研究所;
【基金】：国家重点基础研究发展计划(973计划)(编号:2013CB733406,2013CB733404) 国家高技术研究发展计划(863计划)(编号:2012AA12A306) 中央高校基本科研业务费专项资金(编号:2015KJJCA12) 中央级公益性科研院所基本科研业务费专项资金项目(编号:CAFYBB2016ZD004)~~
【分类号】：TP751
，

本文编号：1753376