MMLA: Multi-Environment, Multi-Species, Low-Altitude Aerial Footage Dataset

2025-04-10

AI Summary

To address the limited generalization capability of real-time wildlife detection in low-altitude UAV imagery, this paper introduces MMLA—the first cross-environment, cross-species aerial wildlife dataset for low-altitude UAVs—spanning two Kenyan protected areas and a conservation center in Ohio, USA, and featuring five species: plains zebra, Grevy’s zebra, giraffe, African wild ass, and African wild dog. We conduct multi-domain benchmarking using YOLOv5m, YOLOv8m, and YOLOv11m, underpinned by an evaluation framework integrating field data collection, fine-grained annotation, and cross-domain analysis. Experiments reveal substantial geographic and species-specific performance degradation, demonstrating constrained generalization of existing methods in realistic野外 settings. This work bridges a critical gap in standardized, generalizable evaluation for low-altitude UAV-based wildlife detection, establishing a foundational benchmark for domain adaptation and few-shot learning research.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Application Domains: Mobility, Driving & Flight
• Computer Vision: Object Detection & Categorization

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Real-time wildlife detection in drone imagery is critical for numerous applications, including animal ecology, conservation, and biodiversity monitoring. Low-altitude drone missions are effective for collecting fine-grained animal movement and behavior data, particularly if missions are automated for increased speed and consistency. However, little work exists on evaluating computer vision models on low-altitude aerial imagery and generalizability across different species and settings. To fill this gap, we present a novel multi-environment, multi-species, low-altitude aerial footage (MMLA) dataset. MMLA consists of drone footage collected across three diverse environments: Ol Pejeta Conservancy and Mpala Research Centre in Kenya, and The Wilds Conservation Center in Ohio, which includes five species: Plains zebras, Grevy's zebras, giraffes, onagers, and African Painted Dogs. We comprehensively evaluate three YOLO models (YOLOv5m, YOLOv8m, and YOLOv11m) for detecting animals. Results demonstrate significant performance disparities across locations and species-specific detection variations. Our work highlights the importance of evaluating detection algorithms across different environments for robust wildlife monitoring applications using drones.

Problem

• Evaluating wildlife detection in low-altitude drone imagery
• Assessing model generalizability across species and environments
• Addressing performance disparities in animal detection algorithms

Innovation

• Multi-environment, multi-species low-altitude drone dataset
• Evaluates YOLO models for wildlife detection
• Highlights performance disparities across locations and species