Design of Vegetable Pest Identification System Based on Improved Alexnet Algorithm and 5G Communication

Abstract

Vegetable pests and diseases are some of the main factors affecting vegetable yield. Accurate monitoring and intelligent identification of vegetable pests and diseases are prerequisites for pest forecasting and integrated control. In this study, a vegetable pest identification system based on an improved Alexnet algorithm and 5G communication is designed. The system uses high-definition cameras and 5G communication modules to form the pest monitoring network. It builds an image recognition model based on the improved Alexnet algorithm to identify vegetable pests, and then it collects pictures for transmission to the terminal. After the experimental test, the pest identification system proposed in this study accounts for only 11.71, 11.91, 30.92, and 31.38% of the identification system of the 4G communication network in terms of transmission delay, transmission jitter, packet loss rate, and packet error rate, respectively. The recognition accuracy of the improved Alexnet algorithm is 18.76% higher than that of the unimproved one. After multiple iterations, it is verified that the recognition accuracy and loss function are better than those of the unimproved Alexnet algorithm. It shows that the identification system proposed can better monitor and identify vegetable pests and diseases, which is beneficial to integrated management.