摘要
建立合成孔径雷达(synthetic aperture Radar,SAR)数据和光学植被指数的定量关系有助于融合这两种数据源,提高山区森林遥感的时序监测能力。为此,以内蒙古大兴安岭根河林区为例,首先分析了归一化植被指数(normalized difference vegetation index,NDVI)、增强型植被指数(enhanced vegetation index,EVI)、绿度植被指数(greenness vegetation index,GVI)和归一化水分指数(normalized difference water index,NDWI)与C波段雷达数据的相关性,接着对比了不同森林干扰下NDVI,NDWI与X,C,L波段雷达数据的相关性差异。结果表明:①极化比(polarization ratio,PR)和干涉相干系数与各植被指数呈显著负相关,PR与NDVI,EVI,GVI线性趋势好(R^2=0.40~0.49),VH的干涉相干系数与各植被指数线性趋势好(R^2=0.43~0.51);②地表类型会影响VH与NDVI线性回归结果,在植被密集的灌草、火烧迹地和森林内线性趋势好(R^2=0.64~0.76);③不同森林干扰下相关性存在差异:在火烧迹地内NDVI与X波段HH和C波段PR呈显著负相关,NDWI与C波段VH呈显著正相关;在未受干扰林地内NDVI和NDWI与C波段PR呈显著正相关;在采伐迹地内L波段PR与NDVI呈显著负相关,L波段VV和VH的PR与NDWI呈显著负相关。
Vegetation index is an important approach in vegetation monitoring and investigation.SAR data are free with weather condition observing data day and night.Building relationships between SAR data and vegetation indices can contribute to fusing two data to improve temporal monitoring in forest of mountain areas.Therefore,the authors made a statistical analysis between vegetation indices including NDVI,EVI,GVI,NDWI and C band SAR data and then made a comparison about difference of correlation between NDVI,NDWI and X,C,L band SAR data in different forest disturbances in Genhe forest region of Da Hinggan Mountains in Inner Mongolia.The results are as follows:①PR and interferometry coefficients both have significant negative correlations with optical vegetation indices,PR has strong linear correlations with NDVI,EVI,GVI(R^2=0.40~0.49),and interferometry coefficients have strong linear correlations with all optical indices(R^2=0.43~0.51).②Ground cover can affect linear regression between VH and NDVI.Scrub-grass land and fires scars with thick vegetation layer and forest land have a strong linear correlation with NDVI(R^2=0.64~0.76).③The correlations are different for different forest disturbances:In fires scars,NDVI has significant negative correlations with X-band HH,and C band PR and NDWI have a significant positive correlation with C band VH.In deforestation areas,L-band PR has significant negative correlations with NDVI,and L band VV and VH have significant positive correlations with NDWI.In undisturbed forest land,C-band PR has significant negative correlations with NDVI and NDWI.
作者
王川
范景辉
林思美
饶月明
黄华国
WANG Chuan;FAN Jinghui;LIN Simei;RAO Yueming;HUANG Huaguo(The Key Laboratory for Silviculture and Conservation of Ministry of Education, Beijing Forestry University, Beijing 100083, China;China Aero Geophysical Survey and Remote Sensing Center for Natrual Resources, Beijing 100083, China)
出处
《国土资源遥感》
CSCD
北大核心
2020年第2期130-137,共8页
Remote Sensing for Land & Resources
基金
国家重点研发计划项目“大兴安岭火烧及采伐迹地植被恢复遥感监测及其辅助决策技术”(编号:2017YFC0504003-4)
国家自然科学基金项目“基于相关生长理论的森林光学微波信息互补机理研究”(编号:41971289)共同资助。
关键词
植被指数
SAR遥感参数
火烧迹地
森林采伐
相关性
vegetation index
SAR remote sensing parameter
fire scar
deforestation
correlation