Optimization of YOLOv8 architecture for UAV object capture tasks: analysis of the trade-off between accuracy, speed and computational resources
https://doi.org/10.46973/0201-727X_2025_2_35
Abstract
This paper presents a comprehensive approach to optimizing the YOLOv8n model for object detection tasks using unmanned aerial vehicles (UAVs) under constrained computational resources. The focus is on quantization (INT8/INT4) and pruning (50%/75%) techniques aimed at reducing the model's computational complexity while maintaining acceptable accuracy. As a result of optimization, the YOLOv8n-Optimized-Drone model was developed, demonstrating a 4-fold increase in processing speed on the Raspberry Pi 5 platform compared to the basic version. The model size was reduced by 3.8 times, which is critical for embedded UAV systems.
A specialized dataset with bounding box markup was created for training and validating the model, taking into account the UAV shooting conditions. Field tests confirmed the effectiveness of the proposed method, which provides a balance between performance, power consumption, and accuracy. Additionally, the influence of different quantization and pruning levels on final metrics was investigated, enabling the determination of the optimal configuration for deployment on low-power devices. The obtained results open prospects for further adaptation of the model to dynamic flight conditions and integration with multi-sensor UAV systems.
About the Author
A. V. Satsuk
Donetsk Institute of Railway Transport (DonIRT)
Russian Federation
Russian Federation
Satsuk Alexander Vladimirovich, Chair “Automation, Telemechanics, Communication and Computer Engineering”, Candidate of Engineering Sciences, Associate Professor
References
1. An improved YOLOv5-based algorithm for common ragweed detection in UAV images / S. Xiaoming, S. Tianzeng, M. Haoming [et al.] // Front. Plant Sci. – 2021. – Vol. 15. – P. 1360419. – DOI 10.3389/fpls.2024.1360419.
2. Real-time Object Detection for UAV Aerial Images Based on Improved YOLOv7 / Y. Zhang, X. Chen, H. Li, Y. Wang // Journal of Electronic Imaging. – 2015. – Vol. 32, No. 6. – P. 063002. – DOI 10.3390/electronics12234886.
3. Satsuk, A. V. Evaluation of the effectiveness of YOLO algorithms for real-time object detection in embedded systems of unmanned vehicles / A. V. Satsuk, R. V. Bely, A. E. Ishchenko // Collection of scientific papers of the Donetsk Institute of Railway Transport. – 2024. – No. 4 (75). – P. 73–82. – ISSN 1993-5579.
4. Redmon, J. A. YOLO9000 : Better, Faster, Stronger / J. Redmon, A. Farhadi // Conference on Computer Vision and Pattern Recognition (CVPR). – 2017. – arXiv :1612.08242v1.
5. Satsuk, A. V. Analysis of computer vision algorithm trackers in matters of tracking moving objects in a video stream / A. V. Satsuk, E. G. Voevoda // Collection of scientific papers of the Donetsk Institute of Railway Transport. – 2023. – No. 71. – P. 54–65. – ISSN 1993-5579.
6. Programmer's Library. JavaScript for deep learning / F. Schollet, E. Nilsson, S. Bayleschi [et al.]. – Saint-Petersburg : Piter, 2021. – 576 p. – ISBN 978-5-4461-1697-3.
7. YOLOv8. Ultralytics. Official YOLOv8 repository / G. Jocher, A. Stoken, A. Chaly [et al.]. – 2023. – URL: https://github.com/ultralytics/ultralytics (date of access: 02/07/2025).
8. Zheng, Z. ??? Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation / Z. Zheng, P. Wang, D. Ren [et al.] // Transactions on Cybernetics. – DOI 10.1109/TCYB.2021.3095305.
9. An Efficient UAV-Based Aerial Image Object Detection Method via Improved YOLOv4 / B. Liu, X. Li, G. Wang [et al.] // Remote Sensing. – 2023. – Vol. 12, No. 21. – P. 3599. – DOI 10.1109/SIU53274.2021.9478027.
10. Satsuk, A. V. Features of the development of unmanned aerial vehicles for the railway transport industry / A. V. Satsuk, A. A. Vorobyov // Collection of scientific papers of the Donetsk Institute of Railway Transport. – 2023. – No. 68. – P. 13–21. – ISSN 1993-5579
11. Satsuk, A. V. Autonomous guidance of UAVs using computer vision : the problem of precise control of rudders / A. V. Satsuk, D. V. Shvalov // Automation in transport. – 2024. – Vol. 10, No. 4. – P. 372–381. – DOI 10.20295/2412-9186-2024-10-04-372-381.
12. Lalak, M. Automated Detection of Atypical Aviation Obstacles from UAV Images Using a YOLO Algorithm / M. Lalak, D. Wierzbicki // Sensors (Basel). – 2022. – Vol. 22, No. 17. – P. 6611. – DOI 10.3390/s22176611.
Review
For citations:
Satsuk A.V. Optimization of YOLOv8 architecture for UAV object capture tasks: analysis of the trade-off between accuracy, speed and computational resources. Vestnik Rostovskogo Gosudarstvennogo Universiteta Putej Soobshcheniya. 2025;(2):35-42. (In Russ.) https://doi.org/10.46973/0201-727X_2025_2_35
Views:
10
JATS XML
Email this article 