지적과 국토정보 (Journal of Cadastre & Land InformatiX)

  • 제52권1호
  • /
  • Pages.81-93
  • /
  • 2022
  • /
  • 2508-3384(pISSN)
  • /
  • 2508-3392(eISSN)
한국국토정보공사 공간정보연구원 (LX Spatial Information Research Institute)
권호 표지
DOI QR코드

DOI QR Code

객체 인식 모델 기반 실시간 교통신호 정보 인식

Real-time traffic light information recognition based on object detection models

  • 주은오 (대전대학교 컴퓨터공학과) ;
  • 김민수 (대전대학교 컴퓨터공학과)
  • Joo, eun-oh (Department of Computer Engineering, Daejeon University) ;
  • Kim, Min-Soo (Department of Computer Engineering, Daejeon University)
  • 투고 : 2022.05.04
  • 심사 : 2022.06.22
  • 발행 : 2022.06.30
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

최근 자율주행 기술에서 차량 주변 객체 인식과 교통표지판 및 차량 신호 인식을 위한 연구가 활발히 수행되고 있으며, 특히 차량 신호 인식은 자율주행 기술에 있어서 핵심 요소로 평가되고 있다. 이에 차량 신호 인식을 위한 다양한 연구가 진행되어 왔으며, 최근에는 딥러닝 기반 객체 인식 모델을 활용한 차량 신호 인식 연구가 크게 증가하고 있다. 또한 AIHub에서 음성, 비전, 자율주행 등을 위한 양질의 국내 인공지능 학습데이터 셋이 공개됨에 따라 이들 데이터를 활용한 국내 환경에 적합한 차량 신호 인식 모델의 개발도 가능하게 되었다. 이에 본 연구에서는 AIHub의 학습데이터와 객체 인식모델 YOLO를 적용한 국내 차량 신호 인식 모델을 개발하였다. 특히 차량 신호의 인식 성능을 개선하기 위하여 YOLOv4와 YOLOv5의 다양한 모델을 적용하였으며 학습데이터의 클래스도 다양하게 분류하여 실험을 수행하였다. 결론적으로 YOLOv5가 YOLOv4보다 차량 신호 인식에 조금 더 적합함을 확인할 수 있었으며, 두 모델의 아키텍처 비교를 통하여 YOLOv5 성능이 우수한 이유를 확인할 수 있었다.

Recently, there have been many studies on object recognition around the vehicle and recognition of traffic signs and traffic lights in autonomous driving. In particular, such the recognition of traffic lights is one of the core technologies in autonomous driving. Therefore, many studies for such the recognition of traffic lights have been performed, the studies based on various deep learning models have increased significantly in recent. In addition, as a high-quality AI training data set for voice, vision, and autonomous driving is released on AIHub, it makes it possible to develop a recognition model for traffic lights suitable for the domestic environment using the data set. In this study, we developed a recognition model for traffic lights that can be used in Korea using the AIHub's training data set. In particular, in order to improve the recognition performance, we used various models of YOLOv4 and YOLOv5, and performed our recognition experiments by defining various classes for the training data. In conclusion, we could see that YOLOv5 shows better performance in the recognition than YOLOv4 and could confirm the reason from the architecture comparison of the two models.

키워드

과제정보

This research was supported by the Daejeon University Research Grants (2021)

참고문헌

  1. Kang JO, Lee YC. 2019. Construction of 3D Spatial Information of Vertical Structure by Combining UAS and Terrestrial LiDAR. Journal of Cadastre & Land InformatiX. 49(2):57-66
  2. Kim HS, Park MG, Son WI, Choi HD, Park SK. 2018. Deep Learnging based Object Detection and Distance Estimation using Mono Camera. Journal of Korean Institute of Intelligent Systems. 28(4):201-209 https://doi.org/10.5391/JKIIS.2018.28.3.201
  3. Na YS, Kim SK, Kim YS, Park JY, Jeong JM, Jo KC, Lee SJ, Cho SJ, Sunwoo MH, Oh JM. 2020, HD Map Usability Verification for Autonomous Car. Transaction of the Korean Society of Automotive Engineers. 28(11):797-808 https://doi.org/10.7467/KSAE.2020.28.11.797
  4. Park SB, Kim JH. 2020. Machine-learning-based Real-Time Multi-object Recognition Method for Urban Autonomous Driving. Jounal of Institute of Control Robotics and Systems. 26(6):499-505. https://doi.org/10.5302/J.ICROS.2020.20.0045
  5. Park SJ, Cho K, Im JH, Kim MC. 2021. Location Tracking and Visualization of Dynamic Objects using CCTV Images. Journal of Cadastre & Land InformatiX. 51(1):53-65. https://doi.org/10.22640/LXSIRI.2021.51.1.53
  6. Seo HD, Kim EM. 2020. Object Classification and Change Detection in Point Clouds Using Deep Learning. Journal of Cadastre & Land InformatiX. 50(2):37-51. https://doi.org/10.22640/LXSIRI.2020.50.2.37
  7. Lee DJ, Kim JH, Han SJ, Choi JD, Park CH. 2016. Traffic light recognition using the LISA traffic light dataset. Conference of The Institute of Electronics and Information Engineers. 775-778.
  8. Lee DH, Kim HI. 2020. Object detection and tracking through cameras and LiDAR fusion in autonomous vehicle. Conference of The Korean Society Of Automotive Engineers. 690-693.
  9. Lee TW, Lim KY, Bae GT, Byun HR, Choi YW. 2015. An Illumination Invariant Traffic Sign Recognition in the Driving Environment for Intelligence Vehicles. Korean Institute of Information Scientists and Engineers. 42(2) 203-212.
  10. Jung CY, Lee DG, Shim HC. 2018. Vison-based Traffic Light Recognition System using Faster RCNN. Conference of The Korean Society Of Automotive Engineers. 868-871.
  11. Jung TH, Kim JH. 2017. Traffic Light Recognition based on HSV/YCbCr Color Model and Morphological Feature. Conference of The Korean Society Of Automotive Engineers. 547-551.
  12. Korea Intelligent Information Society Promotion Agency. 2022. AIHub[Internet]. [https://aihub.or.kr/]. Last accessed 19 April 2022.
  13. Evert Bos. 2019. Including traffic light recognition in general object detection with YOLOv2 [theses]. Delft University of Technology.
  14. Glenn Jocher 2020. YOLOv5[Internet]. [https://github.com/Tianxiaomo/pytorch-YOLOv4.git]. Last accessed 27 March 2022.
  15. Tianxiaomo. 2020. pytroch-YOLOv4[Internet]. [https://github.com/Tianxiaomo/pytorchYOLOv4.git]. Last accessed 27 March 2022.
  16. Tai Huu - Phuong Tran, Cuong CP, Tien PN, Tin TD, Jeon JW. 2016. Real-Time Traffic Light Detection Using Color Density. IEEE International Conference on Consumer Electronics-Asia (ICCE-Asia). 26-28
  17. Tai Huu - Phuong Tran, Jeon JW. 2020, Accurate Real-Time Traffic Light Detection Using YOLOv4. IEEE International Conference on Consumer Electronics - Asia (ICCE-Asia). 1-3
  18. Wang Q, Zhang Qi, Liang X, Wang Y, Zhou C, Vladimir IM. 2022. Traffic Lights Detection and Recognition Method Based on the Improved YOLOv4 Algorithm. sensors. 22(1):200-219.