Abstract

This technical note compares different model sizes of YOLOv5, YOLOv7 and YOLOv8 on the NVIDIA Jetson Nano, Jetson Orin and Raspberry Pi 5 platforms. COCO and a self-compiled in-factory safety dataset were used for evaluation. Mean average precision, frames per second and energy consumption are reported together. YOLOv8n reached 142 FPS on Jetson Orin with TensorRT, while quantisation on low-power platforms provided a 3.1x speed-up with less than 2% precision loss.

Keywords:object detectionYOLOembedded systemsreal-timeedge computing

Full Text

PDF (TR) Download · 8 KB

Declarations

Ethics Approval
Bu çalışma etik kurul onayı gerektirmemektedir.
Conflict of Interest
Yazarlar herhangi bir çıkar çatışması olmadığını beyan eder.

References 4

  1. Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 779–788). https://doi.org/10.1109/CVPR.2016.91
  2. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770–778). https://doi.org/10.1109/CVPR.2016.90
  3. Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., et al. (2021). An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations.
  4. Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the 36th International Conference on Machine Learning (pp. 6105–6114).

How to Cite

Yavaş, E., & Güneş, N. (2026). A Comparative Analysis of YOLO Architectures for Real-Time Object Detection on Embedded Systems. International Journal of Science and Technology Research, 8(1), 51–62. https://doi.org/10.99999/ubtad.2026.14
BibTeX RIS (EndNote, Zotero, Mendeley)

License

CC BY 4.0

© 2026 Emre Yavaş, Nazlı Güneş. This article is distributed under the terms of the CC BY 4.0 license, which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited. License text