Siamese Object Tracking for Unmanned Aerial Vehicle: A Review and Comprehensive Analysis (2205.04281v2)

Abstract: Unmanned aerial vehicle (UAV)-based visual object tracking has enabled a wide range of applications and attracted increasing attention in the field of intelligent transportation systems because of its versatility and effectiveness. As an emerging force in the revolutionary trend of deep learning, Siamese networks shine in UAV-based object tracking with their promising balance of accuracy, robustness, and speed. Thanks to the development of embedded processors and the gradual optimization of deep neural networks, Siamese trackers receive extensive research and realize preliminary combinations with UAVs. However, due to the UAV's limited onboard computational resources and the complex real-world circumstances, aerial tracking with Siamese networks still faces severe obstacles in many aspects. To further explore the deployment of Siamese networks in UAV-based tracking, this work presents a comprehensive review of leading-edge Siamese trackers, along with an exhaustive UAV-specific analysis based on the evaluation using a typical UAV onboard processor. Then, the onboard tests are conducted to validate the feasibility and efficacy of representative Siamese trackers in real-world UAV deployment. Furthermore, to better promote the development of the tracking community, this work analyzes the limitations of existing Siamese trackers and conducts additional experiments represented by low-illumination evaluations. In the end, prospects for the development of Siamese tracking for UAV-based intelligent transportation systems are deeply discussed. The unified framework of leading-edge Siamese trackers, i.e., code library, and the results of their experimental evaluations are available at https://github.com/vision4robotics/SiameseTracking4UAV .

• The paper comprehensively reviews Siamese trackers for UAVs by categorizing methods and analyzing performance across six key benchmarks.
• It evaluates 19 models on an NVIDIA Jetson AGX Xavier, demonstrating above-30 FPS performance crucial for real-time UAV tracking.
• The study highlights challenges like occlusion and low illumination and recommends enhancements in network architectures for improved efficiency.

Review of "Siamese Object Tracking for Unmanned Aerial Vehicle: A Review and Comprehensive Analysis"

The paper "Siamese Object Tracking for Unmanned Aerial Vehicle: A Review and Comprehensive Analysis" provides an exhaustive examination of Siamese networks applied to UAV-based visual object tracking, emphasizing their significance within intelligent transportation systems. Drawing attention to the challenges associated with UAVs such as limited onboard resources and complex environments, this paper presents a systematic exploration of Siamese tracking's potential when deployed on these platforms.

Overview of Siamese Networks in UAV Tracking

Siamese networks have consistently demonstrated efficacy in object tracking by balancing accuracy, robustness, and speed. The architecture utilizes two identical branches to extract features from both the template and the search image, which is particularly advantageous in the context of UAV operations where computational efficiency is paramount. The paper categorizes the state-of-the-art Siamese methods into pioneers, feature modeling, target localization, and other innovative approaches, presenting a well-defined taxonomy and offering a clear insight into the structural variations and enhancements across numerous models.

Numerical and Experimental Insights

A critical component of this paper is the performance analysis spanning six authoritative UAV benchmarks—UAV123@10fps, UAV20L, DTB70, UAVDT, VisDrone-SOT2020-test, and UAVTrack112. Through this rigorous evaluation, the paper illustrates the capabilities of 19 different Siamese trackers on an NVIDIA Jetson AGX Xavier, which is a common UAV processor. The standout models, such as SiamAPN and SiamAPN++, demonstrate the balance between computational efficiency and tracking accuracy, addressing real-time requirements effectively with average speeds above 30 FPS—a threshold necessary for UAV implementations.

Challenges and Potential Improvements

Despite significant advancements, the deployment of Siamese networks in UAV tracking continues to face several challenges, notably dealing with low-resolution, occlusion, viewpoint changes, and low-illumination conditions. The paper's analysis of these attributes highlights the inherent difficulties and presents failure cases that underscore the room for enhancement in feature extraction under adverse scenarios.

Practical Onboard Applications and Future Directions

Beyond theoretical analyses, the paper validates Siamese trackers' feasibility in real-world applications via onboard tests, demonstrating their capability in dynamic environments. This work underscores the need for further optimization in backbone architectures and suggests pathways towards augmenting both performance and speed—considerations critical for UAV tracking in intelligent transportation systems.

Furthermore, the paper identifies areas for future research, recommending the exploration of new network structures (such as vision transformers), enhancements in computational resources, and the utilization of unsupervised domain adaptation for training under challenging conditions like low illumination.

Conclusion

The paper serves as a comprehensive resource for researchers and practitioners in UAV tracking, providing substantive insights into the roles Siamese networks can play in improving tracking reliability under constraints typical of UAV environments. The promising merger of Siamese tracking methodologies with UAV applications presents opportunities for expanding the scope of intelligent transportation systems, with continued research focusing on overcoming computational and environmental hurdles.