Development of AIoT System for facility asparagus cultivation

Abstract

Asparagus is a high economic value crop sensitive to environmental factors and pest outbreaks, so asparagus cultivation in Taiwan is usually grown in greenhouses. This study established an AIoT system specifically for asparagus growth and monitoring of pests and diseases, providing planting guidelines for farmers to grow asparagus and preventing outbreaks of pests and diseases. We have installed many IoT and AI sensors to track environmental changes and pest populations. In the summer, the fans are automatically turned on and off according to the temperature measured by the sensor. The cooling effect is observed on the temperature distribution interpolated with kriging. By correlation analysis, the relationship between environmental factors and asparagus yield and the relationship between environmental changes and pests were obtained, which can help to optimize the planting environment in terms of sunlight and temperature regulation, drip fertigation, irrigation, and pest control strategies. The deep learning model we developed can detect pests and count pest numbers with high accuracy. The model detection results show that with an IOU threshold of 0.5, precision, recall, mAP, and pest count accuracy reach 93.8%, 91.9%, 95.3%, and 95.8%, respectively. Finally, our research and development results can provide farmers with automated and optimized asparagus planting methods, reduce insect pests, optimize the growing environment, and provide good irrigation and soil management to improve yield and quality.