Algorithm for Crop Pest and Disease Recognition Based on Residual Network

Considering the current mainstream convolutional neural network depth is now deeper and problems like many parameters, network degradation. This paper proposes a plant pest and disease detection model (FSCNN) based on residual networks. The model combines the improved residual network with the Batch Normalization (BN) technique, which improves the relationship between image features and pest categories, enhances the model learning ability and accelerates the network convergence speed. At the same time, a Dropout layer is added to prevent model overfitting and improve model generalization ability. The experimental results show that the recognition accuracy of the model reaches 98.64% on the PlantVillage dataset, and the method proposed in this paper features fast convergence speed and high recognition accuracy.