This paper describes a license plate recognition method, exemplified by training and deploying a machine learning model. The study uses the YOLO (“You Only Look Once”) neural network architecture and optical character recognition (OCR) techniques to extract license plate characters for real-time license plate recognition. Experimental tests, including model training, validation, and evaluation, demonstrate the effectiveness of these methods in enhancing automated access control in smart parking systems
- J. Redmon, S. Divvala, R. Girshick and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 2016, pp. 779-788, https://doi.org/10.48550/arXiv.1506.02640
- [Electronic resource] Smart Parking Market Size is projected to reach USD 16.54 Billion by 2030, growing at a CAGR of 13.6%: Straits Research,https://www.globenewswire.com/en/news-release/2023/09/14/2743480/0/en/Smart-Parking-Market-Size-is-projected-to-reach-USD-16-54-Billion-by-2030-growing-at-a-CAGR-of-13-6-Straits-Research.html
- [Electronic resource] AutoRia Dataset, https://nomeroff.net.ua/datasets
- [Electronic resource] Ultralytics YOLOv8 Docs, https://docs.ultralytics.com/models/yolov8
- [Electronic resource] Roboflow Documentation, https://docs.roboflow.com
- [Electronic resource] CUDA Toolkit Documentation, https://docs.nvidia.com/cuda/
- Afif, Mouna & Said, Yahia & Atri, Mohamed. (2020). Computer vision algorithms acceleration using graphic processors NVIDIA CUDA. Cluster Computing. 23. 10.1007/s10586-020-03090-6.
- [Electronic resource] Sambasivarao K, “Non-maximum Suppression (NMS), A technique to filter the predictions of object detectors.”, Towards Data Science, https://towardsdatascience.com/non-maximum-suppression-nms-93ce178e177c
- [Electronic resource] EasyOCR Github page, https://github.com/JaidedAI/EasyOCR
- [Electronic resource] LiteRT overview, https://ai.google.dev/edge/litert