A Method for Identifying Leaf Pests and Diseases Based on an Improved Convolutional Neural Network

In view of the importance of identifying leaf pests and diseases in agricultural production and the shortcomings of the existing technology, this article puts forward a method of identifying leaf pests and diseases based on improved CNN (Convolutional Neural Network). The purpose of this study is to improve the accuracy and robustness of leaf pests and diseases identification by optimizing the CNN model structure, and to provide more efficient and accurate pest monitoring means for agricultural production. In terms of methods, this article designs a multi-scale feature fusion CNN architecture, and introduces innovations such as deep separable convolution, attention mechanism and improved pooling strategy. Through experimental verification and comparative analysis, it is proved that the improved model is superior in the task of leaf pests and diseases recognition. The results show that compared with the unmodified CNN model, the improved model has significantly improved the recognition accuracy, and has strong generalization ability under different pest types and growth conditions. By improving the structure of CNN model, this study successfully improved the accuracy and robustness of leaf pests and diseases identification, and provided strong technical support for pest control in agricultural production.