Pre-stack seismic data random noise suppression method and system

A random noise, pre-stack seismic technology, applied in the field of geophysical exploration technology and deep learning, can solve the problems of unseen research literature, model training takes a long time, manual intervention, etc., to improve training efficiency and avoid training data volume And the high requirements of the neural network scale, to achieve the effect of effective suppression

Pending Publication Date: 2022-05-17
CHINA PETROLEUM & CHEM CORP +1
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Based on the above problems, a random noise suppression technology for seismic data based on the U-NET network was developed. In addition, considering that in the process of intelligent processing, the training of the model takes the longest time, conventional CPU-based training and single-GPU training Failure to make full use of computing resources affects the efficiency of network training. Therefore, the present invention realizes multi-GPU data

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
  • Pre-stack seismic data random noise suppression method and system
  • Pre-stack seismic data random noise suppression method and system
  • Pre-stack seismic data random noise suppression method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0062] Such as Figure 11 As shown, the random noise suppression method of pre-stack seismic data based on U-NET network includes the following steps:

[0063] (1) Prepare data sets using actual seismic data and forward modeling data of pre-stack shots: A key step in deep learning processing is the production of label data sets. In order to ensure the generalization ability of the training network, intelligent noise suppression technology Quality label data source can be composed of two parts: 1. Data analysis and extraction of high-quality denoised seismic data in typical exploration areas: actual data label is an important part of improving the generalization ability of deep network, so data selection should cover a variety of practical Typical detection areas, and select data with high-quality denoising effects to improve the robustness of the data set; 2. Pre-stack shot set forward modeling simulation data: the actual data denoising effect is largely affected by the denois...

Embodiment 2

[0071] Such as Figure 12 As shown, the system includes:

[0072] A data set preparation unit 10 is used to prepare a data set using actual seismic data and pre-stack shot set forward modeling data, and divide the data set into a training data set and a verification data set;

[0073] The network design training unit 20 is connected with the data set preparation unit 10 for designing the network, and utilizes the GPU and the training data set to carry out parallel training on the network to obtain the trained network;

[0074] The verification unit 30 is connected with the network design training unit 20, and is used to verify the trained network by using the verification data set to obtain the verified intelligent random noise suppression network;

[0075] The noise suppression unit 40 is connected with the verification unit 30, and is used to collect pre-stack seismic data, and input the pre-stack seismic data into the verified intelligent random noise suppression network t...

Embodiment 3

[0078] The training input data size is 128x1000, among which, a data set containing 100,000 single-shot results was prepared. Establish the above-mentioned U-NET network, and set the learning rate to 0.001. During the training process, in order to further improve the generalization ability of the network, 5% Gaussian noise is further added for data enhancement. In addition, the batch_size is set to 20, and the number of learning rounds is set to Set at 15 rounds. In addition, the cluster GPU used in this test is RTX 2080Ti with a memory size of 11G, and the number of GPU cards used for training is seven cards. The running status of the GPU cards during the model training process is as follows: image 3 As shown, it can be seen from the figure that all 7 GPU cards have a high utilization rate, which effectively improves the efficiency of network training.

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 a prestack seismic data random noise suppression method and system, and belongs to the field of geophysical exploration technology and deep learning. According to the pre-stack seismic data random noise suppression method, actual seismic data, pre-stack shot gather forward modeling data and a GPU parallel mode are utilized to obtain an intelligent random noise suppression network, the intelligent random noise suppression network is utilized to process the actual pre-stack seismic data, and pre-stack seismic data after random noise suppression are obtained. According to the method, the U-NET network is applied to seismic data denoising, the high requirement of a conventional convolutional network for the training data volume and the neural network scale is effectively avoided, multi-GPU data parallel network training is achieved, the training efficiency is improved, and effective suppression of prestack data random noise is achieved.

Description

technical field [0001] The invention belongs to the field of geophysical exploration technology and deep learning, and specifically relates to a method and system for suppressing random noise of pre-stack seismic data, which is applied to random noise removal processing in denoising processing in petroleum geophysical exploration. Background technique [0002] The existence of noise in seismic data seriously affects the imaging quality of seismic data. How to effectively denoise seismic data is a key step to improve the signal-to-noise ratio of seismic data in seismic data processing. The noise of seismic data is mainly divided into regular noise and random noise. Regular noise has deterministic characteristics and can be effectively suppressed according to the mechanism of noise formation, while random noise only has statistical characteristics and has no deterministic distribution form. It is relatively difficult to suppress noise. At present, common random noise suppressi...

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
IPC IPC(8): G01V1/36
CPCG01V1/36G01V2210/32
Inventor 陶永慧张兵杜泽源
Owner CHINA PETROLEUM & CHEM CORP
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap