Dual Purpose Vision: A Deep Learning Approach to Helmet Compliance Detection and License Plate Recognition in Real Time Traffic Monitoring

Sayma Nasrin Shompa

doi:10.59890/ijaamr.v3i4.9

Authors

Sayma Nasrin Shompa International Islamic University Chittagong

DOI:

https://doi.org/10.59890/ijaamr.v3i4.9

Keywords:

Deep Learning, Multi- Class Classification, Image Processing, Data Augmentation, Transfer Learning

Abstract

This manuscript presents the implementation of a helmet compliance detection system and a motorcycle license plate recognition system in real-time based on the YOLOv8n object detection model. It differentiates between motorcycles and classifies them as "With Helmet," "Without Helmet," and "Plate" to ensure road enforcement. The developed solution was trained on a custom-labeled dataset that was initially imbalanced, preprocessed, and augmented to achieve class balance and improved model generalization. The lightweight YOLOv8n architecture was chosen, which was fine-tuned through transfer learning and hyperparameter optimization, yielding testing precision from 91% to 92% in three classes, recall rates of 87% to 89%, and progressively increased m AP scores up to 93%. In addition, it achieved a precision score of 96% in detecting motorcycle license plates. Findings demonstrated that the model is efficient and effective, accessible in real-time for future developments as an IoT-based traffic enforcement system. Limitations of the project include small object detection and data imbalance, to be resolved in future endeavors

References

Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. [Online]. Available: https://arxiv.org/abs/2004.10934

Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv preprint arXiv:2004.10934. https://arxiv.org/abs/2004.10934

Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. https://arxiv.org/abs/2004.10934

Challenge. International Journal of Computer Vision. https://link.springer.com/article/10.1007/s11263-009-0275-4

Everingham, M., et al. (2010). The PASCAL Visual Object Classes (VOC) Challenge. International Journal of Computer Vision. The PASCAL Visual Object Classes (VOC)

Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010). The PASCAL Visual Object Classes (VOC) Challenge. International Journal of Computer Vision, 88(2), 303–338.

Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 580-587.

Glenn Jocher et al. (2023). YOLO by Ultralytics. [Online]. Available: https://github.com/ultralytics/yolov5

Glenn Jocher et al. (2023). YOLO by Ultralytics. GitHub Repository: https://github.com/ultralytics/yolov5

Jocher, G., et al. (2023). YOLO by Ultralytics. GitHub. https://github.com/ultralytics/yolov5

Liu, H., Chen, Y., Wang, Y., & Wang, J. (2019). Helmet detection based on improved YOLOv3 model. Proceedings of the 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC), 168-172.

Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. [Online]. Available: https://arxiv.org/abs/1804.02767

Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767. https://arxiv.org/abs/1804.02767

Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. https://arxiv.org/abs/1804.02767

Ruder, S. (2017). An overview of multi-task learning in deep neural networks. ArXiv preprint arXiv:1706.05098. https://arxiv.org/abs/1706.05098

Xie, H., Zhang, J., Li, X., & Yin, B. (2020). Helmet wearing detection based on RCNN. IEEE Access, 8, 144517–144525.

Zhang, Y., Xu, Z., & Wang, H. (2021). Edge-based real-time traffic monitoring using deep learning and edge computing. Sensors, 21(10), 3354.

Zhao, Y., Wang, Z., & Liu, Y. (2017). License plate recognition based on deep learning. 2017 2nd International Conference on Image, Vision and Computing (ICIVC), 745-749.