Research on agricultural environmental monitoring Internet of Things based on edge computing and deep learning

• DOI: 10.1515/jisys-2023-0114
• Publication Date: 2024-05-17T13:23:31Z
• Authors: Mo Dong; Haiye Yu; Zhipeng Sun; Lei Zhang; Yuanyuan Sui; Ruohan Zhao
• Keywords: Edge device

Abstract With the continuous advancement of agricultural Internet Things (IoT) technologies, real-time monitoring environments has become increasingly significant. This provides valuable information on pest and disease occurrences corresponding ecological conditions. However, as have extensive coverage, number IoT devices volume will rapidly increase. leads to a surge in network traffic computing demands. To address this issue, article proposes an environmental system that utilizes edge deep learning technologies. It combines Long-Range Wide Area Network (LoRaWAN) for long-range transmission with recognition counting modules. By offloading workloads traditionally processed cloud nodes, proposed effectively reduces pressures IoT. Simulation experiments demonstrate stable LoRaWAN protocol-based data at edge, overall packet loss rate less than 5%, meeting quality requirements. Moreover, investigates method based technology. Pest images, captured by are recognized counted online using TensorFlow framework. Experimental results indicate accuracy 89% recognition. digitally transmitting image cloud, significantly alleviates