The Transactions of the Korea Information Processing Society (정보처리학회 논문지)

Korea Information Processing Society (한국정보처리학회)
권호 표지
DOI QR코드

DOI QR Code

A Study on Leakage Detection Technique Using Transfer Learning-Based Feature Fusion

전이학습 기반 특징융합을 이용한 누출판별 기법 연구

  • 한유진 (대구가톨릭대학교 AI빅데이터공학과 ) ;
  • 박태진 (한국원자력연구원 ) ;
  • 이종혁 (대구가톨릭대학교 AI빅데이터공학과) ;
  • 배지훈 (대구가톨릭대학교 AI빅데이터공학과)
  • Received : 2023.07.24
  • Accepted : 2024.01.23
  • Published : 2024.02.29
Download PDF
⟨ Previous Next ⟩

Abstract

When there were disparities in performance between models trained in the time and frequency domains, even after conducting an ensemble, we observed that the performance of the ensemble was compromised due to imbalances in the individual model performances. Therefore, this paper proposes a leakage detection technique to enhance the accuracy of pipeline leakage detection through a step-wise learning approach that extracts features from both the time and frequency domains and integrates them. This method involves a two-step learning process. In the Stage 1, independent model training is conducted in the time and frequency domains to effectively extract crucial features from the provided data in each domain. In Stage 2, the pre-trained models were utilized by removing their respective classifiers. Subsequently, the features from both domains were fused, and a new classifier was added for retraining. The proposed transfer learning-based feature fusion technique in this paper performs model training by integrating features extracted from the time and frequency domains. This integration exploits the complementary nature of features from both domains, allowing the model to leverage diverse information. As a result, it achieved a high accuracy of 99.88%, demonstrating outstanding performance in pipeline leakage detection.

시간 및 주파수 영역에서 각각 학습한 모델 간에 성능 차이가 발생할 경우, 앙상블을 수행하더라도 개별 모델 간의 성능 불균형으로 인하여 앙상블의 성능이 오히려 저하되는 현상을 확인할 수 있었다. 따라서, 본 논문은 시간 영역과 주파수 영역에서 특징을 추출하고, 이들을 융합한 단계적 학습 방법을 통해 파이프라인 누출 감지의 정확성을 높이기 위한 누출판별 기법을 제안한다. 이 방법은 두 단계의 학습 과정으로 이루어지며, 먼저, 단계 1에서는 시간 영역과 주파수 영역에서 독립적으로 모델 학습을 수행하여 도메인별로 주어진 데이터로부터 중요한 특징들을 효과적으로 추출하도록 하였다. 단계 2에서는 사전학습 완료된 각 모델로부터 해당 분류기를 제거한 후, 두 도메인의 특징들을 서로 융합하고 새로운 분류기를 추가하여 재학습을 수행하였다. 본 논문에서 제안하는 전이학습 기반 특징융합 기법은 시간 및 주파수 영역에서 추출된 특징들을 융합하여 모델 학습을 수행함으로써, 두 영역의 특징이 상호 보완적으로 작용하여 모델이 다양한 정보를 활용함으로 인해 99.88%의 높은 정확도를 달성하여 파이프 누수 감지에 있어 우수한 성능을 입증하였다.

Keywords

Acknowledgement

이 논문은 2022년도 정부(과기정통부)의 재원으로 한국연구재단의 지원을 받아 수행된 연구임(No. RS-2022-00165225).

