1887
PDF

Abstract

Summary

The interferometric optical fiber sensors acquire the phase change of coherent light corresponding to the optical path change caused by the physical quantity under measurement. Thus, the sensing results inescapably wrap due to the usage of inverse trigonometric function during phase demodulation. Phase unwrapping aims to reconstruct true phase from its principal value (wrapped phase). As the existing trade-off between the sample rate and maximum measurement distance, long range distributed optical sensors degrade the validity more commonly and require a new phase unwrapping algorithm especially. In this paper, we propose a 1D phase unwrapping deep-learning network termed view convolution network (VCN) which takes wrapped phase as input and predicts the wrap count. Then the required true phase can be calculated precisely from them. By supervised learning, a 19 million parameters VCN model is validated on two real artificial vibration datasets with 2K and 4K sample rates by DAS system and our model is able to recover most phase points correctly while classical unwrapping function fails. Besides, another advatange is that unwrapping results of our model would not be affected by former errors which propergate in classical methods.

Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.202376037
2023-11-15
2025-02-19
Loading full text...

Full text loading...

/deliver/fulltext/2214-4609/2023/3rd-eage-workshop-on-fiber-optic-sensing-for-energy-applications/37_-_Junyi_Duan.html?itemId=/content/papers/10.3997/2214-4609.202376037&mimeType=html&fmt=ahah

References

  1. K.Wang, Q.Kemao, J.Di, and J.Zhao. [2022] Deep learning spatial phase unwrapping: a comparative review. Adv. Photonics Nexus.
    [Google Scholar]
  2. L.Kong, K.Cui, M.Zhu, and S.Li. [2022] 1d phase unwrapping based on the quasi-gram matrix and deep learning for interferometric optical fiber sensing applications.J. Light. Technol.40, 252–26.
    [Google Scholar]
  3. L.Kong, X.Pan, Z.Ren, and K.Cui. [2022] Robust one-dimensional phase unwrapping algorithm based on lstm network with reduced parameter number.J. Light. Technol.40, 6560–6567.
    [Google Scholar]
/content/papers/10.3997/2214-4609.202376037
Loading
/content/papers/10.3997/2214-4609.202376037
Loading

Data & Media loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error