The Journal of the Korea institute of electronic communication sciences (한국전자통신학회논문지)

Korea Institute of Electronic Communication Science (한국전자통신학회)
권호 표지
DOI QR코드

DOI QR Code

A Study of Kernel Characteristics of CNN Deep Learning for Effective Fire Detection Based on Video

영상기반의 화재 검출에 효과적인 CNN 심층학습의 커널 특성에 대한 연구

  • Received : 2018.11.01
  • Accepted : 2018.12.15
  • Published : 2018.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, a deep learning method is proposed to detect the fire effectively by using video of surveillance camera. Based on AlexNet model, classification performance is compared according to kernel size and stride of convolution layer. Dataset for learning and interfering are classified into two classes such as normal and fire. Normal images include clouds, and foggy images, and fire images include smoke and flames images, respectively. As results of simulations, it is shown that the larger kernel size and smaller stride shows better performance.

본 논문에서는 보안 감시 카메라 영상을 활용하여 화재 검출을 위한 효과적인 심층학습 방안을 제안한다. AlexNet 모델을 기준으로 효과적인 화재 검출을 위한 커널 크기와 커널 이동 간격의 변화에 따른 분류 성능을 비교 분석한다. 학습을 위한 데이터셋은 정상과 화재 2가지 클래스로 분류한다, 정상 영상에는 구름과 안개 낀 영상을 포함하고, 화재 영상에는 연기와 화염을 각각 포함한다. AlexNet 모델의 첫 번째 계층의 커널 크기와 이동 간격에 따른 분류 성능 분석 결과 커널의 크기는 크고, 이동 간격은 작을수록 화재 분류 성능이 우수한 것을 확인할 수 있다.

Keywords

KCTSAD_2018_v13n6_1257_f0001.png 이미지

그림 1. AlexNet 구조 Fig. 1 Structure of AlexNet

KCTSAD_2018_v13n6_1257_f0002.png 이미지

그림 2. VGG-16 구조 Fig. 2 Structure of VGG-16

KCTSAD_2018_v13n6_1257_f0003.png 이미지

그림 3. AlexNet의 커널 이동 간격에 따른 학습 곡선. (a) 이동 간격 4, (b) 이동 간격 3, (c) 이동 간격 2 Fig. 3 Learning curves as strides of AlexNet. (a) stride 4, (b) stride 3, (c) stride 2

KCTSAD_2018_v13n6_1257_f0004.png 이미지

그림 4. AlexNet의 커널 크기에 따른 학습 곡선. (a) 커널 크기 9×9, (b) 커널 크기 7×7 Fig. 4 Learning curves as kernel size of AlexNet. (a) kernel size 9×9, (b) kernel size 7×7

KCTSAD_2018_v13n6_1257_f0005.png 이미지

그림 5. 화재 검출 학습 영상 예 Fig. 5 Training images examples of fire detection

표 1. 화재 검출 데이터셋 구성 Table 1. Configuration of fire detection dataset

KCTSAD_2018_v13n6_1257_t0001.png 이미지

표 2. 이동 간격에 따른 혼동 행렬(커널 크기 11×11) Table 2. Confusion matrix of each stride(kernal size 11×11)

KCTSAD_2018_v13n6_1257_t0002.png 이미지

표 3. 커널 크기에 따른 혼동 행렬(이동 간격 2) Table 3. Confusion matrix of each kernel size(stride 2)

KCTSAD_2018_v13n6_1257_t0003.png 이미지

표 4. 커널 크기와 이동 간격에 따른 평균 정확도 Table 4. Average accuracy as kernel size and stride

KCTSAD_2018_v13n6_1257_t0004.png 이미지

References

  1. National Security Agency, "Fire Safety Policy Implementation Plan for 2017 in accordance with the 1st Basic Fire Safety Policy Basic Plan (2017-2021)," 2017. Announcement No. 2016-344, in Korean.
  2. H. Hu, G. Wang, and Q. Zhang, "Design wireless multi-sensor fire detection and alarm system based on ARM," In Proc. Int. Conf. on Electronic Measurement & Instruments, Beijing, China Aug. 2009, pp. 285-288.
  3. A. Solorzano, J. Fonollosa, L. Fernandez, J. Eichmann, and S. Marco, "Fire detection using a gas sensor array with sensor fusion algorithms," In Proc. Int. Symposium on Olfaction and Electronic Nose(ISOEN), Montreal, Canada, May. 2007, pp. 1-3.
  4. H. Hoff, "Using distributed fibre optic sensors for detecting fires and hot rollers on conveyor belts," In Proc. Int. Conf. for Fibre-optic and Photonic Sensors for Industrial and Safety Applications(OFSIS), Brisbane, Australia, Jan. 2017, pp. 70-76.
  5. K. Lee, Y. Lee, Y. Shim, Y. Song, S. Han, and C. Kang, "Highly sensitive sensors based on metal-oxide nanocolumns for fire detection," J. of Sensors, vol. 17, no. 2 Feb. 2017, DOI:10.3390/s17020303
  6. K. Kim, D. Wang, and S. Han, "Home Security System Based on IoT," J. of the Korea Institute of Electronic Communication Sciences, vol. 12, no. 1, 2017
  7. H. Lee, "The development of resistive leakage current circuit breaker using a ARM processor," J. of the Korea Institute of Electronic Communication Sciences, vol. 12, no. 4, 2017, pp. 615-620. https://doi.org/10.13067/JKIECS.2017.12.4.615
  8. S. Frizzi, R. Kaabi, M. Bouchouicha, J. Ginoux, E. Moreau, and F. Fnaiech, "Convolutional neural network for video fire and smoke detection," In Proc. Annual Conf. of the IEEE Industrial Electronics Society, Florence, Italy, Oct. 2016, pp. 877-882.
  9. W. Sang, "Implementation of image based fire detection system using convolution neural Network," J. of the Korea Institute of Electronic Communication Sciences, vol. 12, no. 2, 2017, pp. 331-336. https://doi.org/10.13067/JKIECS.2017.12.2.331
  10. X. Wu, X. Lu, and H. Leung, "An adaptive threshold deep learning method for fire and smoke detection," In Proc. Int. Conf. on Systems, Man, and Cybernetics(SMC), Banff, Canada, Oct. 2017, pp. 1954-1959.
  11. D. Shen, X. Chen, M. Nguyen, and W. Q. Yan, "Flame detection using deep learning," In Proc. Int. Conf. on Control, Automation and Robotics(ICCAR), Auckland, New Zealand, Apr. 2018.
  12. M. Khan, A. Jamil, and B. S. Wook, "Early fire detection using convolutional neural networks during surveillance for effective disaster management," J. of Neurocomputing, vol. 288, May, 2018, pp. 30-42. https://doi.org/10.1016/j.neucom.2017.04.083
  13. Y. Do, "Visual sensing of fires using color and dynamic features," J. of Sensor Science and Technology, vol. 21, no. 3, 2012, pp. 211-216. https://doi.org/10.5369/JSST.2012.21.3.211
  14. A. Krizhevsky, I. Sutskever, and G. E. Hinton, "ImageNet classification with deep convolutional neural networks," Advances in Neural Information Processing Systems(NIPS), vol. 1, Lake Tahoe, USA, Dec. 2012, pp. 1097-1105.
  15. C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, "Going deeper with convolutions," In Proc. Computer Vision and Pattern Recognition, Boston, USA, Sept. 2014, pp. 1-9.
  16. X. Zhang, J. Zou, K. He, and J. Sun, "Accelerating very deep convolutional networks for classification and detection", IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 38, Nov. 2015. pp.1943-1955