Breaking Smooth-Motion Assumptions: A UAV Benchmark for Multi-Object Tracking in Complex and Adverse Conditions

2026-03-06

AI Summary

Existing UAV multi-object tracking benchmarks predominantly rely on the assumption of smooth motion, failing to capture real-world challenges such as abrupt scale changes, drastic viewpoint shifts, and motion blur induced by aggressive ego-motion. To address this limitation, this work proposes DynUAV, a new benchmark comprising 42 highly dynamic video sequences with over 1.7 million meticulously annotated bounding boxes across categories including vehicles, pedestrians, and construction machinery. DynUAV is the first to systematically model the compound disturbances arising under intense ego-motion. Through large-scale data collection and comprehensive evaluation of state-of-the-art trackers, the benchmark reveals the coupled failure modes of current methods in dynamic scenarios. DynUAV thus provides a challenging and realistic testbed for advancing the robustness of tracking algorithms.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: Motion and Path Planning
• Application Domains: Mobility, Driving & Flight

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

The rapid movements and agile maneuvers of unmanned aerial vehicles (UAVs) induce significant observational challenges for multi-object tracking (MOT). However, existing UAV-perspective MOT benchmarks often lack these complexities, featuring predominantly predictable camera dynamics and linear motion patterns. To address this gap, we introduce DynUAV, a new benchmark for dynamic UAV-perspective MOT, characterized by intense ego-motion and the resulting complex apparent trajectories. The benchmark comprises 42 video sequences with over 1.7 million bounding box annotations, covering vehicles, pedestrians, and specialized industrial categories such as excavators, bulldozers and cranes. Compared to existing benchmarks, DynUAV introduces substantial challenges arising from ego-motion, including drastic scale changes and viewpoint changes, as well as motion blur. Comprehensive evaluations of state-of-the-art trackers on DynUAV reveal their limitations, particularly in managing the intertwined challenges of detection and association under such dynamic conditions, thereby establishing DynUAV as a rigorous benchmark. We anticipate that DynUAV will serve as a demanding testbed to spur progress in real-world UAV-perspective MOT, and we will make all resources available at link.

Problem

• multi-object tracking
• UAV
• ego-motion
• benchmark
• complex trajectories

Innovation

• dynamic UAV
• multi-object tracking
• ego-motion
• motion blur
• benchmark