Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

1D-CNN based distributed optical fiber sensing signal feature learning and classification method

A 1D-CNN, distributed optical fiber technology, applied in the field of feature learning and classification of distributed optical fiber sensing signals based on 1D-CNN, can solve the problem of complex and changeable large-scale monitoring environments, poor environmental adaptability, and time-consuming feature resolution. and other problems, to achieve the effect of excellent recognition effect, high accuracy, and reduced computational complexity

Active Publication Date: 2018-12-04
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF10 Cites 58 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the problem that in the actual application of the existing distributed optical fiber sound and vibration sensing system, due to the complex and changeable large-scale monitoring environment, it is time-consuming and laborious to manually extract the event distinguishable features of the distributed optical fiber sensing signal. , poor adaptability to complex and changing environments, and high false alarm rate of the system; a 1D-CNN-based distributed optical fiber sensing signal feature learning and classification method is provided

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • 1D-CNN based distributed optical fiber sensing signal feature learning and classification method
  • 1D-CNN based distributed optical fiber sensing signal feature learning and classification method
  • 1D-CNN based distributed optical fiber sensing signal feature learning and classification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0076] Taking the application of long-distance pipeline safety monitoring as an example, based on the one-dimensional convolutional neural network (1D-CNN) distributed optical fiber sound and vibration sensing signal feature learning and classification method, the entire signal processing process is as follows: figure 1 As shown, it is mainly divided into three parts:

[0077] The first part is data preparation. Using the distributed fiber optic sound and vibration sensing system hardware based on the phase-sensitive optical time domain reflectometer to collect the sound or vibration signals along the pipeline in the complex background environment of the actual application site (that is, the distributed fiber optic sensing signal), the data collected at each space point The pipeline event signal time series is divided by time period to construct a typical event signal data set.

[0078] In the second part, a one-dimensional convolutional neural network (1D-CNN) is constructed...

Embodiment 2

[0081] The system hardware used for signal acquisition in Embodiment 1 is a distributed optical fiber sound and vibration sensing system based on the phase-sensitive optical time-domain reflectometry (Φ-OTDR) technology of phase demodulation. The system structure and its working principle are as follows: figure 2 shown. The system hardware consists of three parts, detection optical cable, optical signal demodulation equipment, and signal processing host. Detection cables usually use ordinary single-mode communication fibers, which are usually buried along underground pipelines, transmission cables, and urban roads, and can also directly use the spare cores of communication cables laid along pipelines or roads. Optical signal demodulation equipment is the core of the system, and its internal components mainly include optical devices and electrical devices. A continuous coherent optical signal is generated by an ultra-narrow linewidth laser, which is modulated into an optical ...

Embodiment 3

[0086] The one-dimensional time series of each spatial point in the spatio-temporal response signal matrix accumulated in Embodiment 2 is divided into event signals along the time axis by column to construct a typical event signal data set raw_data. In the present invention, pipeline safety monitoring is taken as an example to construct a typical event signal data set raw_data related to pipeline safety. The specific operation process is as follows: For the signal time series of each spatial point, the event signal with a time length of L is sequentially intercepted, such as image 3 Shown in the rectangular box in the middle, as an event signal sample, denoted as X1 ,X 2 ...., etc., according to the type of event that actually occurred, label the event type respectively.

[0087] In the process of pipeline safety monitoring, typical event types usually include: stable environmental noise, human excavation, mechanical excavation, vehicle interference and factory interference ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a 1D-CNN based distributed optical fiber sensing signal feature learning and classification method, which solves of a conventional distributed optical fiber sensing system thatdistinguishable features of an event extracted manually have poor adaptability to complex changing environments and the system is time-consuming and laborious. The method includes the following steps: conducting time division on obtained distributed optical fiber sensing sound and vibration signals of space points and establishing a typical event signal data set; constructing a one-dimensional convolutional neural network (1D-CNN) model, and performing iterative updating training on a network through the typical event signal data training set to obtain a typical event signal feature set; after training different types of classifiers by the typical event signal feature set, selecting the best classifier. During testing, test data is input into an optimal 1D-CNN and distinguishable featuresof an event are obtained, and then a classification result is obtained by inputting the best classifier.

Description

technical field [0001] Based on the 1D-CNN distributed optical fiber sensing signal feature learning and classification method, the artificial intelligence method is used in the feature extraction and classification of distributed optical fiber sound and vibration sensing signals, which is suitable for underground pipeline networks, long-distance pipelines, communications Optical cables, power cables, perimeter and structural safety monitoring and other application fields. Background technique [0002] As a representative of distributed optical fiber sensing technology, phase-sensitive optical time domain reflectometry (Φ-OTDR) uses optical fiber to sense the spatial distribution and temporal change information of physical quantities such as sound waves and vibrations in the environment along the line. This technology has strong long-distance multi-point positioning capabilities At the same time, the sensing sensitivity is high, there are no functional devices in the optical...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F2218/08G06F2218/12G06F18/285G06N20/10G06N20/20G06N5/01G06F17/18G06N3/04
Inventor 吴慧娟陈吉平刘香荣肖垚王梦娇唐波杨明儒邱浩宇饶云江
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com