Phase-sensitive optical time domain reflectometry (Φ-OTDR) is easy to be interfered by ambient noises, and the nonlinear coherent addition of different interferences always makes it difficult to detect real human intrusions and causes high nuisance alarm rates (NARs) in practical applications. In this paper, an effective temporal signal separation and determination method is proposed to improve its detection performance in complicated noisy environments. Unlike the conventional analysis of transverse spatial signals, the time-evolving sensing signal of Φ-OTDR system is at first obtained for each spatial point by accumulating the changing OTDR traces at different moments. Then, its longitudinal temporal signal is decomposed and analyzed by a multi-scale wavelet decomposition method. By selectively recombining the corresponding scale components, it can effectively extract human intrusion signals, and separate the influences of slow change of the system and other environmental interferences. Compared with the conventional differentiation way and fast Fourier transformation denoising method, the SNRs of the detecting signals for the proposed method is always the best, which can be raised by up to ~35 dB for the best case. Moreover, from the decomposed components, different event signals can be effectively determined by their energy distribution features, and the NAR can be controlled to be less than 2% in the test.

High sensitivity of a distributed optical-fiber vibration sensing (DOVS) system based on the phase-sensitivity optical time domain reflectometry (Φ-OTDR) technology also brings in high nuisance alarm rates (NARs) in real applications. In this paper, feature extraction methods of wavelet decomposition (WD) and wavelet packet decomposition (WPD) are comparatively studied for three typical field testing signals, and an artificial neural network (ANN) is built for the event identification. The comparison results prove that the WPD performs a little better than the WD for the DOVS signal analysis and identification in oil pipeline safety monitoring. The identification rate can be improved up to 94.4%, and the nuisance alarm rate can be effectively controlled as low as 5.6% for the identification network with the wavelet packet energy distribution features.

With the rapid development and extensive applications of phase-sensitive optical time-domain reflectometry ( $\Phi$ -OTDR) to long distance pipeline safety monitoring, it is still challenging to find a very efficient way to achieve highly correct recognition and really deep understanding of physical events sensed in a wide dynamic environment, as the vibration signals usually exhibit non-linear and non-stationary characteristics caused by the complicated environments. In this paper, a dynamic time sequence recognition and knowledge mining method based on the hidden Markov models (HMMs) is proposed to solve this problem. Firstly, local structure feature of the signal is extracted in multiple analysis domains in the time sequence order; and then the HMMs are trained, built and used to mine the temporal evolution information and identify the sequential state process of typical events. The experimental results with real field test data show that the average recognition accuracy of this work is as high as 98.2% for frequently encountered five typical events along buried pipelines. All the related performance metrics such as precision, recall, and F-score are better than those traditional machine learning methods such as the SVM, RF, XGB, DT and BN.

Extracting more and more accurate information to understand the detected vibration or acoustic targets better, has always been an important goal in signal recognition for Distributed Acoustic Sensor(DAS) with optical fiber. In this paper, we use one-dimensional Convolution Neural Networks(1D-CNNs) to extract the detailed temporal structure information at each signal node and utilize a bidirectional Long Short Term Memory(BiLSTM) network to dig out the spatial relationship among the different signal nodes, and then propose a novel identification method by treating the spatial- and temporal- information in a different way, which is denoted as the 1DCNNs-BiLSTM model. The experimental results on the field data show better recognition performance can be achieved in the safety monitoring of the buried optical communication cable in urban with DAS. It helps to improve the recognition rate further compared with the other deep-learning methods frequently or possibly used for DAS signal recognition, such as the 1D-CNNs with a single temporal feature extraction, and 1DCNN-CNN and 2D-CNN models with simultaneous spatiotemporal feature learning. To the best of our knowledge, it is the first time to simultaneously extract and utilize the detailed temporal structure feature and the overall spatial connection through a customized deep learning network.

Presently, it is still challenging to obtain satisfied identification results for long-distance safety monitoring with fiber distributed acoustic sensor (DAS) in practical complicated burying environments. Thus, extracting increasingly abundant features has always been the direction of DAS signal recognition. This paper proposes a new recognition method using an end-to-end mCNN-HMM combined model, which can identify the vibration sources more correctly by simultaneously extracting multi-scale structural features and the sequential information of the DAS signals. A modified multi-scale convolution neural network (mCNN) is designed to automatically extract the DAS signals' local structural features from a multilevel perspective and their relationship in the proposed model. A hidden Markov model (HMM) is then used to mine the sequential information of the whole sample's previously extracted features. The test results based on real field data show that it outperforms the HMM model with the all-around hand-crafted features, the CNN-HMM model, and the MS-CNN-HMM model in both the feature extraction ability and the recognition accuracy in the case of little increase in time consumption. Moreover, the Euclidean distance between the posterior probabilities classified correctly and incorrectly is proposed to evaluate the test samples' feature distinguishability for different recognition models. Then the feature extraction capabilities of the models can be measured in an objective parameter.