Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation
We propose a distributed optical fiber sensing event recognition scheme based on Markov Transition Field (MTF) and knowledge distillation. The event recognition algorithm has the advantages of being lightweight, fast, and high accuracy. The event data are converted into images by the MTF algorithm,...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2023-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/10050869/ |
_version_ | 1811160959177195520 |
---|---|
author | Zhenguo Yang Huomin Dong Faxiang Zhang Shaodong Jiang Jinwei Wang Chang Wang Chunxiao Wang |
author_facet | Zhenguo Yang Huomin Dong Faxiang Zhang Shaodong Jiang Jinwei Wang Chang Wang Chunxiao Wang |
author_sort | Zhenguo Yang |
collection | DOAJ |
description | We propose a distributed optical fiber sensing event recognition scheme based on Markov Transition Field (MTF) and knowledge distillation. The event recognition algorithm has the advantages of being lightweight, fast, and high accuracy. The event data are converted into images by the MTF algorithm, which keeps the event signals correlated in the time domain while highlighting the visual differences between different events. A two-stage knowledge distillation model compression method is proposed, which effectively compresses the large-scale model into a lightweight model with optimal learning capability, ensuring the lightweight and efficient recognition of events by the compressed model (student model). The experimental results show that the student model improves the recognition rate of six events by 5.2% and achieves 96.6% event recognition accuracy by the two-stage knowledge distillation method. The size of the student model is only 1.4 MB, the number of parameters is only 0.35 M, and the FLOPs are only 0.17 G. The student model recognizes a single event in 0.129s on a low configuration device, which can meet the requirements of deployment and real-time monitoring of resource-limited devices. |
first_indexed | 2024-04-10T06:06:32Z |
format | Article |
id | doaj.art-b5cdb6e419894c6d95b0100f5ad7d138 |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-10T06:06:32Z |
publishDate | 2023-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-b5cdb6e419894c6d95b0100f5ad7d1382023-03-03T00:00:19ZengIEEEIEEE Access2169-35362023-01-0111193621937210.1109/ACCESS.2023.324823110050869Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge DistillationZhenguo Yang0https://orcid.org/0000-0001-8128-898XHuomin Dong1Faxiang Zhang2https://orcid.org/0000-0002-5555-9608Shaodong Jiang3Jinwei Wang4Chang Wang5Chunxiao Wang6Shandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaShandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaShandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaShandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaShandong Science and Technology Innovation Group, Jinan, ChinaShandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaShandong Computer Science Center, Qilu University of Technology (Shandong Academy of Sciences), Jinan, ChinaWe propose a distributed optical fiber sensing event recognition scheme based on Markov Transition Field (MTF) and knowledge distillation. The event recognition algorithm has the advantages of being lightweight, fast, and high accuracy. The event data are converted into images by the MTF algorithm, which keeps the event signals correlated in the time domain while highlighting the visual differences between different events. A two-stage knowledge distillation model compression method is proposed, which effectively compresses the large-scale model into a lightweight model with optimal learning capability, ensuring the lightweight and efficient recognition of events by the compressed model (student model). The experimental results show that the student model improves the recognition rate of six events by 5.2% and achieves 96.6% event recognition accuracy by the two-stage knowledge distillation method. The size of the student model is only 1.4 MB, the number of parameters is only 0.35 M, and the FLOPs are only 0.17 G. The student model recognizes a single event in 0.129s on a low configuration device, which can meet the requirements of deployment and real-time monitoring of resource-limited devices.https://ieeexplore.ieee.org/document/10050869/Distributed optical fiber sensingMarkov transition fieldknowledge distillationsensor recognitiondeep learning |
spellingShingle | Zhenguo Yang Huomin Dong Faxiang Zhang Shaodong Jiang Jinwei Wang Chang Wang Chunxiao Wang Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation IEEE Access Distributed optical fiber sensing Markov transition field knowledge distillation sensor recognition deep learning |
title | Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation |
title_full | Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation |
title_fullStr | Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation |
title_full_unstemmed | Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation |
title_short | Distributed Optical Fiber Sensing Event Recognition Based on Markov Transition Field and Knowledge Distillation |
title_sort | distributed optical fiber sensing event recognition based on markov transition field and knowledge distillation |
topic | Distributed optical fiber sensing Markov transition field knowledge distillation sensor recognition deep learning |
url | https://ieeexplore.ieee.org/document/10050869/ |
work_keys_str_mv | AT zhenguoyang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT huomindong distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT faxiangzhang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT shaodongjiang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT jinweiwang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT changwang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation AT chunxiaowang distributedopticalfibersensingeventrecognitionbasedonmarkovtransitionfieldandknowledgedistillation |