Identification Wheat Basal Rot And Wheat Take-all Using Hyperspectral Remote Sensing

This research on plant diseases and insect pests in remote sensing detection,Wheat Basal Rot and Wheat Take-all easily confused problems,all of them,and based on the imaging spectra and ground sample observation data,the canopy scale studies two kinds of diseases,spectral response of different scales were analyzed.Based on the analysis of income difference characteristics and screening of the characteristic parameters.Discrimination analysis model is established and the two kinds of disease samples mixed sample library to distinguish,the experimental results show that distinguish degree is good.In an actual Wheat Basal Rot and Wheat Take-all stem erosion of two kinds of condition assessment process,reduced the demand for human subjective experience,improve the recognition of the disease and computing speed.Implements in the imaging hyperspectral level,of unknown white head rate under the condition of two kinds of diseases.From the Wheat Basal Rot and Wheat Take-all of two kinds of different white head disease original spectrogram of subband to first order differential transformation of the original spectrum,choose different wavelengths get second order differential spectral difference band after processing.At the same time to choose spectral characteristic value and the vegetation index.Through IBM SPSS Statistics software,respectively by the stepwise discriminant method selection,the results show that the Db corresponding wavelength,Dy yellow side of first order differential value,NPCI,SIPI,RVSI and WI,application of Fisher Linear Discrimination(FLD)method and the Bayesian Linear Discrimination(BLD)method Karl fischer,Six index to the distinction between the two kinds of disease degree good,combined with the differences between the selected band,it is concluded that the two set of discriminant coefficient,and establish the corresponding discriminant function type,application of Fisher discriminant method and the Bayes Bayes discriminant method,Karl fischer mixed set discrimination analysis for three kinds of sample.Through the establishment of two discriminant function type,application 2 kind of criterion,the mixture of three kinds of samples set of sample for identification.The experimental results showed that 12 categories are distinguished by good effect.Two categories,this paper verifies the criterion of Wheat Basal Rot with erosion two kinds disease,on how to distinguish between the imaging hyperspectral feasibility.