DOI QR코드

DOI QR Code

Traffic Light Recognition Using a Deep Convolutional Neural Network

심층 합성곱 신경망을 이용한 교통신호등 인식

  • Kim, Min-Ki (Dept. of Computer Science, Gyeongsang National University, Engineering Research Institute)
  • Received : 2018.08.07
  • Accepted : 2018.10.18
  • Published : 2018.11.30

Abstract

The color of traffic light is sensitive to various illumination conditions. Especially it loses the hue information when oversaturation happens on the lighting area. This paper proposes a traffic light recognition method robust to these illumination variations. The method consists of two steps of traffic light detection and recognition. It just uses the intensity and saturation in the first step of traffic light detection. It delays the use of hue information until it reaches to the second step of recognizing the signal of traffic light. We utilized a deep learning technique in the second step. We designed a deep convolutional neural network(DCNN) which is composed of three convolutional networks and two fully connected networks. 12 video clips were used to evaluate the performance of the proposed method. Experimental results show the performance of traffic light detection reporting the precision of 93.9%, the recall of 91.6%, and the recognition accuracy of 89.4%. Considering that the maximum distance between the camera and traffic lights is 70m, the results shows that the proposed method is effective.

Keywords

MTMDCW_2018_v21n11_1244_f0001.png 이미지

Fig. 1. Overall process for creating an image of lighting regions.

MTMDCW_2018_v21n11_1244_f0002.png 이미지

Fig. 4. Samples of the training data.

MTMDCW_2018_v21n11_1244_f0003.png 이미지

Fig. 5. Input images of the DCNN classifier.

MTMDCW_2018_v21n11_1244_f0004.png 이미지

Fig. 2. (a) input image, (b) ROI image (c) binary image B1, (d) binary image B2.

MTMDCW_2018_v21n11_1244_f0005.png 이미지

Fig. 3. (a) size filtered image F2, (b) merged image IF.

Table 1. Each class size of the training data

MTMDCW_2018_v21n11_1244_t0001.png 이미지

Table 2. Architecture of the DCNN classifier

MTMDCW_2018_v21n11_1244_t0002.png 이미지

Table 3. Ground truth of test data

MTMDCW_2018_v21n11_1244_t0003.png 이미지

Table 4. Performance of traffic light detection

MTMDCW_2018_v21n11_1244_t0004.png 이미지

Table 5. Performance of traffic light recognition

MTMDCW_2018_v21n11_1244_t0005.png 이미지

Table 6. Performance of the DCNN using augmented data

MTMDCW_2018_v21n11_1244_t0006.png 이미지

Table 7. Performance comparison of traffic light recognition

MTMDCW_2018_v21n11_1244_t0007.png 이미지

References

  1. Z. Cai, Y. Li, and M. Gu, "Real-time Recognition System of Traffic Light in Urban Environment," Proceedings of the 2012 IEEE Symposium on Computational Intelligence for Security and Defense Applications, pp. 1-6, 2012.
  2. P. Jo and J. Lee, "Traffic Light Detection Using Morphometric Characteristics and Location Information in Consecutive Images," Journal of Institute of Control, Robotics and Systems, Vol. 21, No. 12, pp. 1122-1129, 2015. https://doi.org/10.5302/J.ICROS.2015.15.0133
  3. J. Kim and J. Kim, “Performance Improvement of Traffic Signal Lights Recognition Based on Adaptive Morphological Analysis,” Journal of the Korea Institute of Information and Communication Engineering, Vol. 19, No. 9, pp. 2129-2137, 2015. https://doi.org/10.6109/jkiice.2015.19.9.2129
  4. J. Jeong and D. Rho, “Real Time Detection and Recognition of Traffic Lights Using Component Subtraction and Detection Masks,” Journal of the Institute of Electronics Engineers of Korea, Vol. 43, No. 2, pp. 65-72, 2006.
  5. T. Gevers and A. Smeulders, “Color-based Object Recognition,” Pattern Recognition, Vol. 32, No. 3, pp. 453-464, 1999. https://doi.org/10.1016/S0031-3203(98)00036-3
  6. M. Omachi and S. Omachi, "Traffic Light Detection with Color and Edge Information," Proceedings of the IEEE International Conference on Computer Science and Information Technology, pp. 284-287, 2009.
  7. M. D.-Cabrera, P. Cerri, and J. S.-Medina, "Suspended Traffic Lights Detection and Distance Estimation Using Color Features," Proceedings of the IEEE International Conference on Intelligent Transportation Systems, pp. 1315-1320, 2012.
  8. H. Moizumi, Y. Sugaya, M. Omachi, and S. Omachi, "Traffic Light Detection Considering Color Saturation Using In-Vehicle Stereo Camera," Journal of Information Processing, Vol. 24, No. 2, pp. 349-357, 2016. https://doi.org/10.2197/ipsjjip.24.349
  9. M. Kim, “Detection of a Light Region Based on Intensity and Saturation and Traffic Light Discrimination by Model Verification,” Journal of Korea Multimedia Society, Vol. 20, No. 11, pp. 1729-1740, 2017. https://doi.org/10.9717/KMMS.2017.20.11.1729
  10. Z. Ozcelik, C. Tastimur, M. Karakose, and E. Akin, "A Vision Based Traffic Light Detection and Recognition Approach for Intelligent Vehicles," Proceedings of the International Conference on Computer Science and Engineering, pp. 424-429, 2017.
  11. C. Jang, S. Cho, S. Jeong, J. Suhr, and H. Jung, "Traffic Light Recognition Exploiting Map and Localization at Every Stage," Expert Systems with Applications, Vol. 88, pp. 290-304, 2017. https://doi.org/10.1016/j.eswa.2017.07.003
  12. X. Zhou, J. Yuan, and H. Liu, “Real-Time Traffic Light Recognition Based on C-HOG Features,” Computing and Informatics, Vol. 36, No. 4, pp. 793-814, 2017. https://doi.org/10.4149/cai_2017_4_793
  13. M. Weber, P. Wolf, and J. M. Zollner, "Deep TLR: A Single Deep Convolutional Network for Detection and Classification of Traffic Lights," Proceedings of the IEEE Intelligent Vehicles Symposium, pp. 342-348, 2016.
  14. K. Behrendt, L. Novak, and R. Botros, "A Deep Learning Approach to Traffic Lights: Detection, Tracking, and Classification," Proceedings of the IEEE International Conference on Robotics and Automation, pp. 1370-1377, 2017.
  15. V. John, K. Yoneda, Z. Liu, and S. Mita, “Saliency Map Generation by the Convolutional Neural Network for Real-TIme Traffic Light Detection Using Template Matching,” IEEE Transactions on Computational Imaging, Vol. 1, No. 3, pp. 159-173, 2015. https://doi.org/10.1109/TCI.2015.2480006