Research and Application of an Intelligent Monitoring System for Levee Termite Activity Based on Multi-Optical-Path Infrared Triggering and YOLOv11

Monitoring concealed termite activity in river levees remains challenging, particularly when unattended operation and verifiable visual records are required. This study developed a termite baiting station (TBS) and a multi-optical-path infrared triggering (MIRT) method, integrated with image acquisition, wireless communication and platform-based object detection using YOLOv11. Bait guides termites into a monitoring chamber; changes in the received infrared signal that meet predefined criteria initiate image acquisition and transmission, while the camera and communication module remain in low-power states between events. Uploaded images support automated detection and subsequent human verification. From January to June 2026, 268 stations were evaluated along an ~1.5 km section of the Beijiang Levee, China. The average online rate and image upload success rate were 96% and 98%, respectively. All 42 stations with alarms were inspected on site, and termite activity was confirmed at 38 (positive predictive value 90.5%); daily images from the 226 stations without alarms, which were not inspected on site, showed no visible termite activity. These findings support the feasibility of combining event-triggered sensing, remote image analysis and field verification for levee termite monitoring. Further evaluation is needed to quantify missed detections, energy consumption, and performance over longer deployment periods.