Affiliation:
1. Department of Informatics and Telecommunications, University of Ioannina, 451 10 Ioannina, Greece
2. Industrial Systems Insitute, Athena RC, 263 31 Patra, Greece
Abstract
Wildlife conservationists have historically depended on manual methods for the identification and tracking of avian species, to monitor population dynamics and discern potential threats. Nonetheless, many of these techniques present inherent challenges and time constraints. With the advancement in computer vision techniques, automated bird detection and recognition have become possible. This study aimed to further advance the task of detecting wild birds using computer vision methods with drone footage, as well as entirely automating the process of detection and tracking. However, detecting objects from drone footage presents a significant challenge, due to the elevated altitudes, as well as the dynamic movement of both the drone and the birds. In this study, we developed and introduce a state-of-the-art model titled ORACLE (optimized rigorous advanced cutting-edge model for leveraging protection to ecosystems). ORACLE aims to facilitate robust communication across multiple models, with the goal of data retrieval, rigorously using various computer vision techniques such as object detection and multi-object tracking (MOT). The results of ORACLE’s vision models were evaluated at 91.89% mAP at 50% IoU.
Reference44 articles.
1. Paul, P.K., Choudhury, A., Biswas, A., and Singh, B.K. (2022). Drone Applications in Wildlife Research—A Synoptic Review. Environmental Informatics: Challenges and Solutions, Springer Nature Singapore.
2. Possibility of applying unmanned aerial vehicle (UAV) and mapping software for the monitoring of waterbirds and their habitats;Han;J. Ecol. Environ.,2017
3. Environmental variations in a semi-enclosed embayment (Amvrakikos Gulf, Greece)–reconstructions based on benthic foraminifera abundance and lipid biomarker pattern;Naeher;Biogeosciences,2012
4. Building of an edge enabled drone network ecosystem for bird species identification;Das;Ecol. Inform.,2021
5. The breeding biology of the Dalmatian Pelican Pelecanus crispus;Crivelli;Ibis,1998