植被指数模型估算干旱区稀疏光合/非光合植被覆盖度  

Vegetation Index Model for Estimating Sparse Photosynthetic/Non-photosynthetic Vegetation Fractional Cover in Arid Zone

作　　者：骆义峡 姬翠翠 李晓松 徐金鸿 杨雪梅[3] LUO Yixia;JI Cuicui;LI Xiaosong;XU Jinhong;YANG Xuemei(Chongqing Jiaotong University,Chongqing 400074,China;Institute of Remote Sensing and Digital Earth,Chinese Academy of Sciences,Beijing 100094,China;Gansu Desert Control Research Institute,Lanzhou 730070,China)

机构地区：[1]重庆交通大学,重庆400074 [2]中国科学院遥感与数字地球研究所,北京100094 [3]甘肃省治沙研究所,兰州730070

出　　处：《遥感信息》2022年第3期57-64,共8页Remote Sensing Information

基　　金：国家自然科学基金项目(32060373);重庆市教育委员会科学技术研究项目(KJQN202000746)。

摘　　要：定量估算干旱与半干旱地区的稀疏光合与非光合植被覆盖度(f_(PV)和f_(NPV))可为土地荒漠化研究和生态系统研究提供重要信息,同时获取f_(PV)和f_(NPV)有助于为土地退化监测和土地管理提供重要的数据支撑。选取典型干旱区的稀疏植被为研究对象,基于Sentinel-2A多光谱影像及地面实测纯净端元光谱数据获取影像及纯净端元归一化植被指数(normalized difference vegetation index,NDVI)、优化的叶绿素吸收指数(modified chlorophyll absorption ratio index,MCARI)、干枯燃料指数(dead fuel index,DFI)和红边叶绿素指数(red-edge chlorophyll index,CI red-edge),构建DNVI-DFI、MCARI-DFI、CI-DFI 3个像元三分模型,并采用NDVI、DFI和比值土壤指数(ratio soil index,RSI)构建线性植被指数模型。经过地面实测植被覆盖度分析评价构建模型精度,得到如下结论:NDVI-DFI模型相较于MCARI-DFI、CI-DFI模型表现出更好的估算效果,f_(PV)估算的RMSE为0.0590(R^(2)=0.7738),f_(NPV)估算的RMSE为0.0510(R^(2)=0.8);NDVI-DFI模型中融入RSI指数构建的线性植被指数模型可以提高f_(PV)和f_(NPV)的估算精度,f_(PV)估算的RMSE为0.0524(R^(2)=0.7764),f_(NPV)估算的RMSE为0.0444(R^(2)=0.8115),精度分别提高11.2%和12.9%。因此,融入RSI指数的NDVI-DFI线性植被指数模型可以有效地估算典型干旱区稀疏植被的f_(PV)和f_(NPV)。该文的研究为NPV覆盖度的定量估算提供更可靠的理论基础。Quantitative estimation of sparse photosynthetic/non-photosynthetic vegetation fractional cover in arid and semi-arid regions provides important information for studies on land desertification research and ecosystem research.At the same time,to obtain photosynthetic/non-photosynthetic vegetation fractional cover can help provide important data support for land degradation monitoring and land management.In this paper,taking the typical arid area vegetation as the research object,based on Sentinel-2A to calculate NDVI(normalized difference vegetation index),MCARI(modified chlorophyll absorption ratio index),DFI(dead fuel index),CI_(red-edge)(red-edge chlorophyll index)and RSI(ratio soil index)index for constructing NDVI-DFI,MCARI-DFI,CI-DFI tri-endmember linear mixture model and NDVI-DFI-RSI linear index model.Finally,by verifying the vegetation fractional cover measured on the ground,analyzing and evaluating the accuracy of the constructed model,obtain the following conclusions:compared with the MCARI-DFI and CI-DFI models,the NDVI-DFI model shows a better estimation effect,the RMSE estimated by f_(PV)is 0.0590(R^(2)=0.7738),the RMSE estimated by f_(NPV)is 0.0510(R^(2)=0.8).The linear index model can improve the estimation accuracy of photosynthetic/non-photosynthetic vegetation fractional cover by 11.2%and 12.9%,respectively,the RMSE estimated by f_(PV)is 0.0524(R^(2)=0.7764),the RMSE estimated by f_(NPV)is 0.0444(R^(2)=0.8115).Therefore,the NDVI-DFI tri-endmember linear mixture model integrated with the RSI index can effectively estimate the photosynthetic/non-photosynthetic vegetation fractional cover of sparse vegetation in typical arid areas.The research in this paper provides a more reliable theoretical basis for the quantitative estimation of_(NPV)coverage.

关 键 词：Sentinel-2A 光合/非光合植被 像元三分模型 线性指数模型 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]