A Smart-Alarm Detection and Deterrent System for Long-Tailed Macaque Intrusion in Human Residential Zones.

BACKGROUND: Human-macaque conflict has become an increasing concern in urban and residential areas. Existing deterrent systems, particularly ultrasonic-based approaches relying on fixed frequencies or simple motion detection, have limited effectiveness and high false alarm rates. Moreover, most systems lack integration between detection, risk assessment, and adaptive response mechanisms. OBJECTIVE: To develop a smart-alarm system for detecting and mitigating long-tailed macaque intrusion in human residential environments. METHODS: Computer vision-based object detection was integrated with real-time risk assessment using distance and group size as key parameters. The system dynamically activates deterrent sound signals and sends real-time alerts via communication platforms. The system was deployed and evaluated in a real-world setting (Sriracha Campus, Kasetsart University). RESULTS: The system could effectively detect macaques and provide appropriate, timely deterrence responses, thereby reducing human-macaque conflict. CONCLUSIONS: This novel AI-based integrated framework combines detection, behavior analysis, and risk assessment for nonlethal wildlife management in shared human-animal environments.