AI Tracks Mithun Behaviour at ICAR Farm in Nagaland
AI-powered system could support animal health, breeding and round-the-clock livestock monitoring
GG News Bureau
Nagaland, 8th Sept: Researchers at the ICAR-National Research Centre on Mithun (ICAR-NRC on Mithun), Nagaland, have developed an artificial intelligence-based system for real-time, non-contact detection and tracking of Mithun behaviour in a natural farm environment.
Mithun, popularly known as the “Cattle of the Hills”, has significant social, cultural and economic importance for tribal communities across Northeast India. The new technology could help provide continuous behavioural information for animal health, welfare, nutrition, breeding and reproductive management.
The researchers deployed 12 high-definition CCTV cameras across two sheds at the ICAR-NRC on Mithun farm, enabling continuous day-and-night monitoring, including infrared surveillance. A dataset of 3,000 manually annotated images was created covering four behaviours—feeding, standing, lying and mounting.
The AI system combines the YOLOv8n model for behaviour detection with DeepSORT technology to track individual animals and maintain their identities across video frames.
The YOLOv8n model achieved a mean average precision of 99.5% at mAP@0.5, with a recall of 99.6%. The system processed footage at approximately 31 frames per second using an NVIDIA RTX 3060 GPU, demonstrating real-time monitoring capability.
The system was also tested under challenging farm conditions, including partial occlusion, background clutter, uneven and wet ground, shadows, motion blur and nighttime infrared footage.
Researchers said changes in feeding, standing and lying patterns could provide useful information about an animal’s health, comfort and physiological condition, while mounting behaviour could assist reproductive and oestrus management. Automated monitoring could reduce the need for continuous manual observation, including during night hours.
The researchers cautioned that the system has so far been evaluated at a single farm and requires further validation across different farms, geographical regions, seasons, stocking densities and camera arrangements. Severe occlusion can also affect detection and tracking performance.
Future research will explore additional behaviours such as aggression, grooming and disease-related inactivity, along with temporal AI models, edge-device deployment and larger datasets covering diverse farms and seasons.
The study, published in Engineering Research Express, Volume 8 (2026), Article 175213, was conducted by researchers from ICAR-NRC on Mithun in collaboration with NIT Nagaland, Nagaland University and CHRIST (Deemed to be University).
