Keywords
Real-time object detection
Deep learning
Lightweight model
Motion compensation
Submitted : 2025-05-07
Accepted : 2025-05-22
Published : 2025-06-06
Abstract
To address the challenges of low accuracy and insufficient real-time performance in dynamic object detection for UAV surveillance, this paper proposes a novel tracking framework that integrates a lightweight improved YOLOv5s model with adaptive motion compensation. A UAV-view dynamic feature enhancement strategy is innovatively introduced, and a lightweight detection network combining attention mechanisms and multi-scale fusion is constructed. The robustness of tracking under motion blur scenarios is also optimized. Experimental results demonstrate that the proposed method achieves a mAP@0.5 of 68.2% on the VisDrone dataset and reaches an inference speed of 32 FPS on the NVIDIA Jetson TX2 platform. This significantly improves the balance between accuracy and efficiency in complex scenes, offering reliable technical support for real-time applications such as emergency response.
References
Liu C, Li S, Meng F, et al., 2023, A Review of UAV Object Detection Based on Deep Learning. Computer Science and Applications, 13(5): 1092–1099.
Liu G, Zeng X, Dou J, et al., 2024, Review of Advances in Small Object Detection Technology based on Deep Learning. Infrared and Laser Engineering, 53(9): 184–216.
Li X, Zhou Y, Wang F, 2022, Advanced Information Mining from Ocean Remote Sensing Imagery with Deep Learning. Journal of Remote Sensing, 9849645.
Chinese Academy of Sciences, Literature and Information Center, 2024, Scientific Literature Data Mining to Enhance Research Efficiency and Large Model Training, Shijiazhuang Data Bureau, viewed October 10, 2024. https://sjj.sjz.gov.cn/columns/9bcbfa10-a61f-4005-b732-234ea94403d7/202410/10/6fb6230c-96a8-40f1-801d-9c77c4200f61.html
Girshick R, Donahue J, Darrell T, et al., 2014, Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA, 580–587.
Redmon J, Divvala S, Girshick R, et al., 2016, You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016, 779–788.
Bochkovskiy A, Wang CY, Liao HYM, 2020, YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv preprint. https://doi.org/10.48550/arXiv.2004.10934
Xie L, Xue Y, Ye J, 2022, UAV Aerial Target Detection Algorithm Based on Improved YOLOv5. Acta Automatica Sinica, 2284(1): 012024.
Peng M, Liu J, Hu Z, 2023, Research on UAV Target Detection Based on YOLOv5. Computer Engineering and Applications, 59(5): 157–162.
Zheng H, Zhang J, Hu X, et al., 2022, UAV Ground Target Detection Method Based on Improved YOLOv4. Ordnance Automation, 41(7): 89–94.
Wang L, Zou Y, 2022, Overview of Object Detection Algorithms Based on Deep Learning. Computer Science and Exploration, 16(9): 1521–1531.
Jiang W, Wang M, 2022, Research on Ground Target Detection and Tracking Based on YOLOv5. Modern Electronic Technology, 45(23): 123–127.
Sun K, Li Z, Qi L, 2021, Design of UAV Image Recognition System Based on Embedded Platform. Electronic Technology, 34(4): 56–60.
Ge Z, Liu S, Wang F, et al., 2021, YOLOX: Exceeding YOLO Series in 2021. arXiv preprint. https://doi.org/10.48550/arXiv.2107.08430
Redmon J, Farhadi A, 2018, YOLOv3: An Incremental Improvement. arXiv preprint. https://doi.org/10.48550/arXiv.1804.02767
Du D, Qi H, Yu Q, et al., 2018, The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking. European Conference on Computer Vision (ECCV), 370–386.