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

Adversarial neural network high-resolution seismic fault detection method and system

A neural network and high-resolution technology, applied in the field of anti-neural network high-resolution seismic fault detection, can solve the problems of low detection accuracy and resolution of seismic faults, weakening background information and highlighting fault characteristics, etc., to improve prediction ability and general simplification, mitigation of the effects of lower detection accuracy and resolution

Active Publication Date: 2021-11-26
INST OF GEOLOGY & GEOPHYSICS CHINESE ACAD OF SCI
View PDF7 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these methods highlight fault features by weakening the background information, and the detection accuracy and resolution of seismic faults are relatively low

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
  • Adversarial neural network high-resolution seismic fault detection method and system
  • Adversarial neural network high-resolution seismic fault detection method and system
  • Adversarial neural network high-resolution seismic fault detection method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] figure 1 According to the present invention provides a method for detecting a flowchart earthquake fault against a high resolution neural network. like figure 1 , The method includes the following steps:

[0028] Step S102, based on pre-set training set for neural network training goals against, goal after a confrontation be trained neural network; the default training sample set, including seismic data and fault label.

[0029] In an embodiment of the present invention, the target against neural network comprising: a segmentation module, a feature integration modules and differential module; segmentation module based on the training sample set to obtain a predetermined characteristic fault module; alternatively, the segmentation module for the U-Net type network; wherein the fusion module for the modules of faults with the seismic data of FIG characterized global fusion; target discriminator means for identifying characteristic features in the drawings is a global tag fro...

Embodiment 2

[0082] Figure 8 According to the present embodiment provides a high resolution seismic schematic neural network fault detection system against. like Figure 8 Shown, the system comprising: a training device 10 and the detection means 20.

[0083] Specifically, the training device 10, based on a preset target against the training set of the neural network training, after the target against neural network is trained; default training sample data set comprising seismic and tomographic tag; target against neural network comprises: segmentation module, a feature integration modules and differential modules; the modules based on a predetermined segmentation module training set of faults is obtained; wherein the fusion module for the fault characteristics and seismic data fusion module is global features of FIG.

[0084] Detection means 20 for target-based confrontation after training the neural network, the target image earthquake earthquake fault detection.

[0085] Embodiment provides...

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 provides an adversarial neural network high-resolution seismic fault detection method and system, and the method comprises the steps: training a target adversarial neural network based on a preset training sample set, and obtaining a trained target adversarial neural network, wherein the preset training sample set comprises seismic data and fault labels; the target adversarial neural network comprises a segmentation module, a feature fusion module and a discriminator module; the segmentation module is a module for obtaining fault features based on a preset training sample set; the feature fusion module is used for fusing fault features and seismic data into a global feature map, and performing seismic fault detection on the target seismic image based on the trained target adversarial neural network. The technical problem that in the prior art, seismic fault detection precision and resolution are low is solved.

Description

Technical field [0001] The present invention relates to an earthquake fault interpretation techniques, and in particular, to a neural network fault detection and resolution seismic system against. Background technique [0002] Previous studies of neural network fault identification method focuses on local features of the target, and the network is trained using the synthetic data set. The most common method of detection is the use of tomographic variant of the U-Net, U-Net is applied to an image pixel classification neural network. For example, Wuetal. Simplified (2019) U-Net using a large network of fine binding synthetic seismic data to achieve fault detection. Similarly, Liet al. (2019) using two-dimensional cross-sectional real seismic data U-Net trained network for fault detection. Although many fault segmentation result of fuzzy boundaries, but U-Net network seems to be the choice for image segmentation problem. [0003] To further potential features protruding fault, image...

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): G01V1/34G01V1/30G06K9/62G06N3/04G06N3/08
CPCG01V1/345G01V1/30G06N3/04G06N3/08G01V2210/642G01V2210/74G01V2210/66G06F18/253G06F18/214G01V1/301G06V10/82G06N3/094G06N3/045G06N3/0464G01V20/00G01V2210/612G01V2210/6161G06N3/088
Inventor 王彦飞王天琪
Owner INST OF GEOLOGY & GEOPHYSICS CHINESE ACAD OF SCI
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