Skip to content

Dataset for the article titled as "A Novel Intensity and Phase Stacked Analysis of Φ-OTDR System using Deep Transfer Learning and Recurrent Neural Networks"

Notifications You must be signed in to change notification settings

miralab-ai/IzTech-phaseOTDR-dataset

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 

Repository files navigation

IzTech-phaseOTDR-dataset

Dataset for the article below:

Ceyhun Efe Kayan, Kivilcim Yuksel Aldogan, and Abdurrahman Gumus, “Intensity and phase stacked analysis of a Φ-OTDR system using deep transfer learning and recurrent neural networks,” Applied Optics 62, 1753-1764 (2023) (link to paper)

Dataset:

includes Raw Data files for 15 different cases presented in the article.

https://drive.google.com/drive/folders/1U_fTdhzhofmb6wNdpwNNhbNoScKgNt21?usp=sharing

File name format:

[Date][Apzt]V[fvib]Hz.mat

Example: the phase-OTDR traces for vibration amplitude of 3 V (Apzt=3 V) and vibration frequency of 500 Hz (fvib=500 Hz) is recorded as 08Sept2017_1V_500Hz.mat.

Date: 08Sept2017 for all 15 files (the date of the measurements).

Reading Raw Data (main.m) :

Using main.m, one can plot the Phase-OTDR traces by extracting the data from the *.mat file and placing it as rows in a matrix (tracematrix).

→ Just needed to change the filename (example: trname='08Sept2017_2V_500Hz')

In this way, rows in the matrix represent the spatial distribution, and each column presents the variation of the trace at a particular position during the slow-time window. Then, A bandpass filtering is applied for collecting the frequency components centered around the pulse modulator’s frequency shift (an acousto-optic modulator was used in the set up imposing a frequency shift of 𝐹𝐴𝑂𝑀 = 160 MHz (the filter can be implemented by various ways).

Preprocessing:

The following procedures are implemented on tracematrix to obtain the final temporal-spatial matrice:

  • Each term in a Φ-OTDR trace is converted into a complex number by applying the Hilbert transform.
  • By using the magnitude operation, the amplitude matrix is obtained.
  • To obtain the phase matrix, the phase shift of the acousto-optic modulator is subtracted from the angle of each complex term.
  • The differential amplitude matrix is obtained by subtracting the first row of the amplitude matrix from the other columns.
  • The first column of the phase matrix is subtracted from other columns to remove the “initial phase” caused by the laser source.
  • The mode of each term in the phase matrix with 2𝜋 is calculated (not included in the main.m)

Citation

If you use the IzTech-phaseOTDR-dataset in your research, please cite our related publication:

@article{kayan2023intensity,
  title={Intensity and phase stacked analysis of a $\Phi$-OTDR system using deep transfer learning and recurrent neural networks},
  author={Kayan, Ceyhun Efe and Aldogan, Kivilcim Yuksel and Gumus, Abdurrahman},
  journal={Applied Optics},
  volume={62},
  number={7},
  pages={1753--1764},
  year={2023},
  publisher={Optica Publishing Group}
}

About

Dataset for the article titled as "A Novel Intensity and Phase Stacked Analysis of Φ-OTDR System using Deep Transfer Learning and Recurrent Neural Networks"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages