冬小麦典型多参量冠层高光谱反演的光谱指标敏感性研究  

Sensitivity Research on the Canopy Hyperspectral Inversion for the Typical Multi-parameter of Winter Wheat

作　　者：韩茜 张潇元[1,2] 王树东 张立福 张霞 田静国 

机构地区：[1]中国地质大学地球科学与资源学院,北京100083 [2]中国科学院遥感与数字地球研究所高光谱遥感应用技术研究室,北京100101 [3]广西空间信息与测绘重点实验室,广西541004

出　　处：《科学技术与工程》2017年第25期89-97,共9页Science Technology and Engineering

基　　金：国家自然科学基金(41371359;41671362);广西空间信息与测绘重点实验室开放基金(151400727);高分辨率对地观测系统重大专项资助

摘　　要：高光谱遥感反演作物参量多集中在单一参量研究上,缺乏多参量综合反演研究。因此,面向多参量反演,需要对传感器参数的有效设置,以及同一指标对不同参量的适宜程度进行综合研究。以冬小麦为例,通过实测冠层光谱反射率和LAI、叶绿素、氮素含量数据,分析了中心波长、波段宽度、信噪比等指标的变化对各参量定量模型的影响,及光谱指标对LAI、叶绿素和氮素定量模型反演的敏感性和有效性,以及对冬小麦典型参量高光谱遥感反演的光谱指标进行了综合性分析。结果表明:反演冬小麦LAI的最佳植被指数为DVI(R^2=0.457,RMSE=0.614%),对应的最佳指标为:中心波长为768 nm和732 nm,波段宽度在5 nm以内,信噪比大于70 d B;反演冬小麦叶绿素的最佳植被指数为MSR(R2=0.554,RMSE=0.548%),对应的最佳指标为:中心波长为768 nm和736 nm,波段宽度在5 nm以内,信噪比大于70 d B;反演冬小麦氮素的最佳植被指数为NDVI_(g_b)(R^2=0.733,RMSE=0.600%),对应的最佳指标为:中心波长为500 nm和454 nm,波段宽度为5 nm以内,信噪比大于70d B。植被指数SAVI在一定波段范围内可同时反演LAI、叶绿素和氮素;MSAVI、DVI、RDVI和NDVI均可在一定波段范围内同时反演LAI和叶绿素含量,而反演LAI和氮素含量的适宜波段以及反演叶绿素和氮素的适宜波段存在差异。利用高光谱植被指数可实现作物参量的有效反演,且作物参量的定量反演对不同的光谱指标,即中心波长、波段宽度和信噪比具有较强的敏感性。The study on hyperspectral remote sensing inversion of crop parameters mostly concentrate on single parameter,lack of multi-parameter comprehensive inversion. Therefore,for the multi-parameter inversion,it need effective setting of sensor parameters and the comprehensive study on the suitability of the same parameters for different parameters. In this paper,winter wheat was taken as an example to analyze the change of center wavelength,band width and SNR index influence on quantitative model,through the measured canopy reflectance and LAI,chlorophyll,nitrogen data,and the sensitivity and effectiveness of spectral index for inversion of LAI,chlorophyll and nitrogen content with quantitative models,as well as the comprehensive of hyperspectral remote sensing inversion for typical parameters in winter wheat. The result showed that： DVI was the best vegetation index for LAI inversion with the center wavelength of 768 nm and 732 nm,band width of less than 5 nm,SNR greater than 70 d B（ R^2= 0. 457,RMSE = 0. 614%）. MSR was the best vegetation index for chlorophyll content inversion with the center wavelength of 768 nm and 736 nm,band width of less than 5 nm,SNR greater than 70 d B（ R^2=0. 554,RMSE = 0. 548%）. NDVI_（g_b） was the best vegetation index for nitrogen content inversion with the center wavelength of 500 nm and 454 nm,band width of less than 5 nm,SNR greater than 70 d B（ R^2= 0. 733,RMSE =0. 600%）. SAVI could simultaneously invert LAI,chlorophyll and nitrogen content within a certain wavelength range. MSAVI,DVI,RDVI and NDVI could simultaneously invert LAI and chlorophyll content within a certain wavelength range,while the suitable wavelength for inversion of LAI and nitrogen content had no intersection,as well as chlorophyll and nitrogen. The inversion of crop parameters can be realized by using hyperspectral vegetation index,and quantitative inversion of crop parameters had a strong sensitivity with different spectral index—center wavelength,band width,SNR.

关 键 词：高光谱反演 冬小麦 植被指数 叶面积指数 叶绿素 氮素 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]