We are pleased to share that our paper, “Stay Local or Go Global: Geo-Referenced Bounding Boxes for Tracking Wildlife in Thermal Drone Videos,” has been accepted and published in IET Computer Vision.
In this work, we investigate how geo-referenced information can improve wildlife tracking in thermal drone videos. By combining animal detections with drone position, camera orientation, and elevation data, our approach tracks animals not only in image coordinates but also in real-world geographic space.
Evaluated on 225 drone videos, the proposed methods reduced identity switches compared with existing tracking approaches and enable trajectories that can be directly used for spatial analysis and wildlife monitoring.
The paper is available open access here:
https://ietresearch.onlinelibrary.wiley.com/doi/full/10.1049/cvi2.70077