Identification of the Agricultural Pests Based on Deep Learning Models

Pest identification is an important part of farming. It is crucial to accurately identify crop pests. In this background, based on the convolutional neural network, this paper collected 71 types of 35,000 images of pests, using the Inception-v3 and Inception-v4 model in GoogLeNet to build pests recognition model. Finally, based on several evaluation indicators, we combine the traditional methods for comparison and effect verification. The results show that the deep learning model is the best model for identifying pests and diseases in all the models involved in this study. Among them, the Inception-v4 support vector machine classification method has the highest accuracy, the overall classification accuracy rate is 97.3%, and the Inception-v3 classification accuracy is the second. The research can provide references for the identification of agricultural pests.