Recognition And Application Of Strawberry Pests And Diseases Based On Convolutional Neural Network

China is the largest country in strawberry production and consumption,Strawberry planting industry is one of the focal points of agricultural development in China.Reasonable monitoring and control of diseases and insect pests of strawberry is the basic method to increase the total yield of strawberry.The traditional method of manual inspection of diseases and pests relies on the professional knowledge of the inspection personnel,and the inspection process is inefficient,poor real-time,prone to errors caused by various factors.In order to solve the above problems,this paper presents a convolutional neural network based classification algorithm for strawberry disease and pest images,which introduces the attention mechanism and the central-loss function of the classical convolutional neural network,in order to improve the accuracy of image classification,the problem that the key region information of feature is lost greatly in the existing convolutional neural network classification methods is solved,which affects the effect of convolutional neural network classification.The main work of this thesis includes the following two aspects:(1)A depth image classification algorithm based on attention mechanism and center loss function is proposed.Specifically,the attention mechanism is introduced into the classical residual network ResNet34,in order to reduce the information loss in the process of image feature extraction,in this paper,the center loss function is introduced into the classical residual network ResNet34,and a depth image classification algorithm based on attention mechanism and center-loss function is proposed.In order to verify the performance of the proposed algorithm,the algorithm is compared with the classical convolutional neural network,ResNet34,and the experimental results from CIFAR10,Imagenet,MNIST,etc.the Algorithm proposed in this paper has made a certain breakthrough in the field of image classification.(2)The depth image classification algorithm based on attention mechanism and central loss function is applied to the recognition of strawberry diseases and insect pests,and a recognition system of strawberry diseases and insect pests based on convolutional neural network is designed and implemented.In order to extend the application of convolutional neural network in agricultural informatization,a fine tuning of the depth image classification algorithm based on attention mechanism and central loss function is proposed,this system is based on the convolutional neural network architecture,B/S structure,Django Web development framework and SQLite database management system,and is used in the recognition of strawberry pests and diseases,the invention realizes the functions of taking strawberry image by local camera and uploading it,and then recognizing strawberry diseases and insect pests online.