자동차안전학회지 (Journal of Auto-vehicle Safety Association)

한국자동차안전학회 (Korean Auto-vehicle Safety Association)
권호 표지
DOI QR코드

DOI QR Code

저속 특장차의 도심 자율주행을 위한 신호등 인지 알고리즘 적용 및 검증

Implementation and Validation of Traffic Light Recognition Algorithm for Low-speed Special Purpose Vehicles in an Urban Autonomous Environment

  • 투고 : 2021.09.10
  • 심사 : 2022.11.16
  • 발행 : 2022.12.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

In this study, a traffic light recognition algorithm was implemented and validated for low-speed special purpose vehicles in an urban environment. Real-time image data using a camera and YOLO algorithm were applied. Two methods were presented to increase the accuracy of the traffic light recognition algorithm, and it was confirmed that the second method had the higher accuracy according to the traffic light type. In addition, it was confirmed that the optimal YOLO algorithm was YOLO v5m, which has over 98% mAP values and higher efficiency. In the future, it is thought that the traffic light recognition algorithm can be used as a dual system to secure the platform safety in the traffic information error of C-ITS.

키워드

과제정보

본 논문은 중소벤처기업부 '규제자유특구혁신사업육성(R&D)'의 지원을 받아 연구되었음(No. P0012726).

참고문헌

  1. S. Saini, S. Nikhil, K. R. Konda, H. S. Bharadwaj, and N. Ganeshan, 2017, "An efficient vision-based traffic light detection and state recognition for autonomous vehicles", in 2017 IEEE Intelligent Vehicles Symposium (IV), pp. 606~611.
  2. Z. Ouyang, J. Niu, Y. Liu, and M. Guizani, 2019, "Deep CNN-based real-time traffic light detector for self-driving vehicles", IEEE transactions on Mobile Computing, Vol. 19, No. 2, pp. 300~313.
  3. K. Yabuuchi, M. Hirano, T. Senoo, N. Kishi, and M. Ishikawa, 2020, "Real-time traffic light detection with frequency patterns using a high-speed camera", Sensors, Vol. 20, No. 14, p. 4035.
  4. Kulkarni, Ruturaj, Shruti Dhavalikar, and Sonal Bangar, "Traffic light detection and recognition for self driving cars using deep learning", 2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA), IEEE, 2018.
  5. de Mello, Jean Pablo Vieira, et al., "Deep traffic light detection by overlaying synthetic context on arbitrary natural images", Computers & Graphics 94(2021): 76~86. https://doi.org/10.1016/j.cag.2020.09.012
  6. J. Redmon and A. Farhadi, 2018, "Yolov3: An incremental improvement", arXiv preprint arXiv:1804.02767.
  7. A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, 2020, "Yolov4: Optimal speed and accuracy of object detection", arXiv preprint arXiv:2004.10934.
  8. A. S. Glenn Jocher, Jirka Borovec, NanoCode012, Ayush Chaurasia, TaoXie, Liu Changyu, Abhiram V, Laughing, tkianai, yxNONG, Adam Hogan, lorenzomammana, AlexWang1900, Jan Hajek, Laurentiu Diaconu, Marc, Yonghye Kwon, oleg, wanghaoyang 0106, Yann Defretin, Aditya Lohia, ml5ah, Ben Milanko, Benjamin Fineran, Daniel Khromov, Ding Yiwei, Doug, Durgesh, and Francisco Ingham, 2021, "ultralytics/yolov5: v5.0 - YOLOv5-P6 1280 models, AWS, Supervise.ly and YouTube integrations", ed. Zenodo.
  9. S. R. Young, D. C. Rose, T. P. Karnowski, S.-H. Lim, and R. M. Patton, 2015, "Optimizing deep learning hyper-parameters through an evolutionary algorithm", in Proceedings of the Workshop on Machine Learning in High-Performance Computing Environments, pp. 1~5.
  10. H. Liu, T. Taniguchi, Y. Tanaka, K. Takenaka, and T. Bando, 2017, "Visualization of driving behavior based on hidden feature extraction by using deep learning", IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 9, pp. 2477~2489. https://doi.org/10.1109/TITS.2017.2649541
  11. J. Yosinski, J. Clune, A. Nguyen, T. Fuchs, and H. Lipson, 2015, "Understanding neural networks through deep visualization", arXiv preprint arXiv:1506.06579.
  12. J. Han, J. Pei, and M. Kamber, 2011, Data mining: concepts and techniques. Elsevier.
  13. Tzutalin. LabelImg [Online] Available: https://github.com/tzutalin/labelImg