The phase signals, combined with time-domain signal processing methods, are often used for recognition with the phase-sensitive optical time-domain reflectometer (Phi-OTDR). Considering the advanced and sophisticated algorithms prevalent in the field of image processing, a vibration signal imaging method is proposed to enhance the adaptability of phase signals for learning by image network. The phase time series is converted to aMarkov Transition Fields (MTF) matrix, from which the based matrix is extracted by Non-negative Matrix Factorization (NMF) and saved as an RGBimage. One-dimensional (1-D) Convolutional Neural Network (CNN) and 2-D CNN are applied in the experiment to classify the phase signals and images, respectively. The experimental results show that the training convergence efficiency of 2-D CNN using NMF-MTF images is significantly higher than that of 1-D CNN, demonstrating the effectiveness of converting phase signals into images. In addition, the average recognition accuracy for the four fence events is improved by more than 13% by introducing the NMF algorithm on the MTF matrix.

Although pulse compression optical time domain reflectometry (PC-OTDR) exhibits high performance in spatial resolution and dynamic range, it inevitably introduces autocorrelation sidelobes, potentially impacting measurement accuracy. In this letter, an improved CLEAN algorithm is proposed to efficiently suppress sidelobes and enhance the peakto- sidelobe ratio (PSLR) of signals in PC-OTDR. The proposed method introduces an adaptive step factor instead of the traditional fixed factor to reduce the number of iterations. Compared to the traditional method, the proposed method achieves a 2.87 dB improvement of PSLR from a 10 km sensing fiber. In addition, the computation time cost is significantly reduced, which is 1.92 s less than that of the traditional CLEAN algorithm.

To improve the model training efficiency and the classification performance of the phase-sensitive optical time-domain reflectometer (Phi-OTDR) in disturbance events recognition, a preprocessing method based on Markov transition fields (MTF) and auto-encoder (AE) is proposed. The phase time series, derived from demodulation of the original scattering signals, are converted into images by using the MTF method. Subsequently, an auto-encoder is introduced to perform a dimensionality reduction characterization of the MTF images, and the outputs of the encoder will be used as features for classification. The experimental results demonstrate that, compared with directly processing time series using 1-D CNN and classifying MTF images using CNN, the features obtained by the proposed method can accelerate the training process and improve the recognition performance of the classification model. The recognition accuracy for the four classes of events on the fence reaches 95.6%, representing a 12% increase.

A concise and adaptive sidelobe suppression algorithm based on a least mean square (LMS) filter is proposed for pulse-compressed signals of a phase-sensitive optical time-domain reflectometer (Φ-OTDR) system. The algorithm is suitable for the denoising filtering process of phase coding OTDR (PC-OTDR) systems and mitigates the sidelobe effect due to matched filtering. In a simulation experiment, Rayleigh backscattering (RBS) signals including phase-coded pulse signals are generated and decoded to verify that the LMS algorithm can eliminate the sidelobes more effectively than the windowing method and the recursive least squares (RLS) method. Then, the PC-OTDR system is set up and combined with the LMS algorithm for positioning experiments. The results show that the peak side lobe ratio (PSLR) of the signals can reach −15.86 dB, which is 4.26 dB lower than the raw pulse compressed signal.

A vibration recognition method based on MTF-NMF images for Φ-OTDR is proposed, which has higher training speed and better recognition performance, compared to 1-D CNN with phase time series and 2-D CNN with MTF images.

Aiming to enhance the vibration localization capability of phase-sensitive optical time domain reflectometry ( $\Phi $ -OTDR), a regional localization method based on overlapping phases cross-correlation is proposed to improve the positioning accuracy and signal-to-noise ratio (SNR). Due to the linear relationship between the phase signal and the external disturbance, the cross-correlation of the two phase matrices are calculated and the maximum value appears in the full link can represent the location of the vibration. Meanwhile, autocorrelation is introduced to improve the localization effect. The regional localization accuracy achieved by the proposed method is above 98%, and the average SNR is 59.37 dB in actual vibration experiment. The improved localization method decreases the noise peak rate by more than 20% in single-point vibration experiment compared with that before the improvement.

To enhance the capability of phase-sensitive optical time domain reflectometers (Φ-OTDR) to recognize disturbance events, an improved adaptive feature extraction method based on NMF-MFCC is proposed, which replaces the fixed filter bank used in the traditional method to extract the mel-frequency cepstral coefficient (MFCC) features by a spectral structure obtained from the Φ-OTDR signal spectrum using nonnegative matrix factorization (NMF). Three typical events on fences are set as recognition targets in our experiments, and the results show that the NMF-MFCC features have higher distinguishability, with the corresponding recognition accuracy reaching 98.47%, which is 7% higher than that using the traditional MFCC features.

The power line communications (PLC) channel is noisy one, which can be modeled by the Markov process. The order is the key concerning when used the Markov process. the level of complexity will be incurred from using higher order, while the first-order Markov models may lead to the less accurate channel response. In the paper, the first-order Markov channel is under thoroughly discussion, and it can provide a mathematically tractable model for time-varying channels and uses only the received SNR of the symbol immediately preceding the current one. With the first-order Markov chain, given the information of the symbol immediately preceding the current one, any other previous symbol should be independent of the current one. We show that given the information corresponding to the previous symbol, the amount of uncertainty remaining in the current symbol should be negligible. That means the first-order of Markov process is enough when modeled the PLC channel.