Abstract:Unmanned aerial vehicle (UAV) tracking is crucial for autonomous navigation and has broad applications in robotic automation fields. However, reliable UAV tracking remains a challenging task due to various difficulties like frequent occlusion and aspect ratio change. Additionally, most of the existing work mainly focuses on explicit information to improve tracking performance, ignoring potential interframe connections. To address the above issues, this work proposes a novel framework with continuity-aware latent interframe information mining for reliable UAV tracking, i.e., ClimRT. Specifically, a new efficient continuity-aware latent interframe information mining network (ClimNet) is proposed for UAV tracking, which can generate highly-effective latent frame between two adjacent frames. Besides, a novel location-continuity Transformer (LCT) is designed to fully explore continuity-aware spatial-temporal information, thereby markedly enhancing UAV tracking. Extensive qualitative and quantitative experiments on three authoritative aerial benchmarks strongly validate the robustness and reliability of ClimRT in UAV tracking performance. Furthermore, real-world tests on the aerial platform validate its practicability and effectiveness. The code and demo materials are released at <a class="link-external link-https" href="https://github.com/vision4robotics/ClimRT" rel="external noopener nofollow">this https URL</a>.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the tracking reliability issue in unmanned aerial vehicle (UAV) tracking caused by occlusion, view - angle change, etc. Specifically, the existing UAV tracking methods mainly rely on explicit information to improve tracking performance while ignoring the potential connection information between frames. This leads to poor tracking performance in complex scenarios (such as frequent occlusion and view - angle change). To solve these problems, the paper proposes a new framework - Continuous - aware Latent Inter - frame Information Mining (ClimRT) to achieve reliable UAV tracking. The main contributions of the paper include: 1. Proposing an efficient Continuous - aware Latent Inter - frame Information Mining Network (ClimNet) suitable for UAV tracking, which can provide sufficient and reliable target information and enhance the robustness of tracking. 2. Introducing two convolutional blocks to improve the feature extraction and fusion capabilities of continuous - aware latent inter - frame information. 3. Designing a new Location Continuous Transformer (LCT) for the integration of spatio - temporal information and the enhancement of continuity, thereby improving the accuracy of UAV tracking. 4. Conducting a comprehensive evaluation on three authoritative aerial benchmarks to verify the superior performance of ClimRT compared to the existing state - of - the - art (SOTA) trackers. Practical tests also prove the effectiveness and practicality of ClimRT in the real world. Through these innovations, the paper aims to mine the potential information between frames, introduce sufficient spatio - temporal information, enhance the continuity of information, thereby smoothing the mutation process when encountering common target shape deformation or occlusion, extracting more reliable information, and enhancing the robustness of tracking.

Continuity-Aware Latent Interframe Information Mining for Reliable UAV Tracking

Flow-Guided Single Object Tracking Framework in UAV Aerial Video

Separation Fusion Transformer and Efficient Reuse Matching Network for Aerial Tracking

Interframe Saliency Transformer and Lightweight Multidimensional Attention Network for Real-Time Unmanned Aerial Vehicle Tracking

Precision in pursuit: a multi-consistency joint approach for infrared anti-UAV tracking

Tracker Meets Night: A Transformer Enhancer for UAV Tracking

UAV-Ground Visual Tracking: A Unified Dataset and Collaborative Learning Approach

Learning Motion Blur Robust Vision Transformers with Dynamic Early Exit for Real-Time UAV Tracking

Object Tracking in Unmanned Aerial Vehicle Videos via Multifeature Discrimination and Instance-Aware Attention Network

Progressive Representation Learning for Real-Time UAV Tracking

An Occlusion-Aware Tracker With Local-Global Features Modeling in UAV Videos

Robust Scalable Part-Based Visual Tracking for UAV with Background-Aware Correlation Filter.

SGDViT: Saliency-Guided Dynamic Vision Transformer for UAV Tracking

Anti-Occlusion UAV Tracking Algorithm with a Low-Altitude Complex Background by Integrating Attention Mechanism

Part-Based Background-Aware Tracking for UAV with Convolutional Features

AutoTrack: Towards High-Performance Visual Tracking for UAV With Automatic Spatio-Temporal Regularization

PVT++: A Simple End-to-End Latency-Aware Visual Tracking Framework

All-Day Object Tracking for Unmanned Aerial Vehicle

WebUAV-3M: A Benchmark for Unveiling the Power of Million-Scale Deep UAV Tracking

TrackingMamba: Visual State Space Model for Object Tracking

A Unified Transformer-based Tracker for Anti-UAV Tracking