Smart Media Journal (스마트미디어저널)

THE KOREAN INSTITUTE OF SMART MEDIA (한국스마트미디어학회)
권호 표지
DOI QR코드

DOI QR Code

Parking Lot Occupancy Detection using Deep Learning and Fisheye Camera for AIoT System

  • Received : 2023.12.16
  • Accepted : 2024.01.29
  • Published : 2024.02.07
Download PDF
⟨ Previous Next ⟩

Abstract

The combination of Artificial Intelligence and the Internet of Things (AIoT) has gained significant popularity. Deep neural networks (DNNs) have demonstrated remarkable success in various applications. However, deploying complex AI models on embedded boards can pose challenges due to computational limitations and model complexity. This paper presents an AIoT-based system for smart parking lots using edge devices. Our approach involves developing a detection model and a decision tree for occupancy status classification. Specifically, we utilize YOLOv5 for car license plate (LP) detection by verifying the position of the license plate within the parking space.

Keywords

Acknowledgement

This work was supported by Hongik University and Ministry of SMEs and Startups

References

  1. Mingkai Chen and Tianhai Chang, "A parking guidance and information system based on wireless sensor network," 2011 IEEE International Conference on Information and Automation, Shenzhen, China, pp. 601-605, 2011.
  2. E. Sifuentes, O. Casas and R. Pallas-Areny, "Wireless Magnetic Sensor Node for Vehicle Detection With Optical Wake-Up," in IEEE Sensors Journal, vol. 11, no. 8, pp. 1669-1676, Aug. 2011.
  3. Z. Zhang, M. Tao and H. Yuan, "A Parking Occupancy Detection Algorithm Based on AMR Sensor," in IEEE Sensors Journal, vol. 15, no. 2, pp. 1261-1269, Feb. 2015.
  4. Alam, M., Moroni, D., Pieri, G., Tampucci, M., Gomes, M., Fonseca, J., Ferreira, J. and Leone, G.R., 2018. Realtime smart parking systems integration in distributed ITS for smart cities. Journal of Advanced transportation, 2018.
  5. Joseph, J., Patil, R.G., Narahari, S.K.K., Didagi, Y., Bapat, J. and Das, D., 2014. Wireless sensor network based smart parking system. Sensors & Transducers, vol .162, no. 1, pp. 5-10, 2014.
  6. Cai, B.Y., Alvarez, R., Sit, M., Duarte, F. and Ratti, C., 2019. Deep learning-based video system for accurate and real-time parking measurement. IEEE Internet of Things Journal, vol. 6, no. 5, pp.7693-7701.
  7. Cho, W., Park, S., Kim, M.J., Han, S., Kim, M., Kim, T., Kim, J. and Paik, J., 2018, January. Robust parking occupancy monitoring system using random forests. In 2018 International Conference on Electronics, Information, and Communication (ICEIC), pp. 1-4.
  8. S. Albawi, T. A. Mohammed, and S. Al-Zawi, "Understanding of a convolutional neural network," in 2017 international conference on engineering and technology (ICET), pp. 1-6, Ieee, 2017.
  9. G. Jocher, "YOLOv5 by ultralytics." https://github.com/ ultralytics/yolov5, 2020 (accessed sep., 10, 2023).
  10. Z.-Q. Zhao, P. Zheng, S.-t. Xu, and X. Wu, "Object detection with deep learning: A review," IEEE transactions on neural networks and learning systems, vol. 30, no. 11, pp. 3212-3232, 2019.
  11. X. Wu, D. Sahoo, and S. C. Hoi, "Recent advances in deep learning for object detection," Neurocomputing, vol. 396, pp. 39-64, 2020.
  12. L. Jiao, F. Zhang, F. Liu, S. Yang, L. Li, Z. Feng, and R. Qu, "A survey of deep learning-based object detection," IEEE access, vol. 7, pp. 128837-128868, 2019.
  13. A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," Communications of the ACM, vol. 60, no. 6, pp. 84-90, 2017.
