Sea You Later: Metadata-Guided Long-Term Re-Identification for UAV-Based Multi-Object Tracking

Re-identification (ReID) in multi-object tracking (MOT) for UAVs in maritime computer vision has been challenging for several reasons. More specifically, short-term re-identification (ReID) is difficult due to the nature of the characteristics of small targets and the sudden movement of the drone's gimbal. Long-term ReID suffers from the lack of useful appearance diversity. In response to these chal-lenges, we present an adaptable motion-based MOT algorithm, called Metadata Guided MOT (MG-MOT). This al-gorithm effectively merges short-term tracking data into co-herent long-term tracks, harnessing crucial metadata from UAVs, including GPS position, drone altitude, and camera orientations. Extensive experiments are conducted to vali-date the efficacy of our MOT algorithm. Utilizing the chal-lenging SeaDroneSee tracking dataset, which encompasses the aforementioned scenarios, we achieve a much-improved performance in the latest edition of the UAV-based Maritime Object Tracking Challenge with a state-of-the-art HOTA of 69.5% and an IDFI of85.9% on the testing split.