Researchers at the ICAR-National Research Center on Mithun (ICAR-NRC on Mithun), Nagaland, developed an artificial intelligence-based system capable of detecting and tracking Mithun behavior in real time under natural farm conditions.
The technology could help monitor the animals around the clock and provide useful information for their health, welfare, breeding and reproductive management.
The study, published in Engineering Research Express, presents a real-time, non-contact AI framework for automatically detecting and tracking the behavior of Mithun (Bos frontalis), popularly known as the “Cattle of the Hills”. The animal has significant social, cultural and economic importance for tribal communities across Northeast India and plays an important role in livelihoods and food security.
Monitoring livestock behavior is crucial for effective animal management, as changes in feeding, standing, lying and reproductive activities can provide early indications of an animal’s health, comfort, nutrition and physiological condition. Conventional monitoring, however, largely relies on manual observation, making continuous surveillance difficult, particularly at night.
To overcome this limitation, researchers installed 12 high-definition CCTV cameras across two sheds at the ICAR-NRC on Mithun farm in Nagaland. The cameras provided continuous day-and-night surveillance, including infrared coverage.
Using the footage, the researchers created a dataset of 3,000 manually annotated images covering four key behaviours: feeding, standing, lying and mounting.
The AI framework combines the YOLOv8n model for behavior detection with DeepSORT technology to track individual animals across video frames and maintain their identities. This allows the system to identify what an animal is doing while simultaneously tracking individual Mithun in real time.
The YOLOv8n model achieved a mean average precision of 99.5% at mAP@0.5 and a recall of 99.6%. The system processed around 31 frames per second using an NVIDIA RTX 3060 graphics processing unit, demonstrating its suitability for real-time applications.
Researchers also tested the technology under challenging farm conditions, including partial obstruction of animals, background clutter, uneven and wet ground, shadows, motion blur and nighttime infrared footage.
Continuous automated monitoring could help livestock managers identify changes in feeding, standing and lying patterns associated with health, comfort and physiological conditions. Detection of mounting behavior could also assist in reproductive and oestrous management.
The researchers, however, noted that the system has so far been evaluated at a single farm and requires validation across different locations, seasons, stocking densities and camera arrangements. Severe occlusion can also affect detection and tracking performance.
Future research could expand the system to detect behaviors such as aggression, grooming and disease-related inactivity. Researchers may also explore temporal AI models, deployment on edge devices and larger datasets covering different farms and seasons.
The study highlights the growing potential of combining artificial intelligence, computer vision and livestock science to strengthen precision livestock farming through continuous, data-driven monitoring.
Published in Engineering Research Express, Volume 8 (2026), Article 175213, the study was conducted by researchers from ICAR-NRC on Mithun in collaboration with NIT Nagaland, Nagaland University and CHRIST (Deemed to be University).
