Detector-Augmented SAMURAI for Long-Duration Drone Tracking

TL;DR

This paper evaluates and enhances the SAMURAI foundation model for long-term drone tracking in urban surveillance, introducing a detector-augmented extension that improves robustness and accuracy in complex, long-duration scenarios.

Contribution

It is the first to systematically assess SAMURAI for drone tracking and proposes a detector-augmented extension to improve its robustness in urban environments.

Findings

Detector-augmented SAMURAI significantly improves robustness in urban drone tracking.

The extension yields up to +0.393 success rate improvement and -0.475 FNR reduction.

Enhanced performance is especially notable in long-duration sequences with drone re-entry events.

Abstract

Robust long-term tracking of drone is a critical requirement for modern surveillance systems, given their increasing threat potential. While detector-based approaches typically achieve strong frame-level accuracy, they often suffer from temporal inconsistencies caused by frequent detection dropouts. Despite its practical relevance, research on RGB-based drone tracking is still limited and largely reliant on conventional motion models. Meanwhile, foundation models like SAMURAI have established their effectiveness across other domains, exhibiting strong category-agnostic tracking performance. However, their applicability in drone-specific scenarios has not been investigated yet. Motivated by this gap, we present the first systematic evaluation of SAMURAI's potential for robust drone tracking in urban surveillance settings. Furthermore, we introduce a detector-augmented extension of SAMURAI to mitigate sensitivity to bounding-box initialization and sequence length. Our findings demonstrate that the proposed extension significantly improves robustness in complex urban environments, with pronounced benefits in long-duration sequences - especially under drone exit-re-entry events. The incorporation of detector cues yields consistent gains over SAMURAI's zero-shot performance across datasets and metrics, with success rate improvements of up to +0.393 and FNR reductions of up to -0.475.