Work

Conservation in action: Cost-effective UAVs and real-time detection of the globally threatened swamp deer (Rucervus duvaucelii)

Year: 2024

Type: article

Abstract: UAVs have emerged as an effective tool for ecological monitoring due to their high visual resolution and ability to cover large areas with spatial precision. Artificial Intelligence has further improved the capabilities of UAV vision through object detection. While deep learning has shown significant success in pattern recognition, it still faces challenges in real-world scenarios. In this study, we focused on enhancing the potential of Unmanned Aerial Vehicles (UAVs) for wildlife detection and monitoring, specifically focusing on the globally threatened swamp deer (Rucervus duvaucelii). To improve the accuracy of animal recognition with UAVs, we used one stage detector YOLO (You Only Look Once), DETR and Object detection v3 to train a model using more than 50 thousands labeled swamp deer. We have also developed a real-time object detection framework that seamlessly integrates with front-end technologies, enabling live detection results on a website. This framework operates on a local server, synchronizing with consumer-grade UAVs at a rate of 32 frame per second (fps) using a frame sampling technique, notably without requiring a dedicated Graphical Processing Unit (GPU). Our study also highlights the superior performance of Object detection 3 and YOLO v8, achieving a precision score more than 92% compared to DETR (DEtection TRansformer), Yolo v7, v5 and v3. Overall, our study provides an efficient, cost-effective, and accurate detection framework for ecological monitoring that can be non-invasively used in various demographic studies of cervids in different region and habitat types. This deliberate choice of not using GPU not only underscores the commitment to cost-effectiveness but also aligns with the research's purpose, prioritizing accessibility and affordability for wider scientific exploration. This methodology can be replicated and fine-tuned for studying other threatened species of conservation priority.

Source: Ecological Informatics

DOI: 10.1016/j.ecoinf.2024.102913

Authors: Ravindra Nath Tripathi, Karan Agarwal, Vikas Tripathi, Ruchi Badola, Syed Ainul Hussain

Institutions: Wildlife Institute of India, Graphic Era University

Language: English

Topic: Advanced Neural Network Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

SDG: Life on land

Open Access status: gold

Estimated APC paid: $2,200