References

  1. R. Puust, Z. Kapelan, D. Savic, and T. Koppel, "A review of methods for leakage management in pipe networks," Urban Water Journal, Vol.7, No.1, pp.25-45, 2010.  https://doi.org/10.1080/15730621003610878
  2. H. Vanaei, A. Eslami, and A. Egbewande, "A review on pipeline corrosion, in-line inspection (ili), and corrosion growth rate models," International Journal of Pressure Vessels and Piping, Vol.149, pp.43-54, 2017.  https://doi.org/10.1016/j.ijpvp.2016.11.007
  3. M.-S. Kim and S.-K. Lee, "Detection of leak acoustic signal in buried gas pipe based on the time-frequency analysis," Journal of Loss Prevention in the Process Industries, Vol. 22, No.6, pp.990-994, 2009.  https://doi.org/10.1016/j.jlp.2008.08.009
  4. D. B. Yoon, S. S. Moon, and B. S. Yang, "A study on acoustic signal processing method for detecting small leak of piping system," Proc. of the Domestic conference on the Korean Society for Noise and Vibration Engineering, Hoengseong, Korea, pp.139-139, 2016. 
  5. Y. Duan, J. S. Edwards, and Y. K. Dwivedi, "Artificial intelligence for decision making in the era of big data-evolution, challenges and research agenda," International Journal of Information Management, Vol.48, pp.63-71, 2019.  https://doi.org/10.1016/j.ijinfomgt.2019.01.021
  6. Z. Wang, J. Zhang, X. Li, and X. Liu, "Deep learning for smart manufacturing: Methods and applications," IEEE Transactions on Industrial Informatics, Vol.14, No.12, pp.5328-5338, 2018. 
  7. J. F. Arinez, Q. Chang, R. X. Gao, C. Xu, and J. Zhang, "Artificial intel-ligence in advanced manufacturing: Current status and future outlook," Journal of Manufacturing Science and Engineering, Vol.142, No.11, pp.110804, 2020. 
  8. D. Yeo, J.-H. Bae, and J.-C. Lee, "Unsupervised learning-based pipe leak detection using deep auto-encoder," Journal of The Korea Society of Computer and Information, Vol.24, No.9, pp.21-27, 2019. https://doi.org/10.9708/jksci.2019.24.09.021. 
  9. J. S. Park, D. Yeo, Y. Choi, J. Lee, and J.-H. Bae. "Implementation of a deep-learning model for intelligent detection of low-level leaks in plant piping system," Proceeding of KIIT Conference, JeJu, Korea, pp.431-435, 2021. 
  10. J. Yosinski, J. Clune, Y. Bengio, and H. Lipson, "How transferable are features in deep neural networks?," Advances in Neural Information Processing Systems, pp.3320-3328, 2014. 
  11. S. J. Pan and Q. Yang, "A survey on transfer learning," IEEE Transactions on Knowledge and Data Engineering, Vol.22, No.10, pp.1345-1359, 2010.  https://doi.org/10.1109/TKDE.2009.191
  12. Y. Dai, F. Gieseke, S. Oehmcke, Y. Wu, and K. Barnard, "Attentional feature fusion," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp.3560-3569, 2021. 
  13. Y. Choi, J. Lee, and J. Cho, "A technology of micro-leak detection," Proceedings of the Korean Society of Computer Information Conference, Vol.29, No.2, pp.685-687, 2021. 
  14. Yoon Kim, "Convolutional Neural Networks for Sentence Classification," Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp.1746-1751, 2014. 
  15. Y. Wang, C. Yan, T. Oates, and D. Li, "Time series classification from scratch with deep neural networks: A strong baseline," Journal of Systems Science and Complexity, Vol.30, No.3, pp.686-701, 2017.  https://doi.org/10.1109/IJCNN.2017.7966039
  16. Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-Based Learning Applied to Document Recognition," Proceedings of the IEEE, Vol.86, No.11, pp.2278-2324, 1998.  https://doi.org/10.1109/5.726791
  17. S. Ioffe and C. Szegedy, "Batch normalization: Accelerating deep network training by reducing internal covariate shift," Proceedings of the 32nd International Conference on Machine Learning, Vol.37, pp.448-456, 2015. 
  18. B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, "Learning deep features for discriminative localization," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.2921-2929, 2016. 
  19. G. Huang, Z. Liu, L. van der Maaten, and K. Q. Weinberger, "Densely connected convolutional networks," Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.4700-4708, 2017. 
  20. L. Bottou, "Stochastic Gradient Descent Tricks," in Neural Networks: Tricks of the Trade, 2nd ed., G. Montavon, G. Orr, and K.-R. Muller, Eds. Berlin, Germany: Springer, pp.421-436, 2012. 
  21. L. Breiman, "Bagging and boosting," in Proceedings of the Fourteenth National Conference on Artificial Intelligence and Ninth Conference on Innovative Applications of Artificial Intelligence, Vol.1, No.1, pp.799-806, 1996. 
  22. M. A. Ganaie, M. Hu, A. Malik, M. Tanveer, and P. Suganthan, "En-semble deep learning: A review," Engineering Applications of Artificial Intelligence, Vol.115, pp.105151, 2022.