Early Detection And Grading Method Of Soybean Mosaic Disease Based On Hyperspectral Imaging Technology

As a crop with a long planting history,soybean not only has rich nutrients,but also has the effect of resisting cholesterol.At present,soybean yield is reduced due to the decrease of rotation cycle and the inadequate prevention of diseases and insect pests.Soybean mosaic virus(SMV)disease has appeared around the world.That has affected soybean yield seriously.Therefore,it is particularly important to realize the detection of soybean mosaic diseases.The current research on soybean disease detection mainly focuses on the middle and late stages of soybean disease,in order to achieve timely prevention of soybean mosaic disease,this paper studied the nondestructive detection method for the initial diagnosis of soybean mosaic disease and the severity of soybean mosaic disease based on hyperspectral imaging technology.The main research contents are as follows:(1)In order to reduce the impact of mosaic disease on soybean production and explore a theoretical basis for detection and warning of early soybean mosaic disease,this paper provided a method of early detection of soybean mosaic disease based on SPA2-ELM.Hyperspectral image acquisition of soybean leaves and normal leaves after inoculation of SC3,SC7 soybean mosaic virus for seven days.The spectral was preprocessed by SG smoothing and PMSC.Then,the feature wavelengths ware performed by successive projection algorithm(SPA).There were 9 characteristic wavelengths(405?461?522?552?626.6?705?743.4?855?947nm)Further,in order to reduce the amount of calculation,SPA processing was performed again to obtain 4 optimal characteristic wavelengths(461?552?705?855nm).Three models of extreme learning machine(ELM),artificial neural network(ANN),least squares support vector machine(LSSVM)classification algorithm were established based on full-band information and characteristic wavelength information for initial diagnosis of soybean mosaic disease.Experiments showed that the model with SG smoothing was better than the model with piecewise multi-scatter correction preprocessing.The SPA2-ELM model maintained good accuracy under the premise of removing data redundancy.The accuracy of the model training set reached 89.59%.The prediction set accuracy was 87.5%.(2)To further improve the accuracy of the initial diagnosis model of soybean mosaic disease,an early detection method of soybean mosaic disease based on CNN model was proposed.The convolutional layer of the model was two layers,the ReLu nonlinear activation function was added to the convolutional layer,and the Max pooling pooling function was used in the pooling layer.Compared with the LSSVM and ELM,the recognition rate of the CNN model was higher than that of the LSSVM and the ELM.Regardless of the recognition rate of different types of soybean sample sets or the recognition rate of the overall soybean sample sets,the recognition rate of CNN models was higher than that of LSSVM and ELM.The correct rate of the model training set was 94.79%,and the correct rate of the prediction set was 92.08%.(3)In order to better achieve the growth monitoring of soybean mosaic disease,a method for grading detection of soybean mosaic disease based on CNN-SVM model was proposed.soybean leaves with different disease levels and normal growth were divided into 0,1 and 2 grades,and the severity of soybean mosaic disease was graded.A classification model based on convolutional neural network and support vector machine was proposed.The fully connected layer of the convolutional neural network model was connected to the support vector machine to solve the error caused by the small sample.Compared with the convolutional neural network model alone,the detection result was more accurate.The accuracy of the training set reached 96.67%,and the accuracy of the test set reached 94.17%.It was demonstrated the feasibility of using convolutional neural network combined with support vector machine to classify soybean mosaic disease,and it provided a new direction for the more effective detection of soybean mosaic disease based on hyperspectral image.