Development of an autonomous smart trap for precision monitoring of hematophagous flies on cattle

• Hematophagous flies threaten livestock health, impacting productivity and welfare • Smart traps can enable real-time monitoring of hematophagous flies on cattle • Autonomy, flexibility, and environmental monitoring with integrated sensors • AI-powered system achieves high-precision fly recognition • Scalable, accurate, and sustainable approach for pest control in livestock Hematophagous flies pose a significant threat to livestock health and productivity. Traditional pest control methods, which heavily rely on chemical insecticides, present risks such as resistance development and environmental harm. This study presents a novel smart trap for real-time monitoring and identification of Tabanidae, Hippoboscidae (i.e., Hippobosca equina L.), and Muscidae (i.e., Stomoxys calcitrans (L.)) on cattle. The system includes a microcontroller for managing various sensors (light, temperature, humidity, and gas) and power management. The control unit is complemented by a microprocessor responsible for managing and processing images from a camera. The system integrates high-resolution imaging, a convolutional neural network (CNN) for species recognition, and environmental sensors to monitor factors affecting insect behavior. On the test set, the CNN achieved an overall precision of 0.96 and recall of 0.98 in detecting instances, with an overall classification accuracy of 0.96. Equipped, also, by lithium-ion battery and by communication module, the trap can operate autonomously and transmit data, becoming suitable for large-scale deployments. Overall, the tool developed here offers a practical and cheap solution for sustainable and accurate pest monitoring of hematophagous flies attacking cattle in pasture and feedlot.