Precision Pest Monitoring in Asparagus Greenhouses using Deep Learning and IoT
Journal
2025 Asabe Annual International Meeting
Part Of
2025 ASABE Annual International Meeting
Date Issued
2025
Author(s)
Chen, Po-Shao
Lian, Zhen-Yu
Hsieh, Ming-Hsien
Peng, Jui-Chu
Guo, Ming-Chi
Abstract
With its small land area, Taiwan relies heavily on precision agriculture to maximize yield. The cultivation of asparagus faces significant challenges from insect pests, particularly whiteflies and thrips. This research focuses on implementing a real-time pest detection system using deep learning and IoT devices in asparagus greenhouses to increase production and reduce labor. We developed an IoT system with an integrated website featuring a concise user interface. The platform provides pest detection, weather forecasts, yield predictions, and essential environmental data including soil conditions and greenhouse parameters. Our pest monitoring system uses low-power devices that achieve good results with lower-resolution images and can be powered by solar energy, facilitating field deployment for monitoring pests on sticky traps. In two Tainan greenhouses, we installed IoT devices to monitor environmental conditions. For pest identification, images were divided into sixteen sub-images, creating a dataset of 800 images for training YOLO models. Our model achieved an mAP50 of 0.952, ensuring reliable pest population monitoring. The image processing approach maximizes detection accuracy for small insect targets despite resolution limitations. Our energy-efficient pest monitoring device reduces power consumption by 95% compared to conventional systems, enabling long-term operation with minimal solar charging. This platform aids farmers by providing real-time data and recommendations, improving asparagus quality and yield in Taiwan through comprehensive precision agriculture solutions.
Event(s)
2025 American Society of Agricultural and Biological Engineers Annual International Meeting, ASABE 2025
Subjects
asparagus
deep learning
IOT
object detection
pests
Publisher
American Society of Agricultural and Biological Engineers
Type
conference paper