A fault recognition method based on grad-cam attention guidance

A fault identification and attention technology, applied in neural learning methods, seismology, instruments, etc., can solve problems such as inconsistency in human objective understanding and key pixel deviation, and achieve the effect of improving fault occurrence, guiding attention, and increasing attention.

Active Publication Date: 2021-07-20
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF19 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

There may be some deviations in the key pixels of the fault classification made by the neural network, which is inconsistent with the objective understanding of human beings

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
  • A fault recognition method based on grad-cam attention guidance
  • A fault recognition method based on grad-cam attention guidance
  • A fault recognition method based on grad-cam attention guidance

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0033] Prior art related to the present invention

[0034] (1) Convolutional Neural Network (CNN)

[0035] A convolutional neural network is a type of neural network that is used to process data with a grid-like structure. For example, time series data (which can be regarded as a one-dimensional grid formed by regular sampling on the time axis, such as text and speech) and image data (which can be regarded as a two-dimensional pixel grid). Convolutional networks perform well in many application fields such as computer vision and natural language processing. Convolutional neural networks use a mathematical operation known as convolution. Convolution is a special kind of linear operation. A convolutional network refers to a neural network in which at least one layer of the network uses convolution operations to replace general matrix multiplication operations. Convolutional networks usually also contain pooling layers.

[0036] In two-dimensional image data, the specific st...

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 fault recognition method based on grad-CAM attention guidance, comprising the following steps: S1, obtaining the attention map of the convolutional neural network through grad-CAM; S2, adding attention to the objective function of the convolutional neural network Trying to use the cross-entropy loss function of the attention map marked by geoscience experts to obtain a new convolutional neural network objective function; S3, using the objective function obtained in step S2 to train a fault recognition model. On the basis of a typical deep learning framework, the present invention introduces an attention guidance mechanism, which can effectively increase the attention of the network on the fault and its neighboring pixels, and can effectively guide the neural network to make fault classification judgments, and can effectively improve fault detection. When the recognition result is broken, a recognition result with better continuity is obtained.

Description

technical field [0001] The invention belongs to the technical field of seismic data identification, in particular to a fault identification method based on grad-CAM attention guidance. Background technique [0002] Seismic data interpretation is an important step in oil and gas exploration, and fault identification is an important part of seismic data interpretation. A fault is a structure in which the crust is fractured by force, and the rock blocks on both sides of the fracture surface undergo significant relative displacement. Faults disrupt the continuity of rock formations. The nature of the fault, the degree of fragmentation and compaction, and the contact relationship between the lithological combinations on both sides of the fault plane are closely related to the migration, accumulation and destruction of oil and gas. The same fault plays different roles in the deep and shallow parts; in the course of historical development, it may also play two opposite roles of s...

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 Patents(China)
IPC IPC(8): G01V1/28G01V1/30G06N3/04G06N3/08
CPCG01V1/282G01V1/306G06N3/08G01V2210/6161G01V2210/624G06N3/045
Inventor 姚兴苗李岱周成胡光岷
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products