|
|
|
You can get e-magazine links on WhatsApp. Click here
|
|
|
|
|
|
ICAR-NRC on Mithun develops AI system for real-time behaviour tracking
|
|
Friday, 11 September, 2026, 08 : 00 AM [IST]
|
|
Our Bureau, Mumbai
|
Researchers at the ICAR-National Research Centre on Mithun (ICAR-NRC on Mithun), Nagaland, have developed an artificial intelligence (AI)-based system for real-time detection and tracking of Mithun behaviour in a natural farm environment.
The technology is designed to enable continuous, non-contact monitoring of animals and could support livestock health, welfare, breeding and reproductive management. The study has been published in Engineering Research Express.
Mithun (Bos frontalis), popularly known as the “Cattle of the Hills”, holds significant social, cultural and economic importance for tribal communities across Northeast India and plays an important role in regional livelihoods and food security.
The AI framework automatically detects and tracks four key behaviours—feeding, standing, lying and mounting. Researchers noted that monitoring such behavioural patterns can provide early indications of changes in an animal’s health, comfort, nutrition and physiological condition. Mounting behaviour can also provide useful information for reproductive and oestrus management.
To develop the system, researchers installed 12 high-definition CCTV cameras across two sheds at the ICAR-NRC on Mithun farm in Nagaland. The cameras enabled continuous day-and-night surveillance, including infrared monitoring during low-light conditions. The team subsequently created a dataset of 3,000 manually annotated images covering the four target behaviours.
The system combines the YOLOv8n object-detection model with DeepSORT tracking technology. While YOLOv8n identifies the behaviour being performed, DeepSORT tracks individual animals across video frames and assigns persistent identities, enabling real-time behavioural monitoring of individual Mithun.
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 its potential for real-time deployment.
The researchers also evaluated the framework under challenging farm conditions, including partial occlusion, background clutter, uneven and wet ground, shadows, motion blur and nighttime infrared footage.
According to the researchers, automated behavioural monitoring could reduce the dependence on continuous manual observation, which is labour-intensive and difficult to maintain around the clock. Changes in feeding, standing and lying patterns could help livestock managers identify potential health or welfare concerns, while automated detection of mounting behaviour could assist reproductive management.
However, the researchers said further validation is required before wider deployment. The current system has been evaluated at a single farm and needs to be tested across different farms, geographical regions, seasons, stocking densities and camera configurations. The study is also limited to four behaviours, while severe occlusion can affect detection and tracking performance.
Future research will focus on expanding the system to detect behaviours such as aggression, grooming and disease-related inactivity. The researchers are also exploring temporal AI models, edge-device deployment and larger datasets representing different farms and seasonal conditions.
The study highlights the potential of combining artificial intelligence, computer vision and livestock science to advance precision livestock farming. The technology could provide farmers and livestock managers with continuous, data-driven insights into animal behaviour while strengthening approaches to health, welfare and reproductive management.
The research was conducted by scientists from ICAR-NRC on Mithun, Nagaland, in collaboration with NIT Nagaland, Nagaland University and CHRIST (Deemed to be University).
|
|
|
|
|
|
|
|
|
|