Laser & Optoelectronics Progress, Volume. 58, Issue 13, 1306010(2021)
Research Status of Brillouin Signal Analysis Method Based on Machine Learning
Fig. 3. Fiber temperature distribution diagram extracted by ANN under different frequency sweep intervals, where temperature of fiber at 41 m of tail is (a) 21.00 ℃ (room temperature), (b) 29.90 ℃, (c) 39.14 ℃, and (d) 48.63 ℃[30]
Fig. 6. Measured BGS distribution and temperature distribution extracted by SVM-0.5 °C[32]
Fig. 7. Uncertainty and RMSE comparison curves under different SNR[32]. (a1) (a2) 11.5 dB; (b1) (b2) 6.1 dB
Fig. 8. Structure and extraction principle of DNN[35]. (a) DNN structure; (b) schematic of extracting temperature and stress at the same time by DNN
Fig. 9. Experimental device of LEAF and comparison curves of extraction results of different methods[35]. (a) Experimental device; (b) comparison curves of temperature and stress under DNN and equation solving method
|
Get Citation
Copy Citation Text
Liang Wang, Hao Wu, Ming Tang, Deming Liu. Research Status of Brillouin Signal Analysis Method Based on Machine Learning[J]. Laser & Optoelectronics Progress, 2021, 58(13): 1306010
Category: Fiber Optics and Optical Communications
Received: Feb. 26, 2021
Accepted: May. 28, 2021
Published Online: Jul. 14, 2021
The Author Email: Hao Wu (wuhaoboom@qq.com), Ming Tang (tangming@mail.hust.edu.cn)