  14. M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, "Mobilenetv2: Inverted residuals and linear bottlenecks," in Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 4510-4520, 2018.
  15. A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, "Yolov4: Optimal speed and accuracy of object detection," arXiv preprint arXiv:2004.10934, 2020.
  16. Y. Xiong, H. Liu, S. Gupta, B. Akin, G. Bender, Y. Wang, P.-J. Kindermans, M. Tan, V. Singh, and B. Chen, "Mobiledets: Searching for object detection architectures for mobile accelerators," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 3825-3834, 2021.
  17. J. Redmon and A. Farhadi, "Yolov3: An incremental improvement," arXiv preprint arXiv:1804.02767, 2018.
  18. A. Ghosh, S. A. Al Mahmud, T. I. R. Uday, and D. M. Farid, "Assistive technology for visually impaired using tensor flow object detection in raspberry pi and coral usb accelerator," in 2020 IEEE Region 10 Symposium (TENSYMP), pp. 186-189, IEEE, 2020.
  19. A. Yazdanbakhsh, K. Seshadri, B. Akin, J. Laudon, and R. Narayanaswami, "An evaluation of edge tpu accelerators for convolutional neural networks," arXiv e-prints, pp. arXiv-2102, 2021.
  20. V. Mazzia, A. Khaliq, F. Salvetti, and M. Chiaberge, "Realtime apple detection system using embedded systems with hardware accelerators: An edge ai application," IEEE Access, vol. 8, pp. 9102-9114, 2020.
  21. C.-Y. Wang, H.-Y. M. Liao, Y.-H. Wu, P.-Y. Chen, J.-W. Hsieh, and I.-H. Yeh, "Cspnet: A new backbone that can enhance learning capability of cnn," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pp. 390-391, 2020.
  22. Z. Li and F. Zhou, "Fssd: feature fusion single shot multibox detector," arXiv preprint arXiv:1712.00960, 2017.
  23. D. Biswas, H. Su, C. Wang, A. Stevanovic, and W. Wang, "An automatic traffic density estimation using single shot detection (ssd) and mobilenet-ssd," Physics and Chemistry of the Earth, Parts A/B/C, vol. 110, pp. 176-184, 2019.
  24. J. Redmon and A. Farhadi, "Yolov3: An incremental improvement," arXiv preprint arXiv:1804.02767, 2018.
  25. D. Wu, S. Lv, M. Jiang, and H. Song, "Using channel pruning-based yolo v4 deep learning algorithm for the realtime and accurate detection of apple flowers in natural environments," Computers and Electronics in Agriculture, vol. 178, p. 105742, 2020.
  26. B. Jiang, R. Luo, J. Mao, T. Xiao, and Y. Jiang, "Acquisition of localization confidence for accurate object detection," in Proceedings of the European conference on computer vision (ECCV), pp. 784-799, 2018.
  27. "Post-training quantization." https://www.tensorflow.org/lite/performance/posttrainingquantization(accessed sep., 20, 2023).
  28. Z.-C. Xue, N. Xue, and G.-S. Xia, "Fisheye distortion rectification from deep straight lines," arXiv preprint arXiv:2003.11386, 2020.
  29. N. Wojke, A. Bewley, and D. Paulus, "Simple online and realtime tracking with a deep association metric," in 2017 IEEE international conference on image processing (ICIP), pp. 3645-3649, IEEE, 2017.
  30. M. Tan, R. Pang, and Q. V. Le, "Efficientdet: Scalable and efficient object detection," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 10781-10790, 2020.
  31. WongKinYiu,"Yolov4. "https://github.com/WongKinYiu/PyTorch YOLOv4, 2020.
  32. M. Tan, R. Pang, and Q. V. Le, "Efficientdet: Scalable and efficient object detection," in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 10781-10790, 2020.