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

Low-frequency seismic data reconstruction method based on multi-seismic-source convolutional neural network

A convolutional neural network and seismic data technology, applied in the field of low-frequency seismic data reconstruction, can solve the problems of ignoring the application of passive source seismic signals, unknown source wavelet and source position, weak signal energy, etc., reaching the level of instrument and signal processing Reduced requirements, reduced collection costs, and high accuracy

Active Publication Date: 2021-06-18
JILIN UNIV
View PDF7 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, passive source seismic data is characterized by low signal-to-noise ratio, weak signal energy, unknown source wavelet and source location, etc.
As a result, for a long time, in the field of exploration, people tend to ignore the application of passive source seismic signals

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
  • Low-frequency seismic data reconstruction method based on multi-seismic-source convolutional neural network
  • Low-frequency seismic data reconstruction method based on multi-seismic-source convolutional neural network
  • Low-frequency seismic data reconstruction method based on multi-seismic-source convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] This embodiment discloses a low-frequency seismic data reconstruction method based on a multi-source convolutional neural network, figure 1 For a specific work flow chart, the following steps are included:

[0041] a. Seismic data preprocessing: perform static correction processing on the original active source seismic data to correct the influence of the undulating surface on the reflection coaxial axis; perform denoising processing on the data to remove microseismic, background noise and other random noise; remove interference waves, Including sound waves, surface waves, industrial electrical interference, ghost reflections, multiple reflections, side waves, bottom waves, reverberation and ringing, etc. Finally, high-quality observed seismic data are obtained;

[0042] b. Process the processed active source data and the original passive source data separately: Passive source seismic data is low-pass filtered through a Butterworth low-pass filter to obtain passive sou...

Embodiment 2

[0054] The following is a detailed introduction to Embodiment 2 of a seismic data reconstruction method based on a multi-source convolutional neural network provided by this application. Embodiment 2 is implemented based on the foregoing Embodiment 1, and is carried out to a certain extent on the basis of Embodiment 1. expansion.

[0055] We demonstrate the success of the entire convolutional neural network on the Marmousi velocity model. Figure 5(a) is the simulated passive source seismic data obtained by cross-correlation method, and Figure 5(b) is the passive source low-frequency seismic data obtained after low-pass filtering. Figure 6(a) and (b) show the active source seismic data obtained from forward modeling and their corresponding low-frequency data. The low-frequency part of the simulated passive source seismic data is input into a pre-trained convolutional neural network to perform low-frequency reconstruction on the active source seismic data. Finally, the obtaine...

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 relates to a low-frequency seismic data reconstruction method based on a multi-seismic-source convolutional neural network. The method comprises the steps of preprocessing seismic data to obtain high-quality active source seismic data; creating a training data set: carrying out low-pass filtering and block processing on the processed passive source and active source data, and respectively taking the processed passive source and active source data as input data and labels; establishing a neural network model: establishing a convolutional neural network model based on multiple seismic sources, wherein the convolutional neural network model is used for reconstructing active source low-frequency seismic data; and training: inputting the processed data and labels into a convolutional neural network model. According to the method, under the condition that active source low-frequency data are lacked, reconstruction is achieved through passive source data, the accuracy of obtained low-frequency information is high, the requirement for an instrument is not high, the collection cost of the low-frequency information is reduced, high-quality seismic data are provided for the following seismic inversion process after simple denoising, various problems caused by low-frequency information missing are improved, and more detailed information can be recovered.

Description

technical field [0001] The invention belongs to the technical field of seismic data reconstruction, and in particular relates to a low-frequency seismic data reconstruction method based on a convolutional neural network combining active sources and passive sources. Background technique [0002] As the focus of oil and gas resource exploration gradually shifts to complex underground structures and deep areas, the requirements for imaging quality of underground structures are getting higher and higher. Accurate velocity modeling is the key to high-precision imaging. Full waveform inversion is currently recognized as the most accurate velocity modeling method in the field of seismic exploration. However, its application still faces many challenges due to deficiencies such as insufficient low-frequency information of seismic data, dependence on initial models, cycle-skipping problems, and computational efficiency problems. Among them, how to obtain broadband seismic data contain...

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): G01V1/30G01V1/36
CPCG01V1/306G01V1/364G01V2210/27G01V2210/624
Inventor 尹语晨韩立国张盼尚旭佳赵炳辉
Owner JILIN UNIV
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