Hyperspectral identification of cotton verticillium disease severity

Hyperspectral remote sensing provides fine spectral information for diagnosing crop disease severity, and in this paper the application of the hyperspectral remote sensing in identifying cotton verticillium disease severity was investigated. The wavelet transform was employed to extract the principal information and reduce the dimensions of the hyperspectral reflectance data, which were measured for cotton blades in different disease severity. Then, four identification models were built using discriminant analysis, back propagation (BP) neural network, genetic back propagation (GA-BP) neural network and support vector machine (SVM). The effects of the four models were examined and it was indicated that the SVM approach was the best.