Crop Identification And Application Based On Convolutional Neural Network

Agriculture is the economic foundation of our country.The identification and control of crop diseases is an essential part of agricultural production activities,which can directly affect the economic benefits of agriculture.Therefore,the identification methods and accuracy of crop diseases are of great significance for agricultural development.The traditional methods of crop disease recognition are mainly based on artificial detection,which have low efficiency and can not realize automatic recognition,and the methods based on image feature engineering have many problems,such as heavy workload,difficult to identify accurately in different complex environments and disease diversity.The purpose of this paper is to find a better method of crop disease identification to solve the above problems.In recent years,deep learning is widely used in the field of image recognition,and convolutional neural network is one of the typical representatives of deep learning.This network not only has a strong learning ability,but also can realize the automatic extraction of image features and achieve a good recognition effect.Therefore,based on deep learning and focusing on convolutional neural network,this paper develops the recognition of crop diseases Research.The main research work of this paper includes the following three parts:(1)In view of the problems that insufficient feature extraction or feature loss in the process of crop disease identification by the single networks under normal circumstances,a method of crop disease recognition based on deep fusion convolution neural network is proposed.This method cascades the effective modules in the Inception V3 and DenseNet169 models,then extracts the diversity and deep disease features in the target data by using their feature extraction capabilities,and then merges and splices the extracted features by feature fusion strategy.In order to further improve the experimental results,then a series of data enhancement and image preprocessing technologies are used for the target data set,and the training optimization method is given.Finally,experiments are carried out on the target data set and compared with various methods.The experimental results show that the method used in this paper is better than other methods in crop disease identification.(2)in view of the problem of crop disease identification needs under actual development,a crop disease identification method combining migration learning with MobileNet is proposed.This method uses ImageNet as auxiliary data set to pre-train MobileNet model,and uses the model-based transfer learning method to transfer the model and parameters,and finally uses the transferred model to carry out training recognition on the target data set and compares with various methods.The experimental results show that the method proposed in this paper has better practical application value than other methods.(3)In order to provide a convenient tomato disease identification method,the system uses the MobileNet model for migration learning as the identification model,and uses the Flask framework to build a simple tomato disease identification system,which realizes the functions of system login and registration,image upload,disease identification,image recognition management and user management,etc.And it also provides a new way for tomato planters and related scientific researchers engaged in tomato research to identify tomato diseases and insect pests.