Laser & Infrared, Volume. 54, Issue 2, 312(2024)
Recognition of shallow buried state of submarine cable based on fiber optic temperature hybrid domain features
In response to the non-stationary and nonlinear characteristics of fiber optic temperature signals, as well as the limitations of using fiber optic temperature difference to identify the shallow burial position of submarine cables during the temperature equilibrium time period when the surface temperature of the seabed and the temperature at the depth of the seabed are approximately equal, a shallow burial state identification method is proposed based on optimized VMD mixed domain features and LSTM for identifying the two states of deep and shallow burial of the cables. Firstly, a parameter optimized VMD is used to decompose the fiber temperature signal and extract the component with the highest correlation coefficient between the intrinsic modal components of each order and the original signal. Secondly, the time-domain and frequency-domain features of the original temperature signal are extracted, and a mixed-domain feature set is constructed by combining the time-domain and frequency-domain features as well as the energy and entropy features of the selected IMF, and the CDET is used for sensitive feature selection. Finally, an LSTM structure is designed, the training sets are inputted into the network for training, the test set verifies the effectiveness of the network with the test set, and achieve shallow burial state recognition of submarine cables. Through on-site collection of submarine cable fiber temperature data for verification, the testing accuracy reaches 100%, and the results show that this method can accurately identify the shallow burial state of submarine cables.
Get Citation
Copy Citation Text
JIANG Kun, ZHANG Shuai, FU Xiang, SHI Er-zhen, AN Bo-wen, CHEN Yuan-lin, CUI Gui-yan. Recognition of shallow buried state of submarine cable based on fiber optic temperature hybrid domain features[J]. Laser & Infrared, 2024, 54(2): 312
Category:
Received: Jul. 30, 2023
Accepted: Jun. 4, 2025
Published Online: Jun. 4, 2025
The Author Email: CUI Gui-yan (1363216917@qq.com)