摘要
探讨以不同的植被指数建立的高光谱模型对玉米叶面积指数LAI的反演精度。实测不同水肥耦合作用下,玉米冠层的高光谱反射率与叶面积指数（Leaf Area Index）数据,采用高光谱红光波段（631～760 nm）与近红外波段（760～1 074 nm）逐波段构建NDVI、RVI、DVI、TSAVI、PVI植被指数,分别找出与LAI具有最佳相关性波段组合的植被指数,建立玉米LAI估算模型。结果显示,与LAI具有最佳相关性的波段组合分别是NDVI（R760,R990）、RVI（R760,R1001）、DVI（R677,R1070）、TSAVI（R760,R975）、PVI（R658,R966）,它们反演玉米LAI的确定性系数分别：R^2〉0.72、R^2〉0.74、R^2=0.95、R^2〉0.79、R^2〉0.95。结果表明,在玉米的整个生长季的47个样本中,通过PVI和DVI方式建立的遥感估算模型能够较为准确地估算玉米LAI,TSAVI次之,NDVI、RVI稍差。
An experiment was carried out to evaluate the precision of hyperspectral reflectance model to monitor corn leaf area index（LAI）.Corn were cultivated under water-fertilizer coupled control condition,and corn LAI was collected simultaneously with LI-COR LAI-2000,and Corn canopy reflectance data were collected with ASD spectroradiometer（350～1 074nm）.Firstly,each band of NIR and red were applied to establish five Vegetation Indices;secondly,find out the best band for each kind of Vegetation Index respectively;finally,five Vegetation Indices with the best reflectance band were applied to regress against corn LAI.The result shows that the best Vegetation Indices with reflectance which could be applied to regress against corn LAI were DVI and PVI,however,TSAVI was not as well as DVI and PVI,but better than NDVI and RVI.
出处
《遥感技术与应用》
CSCD
2007年第5期586-592,共7页
Remote Sensing Technology and Application
基金
中国科学院知识创新重要方向性项目(KZCX3-SW-356)
中国长春净月潭遥感站网络台站基金资助
