ANALISIS OPTIMASI MODEL YOLOv8 UNTUK DETEKSI JENIS AWAN PADA JETSON NANO
Abstract
Penerapan deteksi jenis awan berbasis deep learning pada perangkat edge menghadapi keterbatasan sumber daya komputasi yang dapat menurunkan efisiensi inferensi. Penelitian ini menganalisis optimasi model YOLOv8 untuk deteksi 10 jenis awan pada NVIDIA Jetson Nano melalui konversi model ke format TorchScript, ONNX, NCNN, dan MNN dengan presisi FP32 dan FP16. Dataset yang digunakan terdiri atas 4.916 citra, kemudian melalui augmentasi menghasilkan 11.791 citra pelatihan. Model YOLOv8s dilatih selama 100 epoch menggunakan batch size 16, learning rate 0,001, optimizer AdamW, dan pre-trained weight YOLOv8s. Model menghasilkan precision 0,86557, recall 0,90721, F1-score 0,887, mAP50 0,92518, dan mAP50-95 0,73288, dengan bobot terbaik diperoleh pada epoch ke-42. Pengujian pada Jetson Nano menunjukkan NCNN FP16 memberikan performa komputasi terbaik dengan processing time 761,99 ms dan FPS 1,31, meningkat sekitar 6,9 kali dibandingkan baseline PyTorch sebesar 5.238,57 ms dan 0,19 FPS. Nilai mAP50 tetap 0,9927, sedangkan mAP50-95 sebesar 0,8683. Hasil tersebut menunjukkan bahwa NCNN FP16 merupakan format yang paling optimal untuk implementasi YOLOv8 pada Jetson Nano karena mampu meningkatkan efisiensi inferensi dengan penurunan akurasi yang relatif kecil.
Downloads
References
M. N. Fikriansyah, H. A. Nugroho, and M. Sinambela, “Low Cloud Type Classification System Using Convolutional Neural Network Algorithm,” 2022 7th Int. Conf. Informatics Comput. ICIC 2022, pp. 1–6, 2022, doi: 10.1109/ICIC56845.2022.10006907.
S. Kopeć, G. Duniec, B. Bochenek, and M. Figurski, “Artificial neural networks in automatic image classifications of cloud from ground‐based observations using deep learning models,” Q. J. R. Meteorol. Soc., vol. 150, no. 765, pp. 5206–5224, 2024.
M. Wang, S. Zhou, Z. Yang, and Z. Liu, “Clouda: A ground-based cloud classification method with a convolutional neural network,” J. Atmos. Ocean. Technol., vol. 37, no. 9, pp. 1661–1668, 2020.
J. Zhang, P. Liu, F. Zhang, and Q. Song, “CloudNet: Ground‐based cloud classification with deep convolutional neural network,” Geophys. Res. Lett., vol. 45, no. 16, pp. 8665–8672, 2018.
Z. Q. Zhao, P. Zheng, S. T. Xu, and X. Wu, “Object Detection with Deep Learning: A Review,” IEEE Trans. Neural Networks Learn. Syst., vol. 30, no. 11, pp. 3212–3232, 2019, doi: 10.1109/TNNLS.2018.2876865.
J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788. doi: 10.1109/CVPR.2016.91.
J. Terven, D.-M. Córdova-Esparza, and J.-A. Romero-González, “A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS,” Mach. Learn. Knowl. Extr., vol. 5, no. 4, pp. 1680–1716, 2023, doi: 10.3390/make5040083.
Ultralytics, “Ultralytics YOLOv8 Documentation,” 2025. [Online]. Available: https://docs.ultralytics.com/models/yolov8/
M. Bakirci, “Utilizing YOLOv8 for enhanced traffic monitoring in intelligent transportation systems (ITS) applications,” Digit. Signal Process., vol. 152, p. 104594, 2024, doi: 10.1016/j.dsp.2024.104594.
U. A. Hasib, M. A. Raihan, J. Yang, U. A. Bhatti, C. S. Ku, and L. Y. Por, “YOLOv8 framework for COVID-19 and pneumonia detection using synthetic image augmentation,” Digit. Heal., vol. 11, 2025, doi: 10.1177/20552076251341092.
D. Kumar and N. Muhammad, “Object detection in adverse weather for autonomous driving through data merging and YOLOv8,” Sensors, vol. 23, no. 20, p. 8471, 2023.
X. Li, Y. Wang, H. Zhang, J. Liu, M. Chen, and L. Zhao, “Crop pest identification and real-time monitoring system design based on improved YOLOv8s,” Sensors, vol. 26, no. 2, p. 404, 2026, doi: 10.3390/s26020404.
Y. Mao, C. You, J. Zhang, K. Huang, and K. B. Letaief, “A survey on mobile edge computing: The communication perspective,” IEEE Commun. Surv. Tutorials, vol. 19, no. 4, pp. 2322–2358, 2017.
W. Shi, J. Cao, Q. Zhang, Y. Li, and L. Xu, “Edge Computing: Vision and Challenges,” no. c, pp. 1–10, 2016.
P. S. Mittapalli, M. R. N. Tagore, P. A. Reddy, G. B. Kande, and Y. M. Reddy, “Deep learning based real-time object detection on jetson nano embedded gpu BT - Microelectronics, Circuits and Systems: Select Proceedings of Micro2021,” Singapore: Springer Nature Singapore, 2023, pp. 511–521.
T. P. Swaminathan, C. Silver, T. Akilan, and J. Kumar, “Benchmarking deep learning models on nvidia jetson nano for real-time systems: An empirical investigation,” Procedia Comput. Sci., vol. 260, pp. 906–913, 2025.
WMO, “Definition of a cloud,” 2017, World Meteorological Organization. [Online]. Available: https://cloudatlas.wmo.int/en/definition-of-a-cloud.html
T. Ahmad, Y. Ma, M. Yahya, B. Ahmad, S. Nazir, and A. Haq, “Object Detection through Modified YOLO Neural Network,” vol. 2020, 2020, doi: 10.1155/2020/8403262.
P. Hidayatullah, N. Syakrani, M. R. Sholahuddin, T. Gelar, R. Tubagus, and P. N. Bandung, “YOLOv8 to YOLO11 : A Comprehensive Architecture In-depth Comparative Review,” 2025.
“Jetson Nano,” 2026, NVIDIA Corporation. [Online]. Available: https://developer.nvidia.com/embedded/jetson-nano
P. Micikevicius, S. Narang, J. Alben, and et al., “Mixed Precision Training,” in International Conference on Learning Representations (ICLR), 2018. [Online]. Available: https://arxiv.org/abs/1710.03740
“IEEE Standard for Floating-Point Arithmetic,” 2019. doi: 10.1109/IEEESTD.2019.8766229.
A. Paszke et al., “PyTorch: An imperative style, high-performance deep learning library,” Adv. Neural Inf. Process. Syst., vol. 32, pp. 8024–8035, 2019, [Online]. Available: https://arxiv.org/abs/1912.01703
A. D. Hirve, “Introduction to ONNX (Open Neural Network Exchange) framework,” 2023. [Online]. Available: https://www.researchgate.net/publication/400950696
Tencent, “NCNN: A high-performance neural network inference computing framework optimized for mobile platforms,” 2017, GitHub. [Online]. Available: https://github.com/Tencent/ncnn
X. Jiang et al., MNN: A Universal and Efficient Inference Engine. 2020. doi: 10.48550/arXiv.2002.12418.
L. He, Y. Zhou, L. Liu, W. Cao, and J. Ma, “Research on object detection and recognition in remote sensing images based on YOLOv11,” pp. 1–25, 2025.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
