Automated Activity Analysis of Pregnant, Pre-partum, and Post-partum Dromedary Camels Using YOLOv8 and SAMURAI

A

Authors: Nabil Mansour, Fawaghy Alhashmi, Hisham Cholakkal, Mostafa Nasef, and Fouad LamghariInstitutions: Fujairah Research Centre (FRC), MBZUAI, and Bulaida Farms (Raibal Group)

LINK : Frontiers | Automated activity Analysis of Pregnant, Pre-partum, and Post-partum Dromedary female Camels Using YOLOv8 and SAMURAI Tracking

Overview

This groundbreaking study demonstrates the first successful use of YOLOv8 and SAMURAI tracking for automated, continuous, and non-invasive behavioral monitoring of dromedary camels across pregnancy, pre-partum, and post-partum stages. Conducted at Bulaida Farm, Fujairah, the project establishes a new era in precision camel farming through AI-driven activity analysis.

Methods

Two experiments were conducted:

• Experiment 1: Short 15-minute video samples from 25 pregnant camels analyzed using YOLOv8 + SAMURAI.

• Experiment 2: Continuous 24-hour monitoring of 12 camels fitted with color-coded neck collars to capture behaviors before and after calving.

Cameras provided 360° coverage with overlapping fields, and SAMURAI ensured cross-camera identity tracking using motion-aware memory and re-identification algorithms.

Key Findings

• Pregnant camels displayed distinct circadian rhythms: more feeding during daylight, resting at night.

• Pre-partum camels showed restlessness—standing longer (≈ 700 min/day) and eating less (≈ 128 min/day).

• Post-partum camels regained normal behavior, increasing sitting, sleeping, and feeding within 24 hours after calving.

• The AI model achieved reliable detection and classification (eating, drinking, standing/walking, sitting, sleeping) across conditions.

Impact

The system accurately identified behavioral markers linked to reproductive phases, enabling:

• Early detection of health or welfare issues

• Non-invasive reproductive management

• Continuous precision monitoring in intensive camel farming systems

Acknowledgments & Funding

Supported by the Ministry of Culture (KSA), King Faisal University, Fujairah Research Centre, and MBZUAI, under Project No. 1031421:“Leveraging Deep Learning and Automated Monitoring Systems for Early Detection of Health and Performance in Camel Farms.”The authors thank H.H. Sheikh Mohammed bin Hamad Al-Sharqi, Crown Prince of Fujairah, for his continuous support.

Significance

This marks a world-first AI application in camel reproductive monitoring—demonstrating how computer vision and deep learning can replace manual observation and wearable sensors with scalable, contactless, and objective behavioral analysis, paving the way for next-generation smart livestock systems in the UAE and beyond.