玉米叶片花青素相对含量高光谱遥感反演  

Nondestructive Inversion of Anthocyanins Content in Maize Leaves Using Hyperspectral Remote Sensing

作　　者：刘秀英[1,2,3] 刘晨洲 吴姗微[1,2,3] 石兆勇 常庆瑞 LIU Xiuying;LIU Chenzhou;WU Shanwei;SHI Zhaoyong;CHANG Qingrui(College of Agronomy,Henan University of Science and Technology,Luoyang,Henan471023,China;Luoyang Key Laboratory of Symbiotic Microorganism and Green Development,Luoyang,Henan471023,China;Luoyang Key Laboratory of Plant Nutrition and Environmental Ecology,Luoyang,Henan471023,China;College of Resources and Environment,Northwest A&F University,Xianyang,Shanxi 712100,China)

机构地区：[1]河南科技大学农学院,河南洛阳471023 [2]洛阳市共生微生物与绿色发展重点实验室,河南洛阳471023 [3]洛阳市植物营养与环境生态重点实验室,河南洛阳471023 [4]西北农林科技大学资源环境学院,陕西咸阳712100

出　　处：《遥感信息》2018年第6期1-8,共8页Remote Sensing Information

基　　金：国家高技术研究发展计划(863计划)项目(2013AA102401-2);河南省科技攻关项目(182102110206);河南科技大学博士科研启动基金(13480074)

摘　　要：针对传统植物叶片色素测量存在的问题,提出了一种采用高光谱遥感技术反演玉米叶片花青素相对含量的方法。以含花青素的玉米叶片为对象,获取玉米叶片的花青素相对含量及高光谱反射率,分析350～1 000nm波段范围内玉米叶片的反射光谱特征,建立基于敏感波段、已有光谱指数、新光谱指数的玉米叶片花青素相对含量反演模型,并进行分析、比较。结果表明,玉米叶片花青素相对含量的敏感波段为548nm;新光谱指数为521和698nm组成的比值指数(RI(521,698))、554和704nm组成的差值指数(DI(554,704))、557和701nm组成的归一化指数(NI(557,701))。基于RI(521,698)建立的一元线性和一元二次模型及DI(554,704)建立的一元二次模型的拟合R2均大于0.78,预测R2均大于或等于0.73,RMSE小于0.12,是进行玉米叶片花青素相对含量高光谱遥感反演的最优模型,说明在一定的精度范围内,新光谱指数具有无损反演玉米叶片花青素相对含量的潜力。Anthocyanins content in maize leaves is an important indicator of fertilizer deficiency,insect pest and extreme temperature.In order to investigate the feasibility of nondestructive inversion of total anthocyanins content in maize leaves using hyperspectral remote sensing,maize leaves with anthocyanins were collected.The anthocyanins content and reflectance spectra of maize leaves were measured.The reflectance spectra featrures of maize leaves were analyzed.The predictive models were constructed based on the sensitive band,existing spectral indices and new spectral indices using simple statistical regression.The results showed that the sensitive band of anthocyanins content in maize leaves was 548nm.The new spectral indices were the ratio index composed by 521nm and 698nm(RI(R521,R698)),the difference index composed by 554nm and 704nm(DI(R554, R704)),and the normalized index composed by 557nm and 701nm(NI(R557,R701)).The linear models and quadratic models based on RI(R521,R698)and the quadratic model based on DI(R554,R704)were optimum models for predicting anthocyanins content in maize leaves.The fitting and validation coefficients of determination R2 were higher than 0.78,the validation coefficients of determination R2 were higher than 0.73or equal to 0.73,and the root mean square error RMSE was less than 0.12.It could be concluded that new spectral indices have significant potential in the nondestructive inversion of total anthocyanins content in maize leaves.

关 键 词：花青素 高光谱 光谱指数 反演模型 玉米叶片 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S3[自动化与计算机技术—控制科学与工程]