Research On Soybean Pest Image Recognition Method Based On Optimized Convolutional Neural Network

Soybean is one of the important economic crops in China.It is often affected by various diseases and insect pests during planting.Therefore,the timely acquisition,rapid identification and scientific control of soybean plant diseases and insect pests are of great significance for guiding agricultural production.The application of image processing technology in the diagnosis and treatment of crop diseases and insect pests has always been a research focus in the process of agricultural informatization,especially the application of convolutional neural networks in agricultural field image recognition in deep learning algorithms,showing that the algorithm has been applied to crop diseases and insect pests.Compared with traditional image recognition methods,this algorithm simplifies the pre-processing work of manual feature selection and data mining.In order to solve the problem of effective identification of soybean pest images,in order to solve the problem of effective identification of soybean pest images,this paper conducted a research on soybean pest image recognition methods based on optimized convolution neural network The main research contents and achievements are as follows:(1)Several algorithms of machine learning are used to compare the image recognition methods of soybean pests.Because of the diverse characteristics of different images,sometimes,the color discrepancy between pests and plants are not obvious,which can cause interference for feature extraction.Taking aphids as an example,this paper proposes the extraction methods of soybean pest features combining the color information and texture features in the image,and uses the classical SVM,BP neural network and K-means classification methods respectively,which can effectively extract aphids from the complex background environment and compare the recognition results.(2)The attention mechanism is introduced into the field of computer vision,and the image recognition model of soybean pests is constructed based on the attention convolutional neural network.According to the feature vector set of the acquired farmland soybean aphid images,a convolutional neural network model based on attention mechanism is proposed by using the optimized convolutional neural network algorithm and connecting the convolutional neural network for coarse-scale image as input to the convolutional neural network with fine-scale images as input.The optimized network model can be effectively trained through alternating training method,and the collected soybean aphid data can be focused to the effective feature area of the image,so as to solve the problem of soybean aphid identification,which achieves the accurate identification of soybean aphids and obtains the ideal results,more importantly provides an significant basis for precise application.(3)An intelligent image recognition and control system for soybean pests is developed and verified.Relying on the soybean disease identification method and based on the convolutional neural network,an intelligent image recognition and control system for soybean pests is developed,and the Tensorflow Model trained by Python is used for image recognition to complete the function implementation and verification of business modules such as system users management,information of common pests,identification and diagnosis,the pest control,expert consultation and other business modules.