Fig. 1. Application of distributed fiber optic strain/temperature monitoring system

Fig. 2. SNR and BFS uncertainty before and after NLM denoising^[18]. (a) SNR; (b) BFS uncertainty

Fig. 3. Degradations of temperature and spatial resolution as functions of SNR improvement at different sampling rates^[19]. (a)(d) 500 MSa/s; (b)(e) 300 MSa/s; (c)(f) 200 MSa/s

Fig. 4. Single Brillouin gain spectrum denoising using autoencoder-based neural networks^[31]

Fig. 5. Training process of denoised CNN^[32]

Fig. 6. Principle of temperature information extraction using SVM^[47]

Fig. 7. Diagram of training and test phases required for temperature extraction based on ANN^[63]

Fig. 8. Structure of DNN with two autoencoder hidden layers^[74]

Fig. 9. Architecture of proposed CNN^[75]