基于EFAST方法的苹果叶片叶绿素含量估算

Estimation of Chlorophyll Content in Apple Leaves Using Hyperspectral Data Based on EFAST Method

为了快速、准确地估算叶绿素含量,使用2012年和2013年在山东省肥城市潮泉镇获取的整个生育期苹果叶片叶绿素含量和配套的光谱数据,利用PROSPECT模型和EFAST方法探讨了对叶绿素含量敏感的波段,然后采用经验统计方法实现了单波段高光谱对苹果叶片叶绿素含量的监测。结果表明:以571 nm和697 nm波段光谱参数为自变量所建立的估测模型拟合精度较高,其决定系数(R2)分别为0.71和0.69,均方根误差(RMSE)分别为1.14、1.17 mg/dm^2,相对误差(RE)分别为-1.07%和-1.01%。以PROSPECT模型和EFAST方法整合筛选的敏感波段建立的估算模型监测叶绿素含量效果较好,为利用高光谱技术监测苹果长势提供了理论依据。

In order to estimate chlorophyll content quickly and exactly,chlorophyll content parameters and the concurrent spectral reflectance of apple leaves were acquired in Chaoquan town,Feicheng city,Shandong province,during 2012 and 2013 apple growth seasons. Sensitive wavebands to chlorophyll content were selected using PROSPECT model and EFAST( extended Flourier amplitude sensitivity test)method,and then estimation models of chlorophyll content were built using empirical statistical methods.Results showed that the fitting accuracy of the estimation models using wavebands of 571 nm and 697 nm was higher. The determination coefficients( R2) were 0. 71 and 0. 69,the root mean square errors( RMSE)were 1. 14,1. 17 mg/dm^2,and the relative errors( RE) were-1. 07% and-1. 01%. The model established by using PROSPECT model and EFAST method can predict the apple leaf chlorophyll content better,providing a theoretical basis for monitoring apple growth conditions by using hyperspectral technology.