Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Analysis of Deep Learning Model for the Development of an Optimized Vehicle Occupancy Detection System

최적화된 차량 탑승인원 감지시스템 개발을 위한 딥러닝 모델 분석

  • Lee, JiWon (Department of Computer Engineering, Dong-Eui University) ;
  • Lee, DongJin (Department of Computer Engineering, Dong-Eui University) ;
  • Jang, SungJin (Department of Computer Engineering, Dong-Eui University) ;
  • Choi, DongGyu (Department of Computer Engineering, Dong-Eui University) ;
  • Jang, JongWook (Department of Computer Engineering, Dong-Eui University)
  • Received : 2020.11.30
  • Accepted : 2020.12.28
  • Published : 2021.01.31
Download PDF
⟨ Previous Next ⟩

Abstract

Currently, the demand for vehicles from one family is increasing in many countries at home and abroad, reducing the number of people on the vehicle and increasing the number of vehicles on the road. The multi-passenger lane system, which is available to solve the problem of traffic congestion, is being implemented. The system allows police to monitor fast-moving vehicles with their own eyes to crack down on illegal vehicles, which is less accurate and accompanied by the risk of accidents. To address these problems, applying deep learning object recognition techniques using images from road sites will solve the aforementioned problems. Therefore, in this paper, we compare and analyze the performance of existing deep learning models, select a deep learning model that can identify real-time vehicle occupants through video, and propose a vehicle occupancy detection algorithm that complements the object-ident model's problems.

현재 국내외 여러 국가에서 한 가정의 차량의 수요가 증가하여 차량의 탑승 인원은 적어지고 도로의 차량 수는 증가하고 있는 추세이다. 이에 따른 문제점인 교통 체증을 해결하기 위해 이용 가능한 다인승 전용차로 제도가 시행되고 있다. 이 제도는 경찰들이 빠르게 움직이는 차량을 직접 눈으로 감시하여 불법 차량을 단속하는 실정이며, 이는 정확성이 낮고 사고의 위험성을 동반된다. 이러한 문제점을 해결하기 위해 도로 현장의 영상을 이용한 딥러닝 객체 인식 기술을 적용한다면 앞서 말한 문제점들이 해결될 것이다. 따라서, 본 논문에서는 기존의 딥러닝 모델의 성능을 비교·분석하여, 영상을 통해 실시간 차량 탑승 인원을 파악할 수 있는 딥러닝 모델을 선정하고 객체 인식 모델의 문제점을 보완한 차량 탑승 인원 감지 알고리즘을 제안한다.

Keywords

References

  1. 1 car per 2.1 Korean citizens [Internet]. Available: http://www.goodnews365.net/news/articleView.html?idxno=151618.
  2. Korea Expressway Corporation. Bus lane system of Korea [Internet]. Available: http://www.roadplus.co.kr/useguide/bus/busView.do.
  3. Issuein Korea. Increase in passenger car traffic per person [Internet]. Available: http://www.iikorea.co.kr/news/articleView.html?idxno=128.
  4. D. J. You, M. Y. Kim, and J. W. Jang, "The Implementation of Passenger Count Confirmation System for Bus Lane based on Embedded System," Korea Institute of information and Communication Engineering, pp. 199-202, Oct. 2019.
  5. M. Y. Kim and J. W. Jang, "A Study of The Unmanned System Design of Occupant Number Counter of Inside A Vehicle for High Occupancy Vehicle Lanes," Procedings of the Korean Institute of Information and Commucation Sciences Conference, pp. 49-51, Oct. 2018.
  6. K. P. Kumar, "ESTIMATION OF TRAFFIC MANAGEMENT AND ROAD SAFETY," Asia-pacific Journal of Convergent Research Interchange, HSST, ISSN : 2508-9080, vol. 3, no. 2, pp. 21-28, Jun. 2017. https://doi.org/10.21742/APJCRI.2017.06.03
  7. J. Y. Lee and S. C. Kang, "Final report on overseas advanced technology review seminars for future technology development that collect the number of people on board," Ministry of Land, Land transportation technology promotion research project, R&D 16CTAP-C110346-01-000000, pp. 12-24, 2017.
  8. D. Dubey and G. S. Tomar, "Echelon Based Pose Generalization of Facial Images Approaches," Asia-pacific Journal of Convergent Research Interchange, HSST, ISSN : 2508-9080, vol. 3, no. 1, pp. 63-75, Mar. 2017.
  9. G. Y. Jang, H. H. Hwang, and C. H. Lee, "Vehicle Occupant Detection System by Infrared Sensor," Korea Automotive Engineering Association Autumn Conference and Exhibitione, pp. 708, 2019.
  10. J. Deng, J. Guo, E. Ververas, I. Kotsia, and S. Zafeiriou, "RetinaFace: Single-Shot Multi-Level Face Localisation in the Wild," 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, pp. 5202-5211, 2020. doi: 10.1109/CVPR42600.2020.00525.
  11. TensorFlow-2.x-YOLOv3 [Internet], Available: https://github.com/pythonlessons/TensorFlow-2.x-YOLOv3.
  12. Object Detection on COCO test-dev [Internet]. Available: https://paperswithcode.com/sota/object-detection-on-coco.
  13. M. Tan, R. Pang, and Q. V. Le, "EfficientDet: Scalable and Efficient Object Detection," 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), Seattle, WA, USA, pp. 10778-10787, 2020. doi: 10.1109/CVPR 42600.2020.01079.
  14. Github. Yolov5 [Internet] Available: https://github.com/ultralytics/yolov5.