摘要
植被病叶色素含量的变化在光谱上会有响应,因此通过光谱测度的方法可以对植被病叶色素含量进行反演,从而能够对作物病害进行即时监测。偏最小二乘法(PLS)的成分提取不仅是对光谱反射率信息的概括,也是对色素含量信息的解释,同时成分提取可以避免变量之间的多重相关性,减少信息冗余。本文通过对加工番茄细菌性斑点病病叶色素含量和光谱反射率的测定,对400nm~800nm的原始光谱进行一阶微分、二阶微分、反对数光谱变换。利用色素含量在原始光谱、一阶、二阶和反对数光谱的敏感波段建立归一化指数,采用多波段PLS算法对色素含量进行归一化指数估测并检验。结果表明,新建归一化指数组成的PLS模型对色素Chl.a、Chl.b、Cars含量的估测精度高于传统的归一化指数,R2都大于0.8,RMSE小于0.05。该模型的应用对加工番茄细菌性斑点病有更强的监测能力。
Determine the leaves of the processing tomatoes bacterial spot disease and spectral reflectance.The original spectum are transformed to one-order differential spectrum;second-order differential spectrum;against several spectrum in 400-800nm.The sensitive spectrum in original spectrum;one-order differential spectrum;second-order differential spectrum;against several spectrum make the normalized index.Optimizing the normalized index estimate pigment contents and inspection using PLS(partial least-squares)algorithm.Results show that the estimation precision of Chl.a,Chl.b,Cars in new normalization index of PLS model is higher than that of traditional estimation,the value of R2 is more than 0.8,the value of RMSE is less than 0.05.Research result show that it has the monitoring ability that the new normalization index of PLS model for processing tomato bacterial spot disease pigment content.
出处
《遥感信息》
CSCD
2012年第4期50-56,共7页
Remote Sensing Information
基金
国家自然科学基金项目(30800733)资助
国家科技支撑项目(2007DAH121301)项目资助
关键词
加工番茄
病害胁迫
色素含量
高光谱
估测
tomato processing
disease stress
pigment content
hyperspectral
estimation