Design and Implementation of a Real-Time Visual Tracking System for UAVs Based on PSDK
Sensors · 31.03.2026 · 26 (7) · 2145
DOI: 10.3390/s26072145 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 31.03.2026. Verfasst von Ranjun Yang und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Guangxi Science and Technology Department, Education Department of Guangxi Zhuang Autonomous Region, Joint International Research Laboratory of Spatio-temporal Information and Intelligent Location Services.
Abstract (Originalsprache)
This paper presents the design and implementation of a real-time visual tracking system for unmanned aerial vehicles (UAVs), based on the DJIPayload Software Development Kit (PSDK), addressing the challenge of balancing high precision with low latency on resource-constrained edge platforms. By utilizing DJI PSDK to abandon the Robot Operating System (ROS) layer and its associated serialization overhead, the proposed Middleware-Free Architecture reduces end-to-end latency by over 60% to approximately 30 ms. To address computational constraints, a Lightweight Asymmetric De-coupled Visual Servoing (ADVS) strategy is proposed. It adopts orthogonal kinematic de-coupling to bypass Jacobian matrix inversion and integrates a non-linear dead-zone mechanism with dynamics-aware gain scheduling to compensate for sensing anisotropy and gravitational nonlinearity. Simultaneously, a Geometry-Aware Fusion strategy is employed to reject visual outliers, while a Finite State Machine (FSM) strictly enforces temporal consistency. Field experiments in various scenarios verify the system’s stability and tracking capability. Specifically, the platform maintains a robust lock on targets at speeds up to 23 m/s across dynamic maneuvers. The successful implementation of this system confirms that high-performance edge tracking does not rely solely on the scaling of visual model complexity but can also be effectively achieved through the architectural minimization of latency combined with the optimization of theoretically grounded robust control strategies.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China.
- Joint International Research Laboratory of Spatio-Temporal Information and Intelligent Location Services, Guilin University of Electronic Technology, Guilin 541004, China.
- Center for Advanced Industrial Technology, Universiti Malaysia Pahang Al Sultan Abdullah, Pekan 26600, Malaysia.
ORCID-Profile: Ningbo Xie, Qinlin Li, Kefei Liao
Förderung
- Guangxi Science and Technology Department
- Education Department of Guangxi Zhuang Autonomous Region
- Joint International Research Laboratory of Spatio-temporal Information and Intelligent Location Services
- Guangxi Science and Technology Department Project
- Guangxi University Young and Middle-aged Teachers’ Basic Scientific Research Ability Improvement Project
Zitieren
APA
Yang, R., Xie, N., Li, Q., Liao, K., Lang, J., & Ghazali, K. H. B. (2026). Design and Implementation of a Real-Time Visual Tracking System for UAVs Based on PSDK. Sensors, 26(7), 2145. https://doi.org/10.3390/s26072145
Vancouver
Yang R, Xie N, Li Q, Liao K, Lang J, Ghazali KHB. Design and Implementation of a Real-Time Visual Tracking System for UAVs Based on PSDK. Sensors. 2026;26(7):2145. doi:10.3390/s26072145
BibTeX
@article{Yang2026,
author = {Yang, Ranjun and Xie, Ningbo and Li, Qinlin and Liao, Kefei and Lang, Jie and Ghazali, Kamarul Hawari Bin},
title = {Design and Implementation of a Real-Time Visual Tracking System for UAVs Based on PSDK},
journal = {Sensors},
year = {2026},
volume = {26},
number = {7},
pages = {2145},
doi = {10.3390/s26072145}
}
EUVIMED