Distributed optical fiber vibration and sound wave sensing signal identification method based on SCNN

Abstract

The invention discloses an SCNN-based distributed optical fiber vibration and sound wave sensing signal identification method. The method comprises the steps of data preparation: constructing different types of distributed optical fiber vibration and sound wave sensing event signal data sets; signal preprocessing: the event signals are subjected to signal preprocessing after being segmented, the signal preprocessing comprises time-frequency transformation, cutting, Gaussian difference filtering and time-frequency feature data set construction, and the time-frequency feature of each event signal comprises a pair of positive and negative time-frequency feature maps after Gaussian difference filtering; constructing and training an unsupervised pulse convolutional neural network SCNN as a feature extraction network based on the time-frequency feature data set; and identification and classification: converting the signal features extracted by the SCNN into feature vectors, and inputting the feature vectors into an SVM classifier for supervised training and classification. According to the method, the over-fitting resistance and generalization resistance of a mainstream supervision recognition model CNN are effectively improved in practical application, and the real-time performance of an unsupervised recognition model SNN in distributed optical fiber sensing signal recognition is effectively improved.

Description

technical field [0001] The invention relates to the application field of distributed optical fiber sensing, in particular to a distributed optical fiber vibration and acoustic wave sensing signal identification method based on a pulsed convolutional neural network (SCNN). Background technique [0002] Optical fiber distributed vibration sensing system and acoustic wave sensing system (DVS / DAS) based on phase-sensitive optical time domain reflectometry (Φ-OTDR), using communication fiber to sense the temporal changes and spatial distribution of physical quantities such as vibration and acoustic waves in the environment along the line It has strong long-distance multi-point positioning ability, high sensing sensitivity, no functional components in the optical fiber, long life, single-ended detection, easy engineering construction and maintenance, so it is an important technical means to realize large-scale environmental safety monitoring. It plays an important role in the appl...

AI Technical Summary

Problems solved by technology

At present, when a supervised deep learning network represented by a convolutional neural network (CNN) performs signal recognition, it requires a high amount of training data. The over-fitting problem of the actual system leads to poor recognition ability of the actual system; the unsupervised spiking neural network (SNN) based on the unsupervised spiking neural network (SNN) proposed in the invention patent of the publication number CN112749637A has significa

Images