Benchmarking Deep Trackers on Aerial Videos

TL;DR

This paper evaluates ten deep learning-based visual object trackers on aerial videos, revealing significant performance challenges due to unique aerial tracking conditions like small targets and camera motion.

Contribution

It provides a comprehensive benchmark of deep trackers on aerial datasets, highlighting their strengths and weaknesses in aerial tracking scenarios.

Findings

Trackers perform worse on aerial datasets than ground-level videos.

Aerial tracking challenges include small target size and camera motion.

Different tracking approaches have varying effectiveness in aerial conditions.

Abstract

In recent years, deep learning-based visual object trackers have achieved state-of-the-art performance on several visual object tracking benchmarks. However, most tracking benchmarks are focused on ground level videos, whereas aerial tracking presents a new set of challenges. In this paper, we compare ten trackers based on deep learning techniques on four aerial datasets. We choose top performing trackers utilizing different approaches, specifically tracking by detection, discriminative correlation filters, Siamese networks and reinforcement learning. In our experiments, we use a subset of OTB2015 dataset with aerial style videos; the UAV123 dataset without synthetic sequences; the UAV20L dataset, which contains 20 long sequences; and DTB70 dataset as our benchmark datasets. We compare the advantages and disadvantages of different trackers in different tracking situations encountered in aerial data. Our findings indicate that the trackers perform significantly worse in aerial datasets compared to standard ground level videos. We attribute this effect to smaller target size, camera motion, significant camera rotation with respect to the target, out of view movement, and clutter in the form of occlusions or similar looking distractors near tracked object.