Repeatability enhancement of time-lapse seismic data via a convolutional autoencoder

Jun, Hyoung Oh; Cho, Yongchae

doi:10.1093/gji/ggab397

Cited by 13 publications

(8 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It also deals only with seismic data, unlike the work of Yuan et al (2020) andZhou et al (2019) that inherently require the network to learn the physics of mapping the data to velocities and the CO 2 mass. Our approach is close to the work implemented by Jun and Cho (2021) except they only focus on reducing the 4D noise in a local image batch and their method is prone to fail when a large time shift exists between the vintages.…”

Section: Introductionmentioning

confidence: 94%

“…These studies require generating training datasets that is diverse enough to generalize the network, which is a challenging task as many variables need to be considered. Jun and Cho (2021) improved the repeatability of post-stack seismic images by training the network to remove the non-repeatable noise in the part of the data that do not have reservoir information (like early arrivals), then inferring on the reservoir area.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Time-lapse data matching using a recurrent neural network approach

et al. 2022

View full text Add to dashboard Cite

Time-lapse seismic data acquisition is an essential tool to monitor changes in a reservoir due to fluid injection, such as CO2 injection. By acquiring multiple seismic surveys in the exact same location, we can identify the reservoir changes by analyzing the difference in the data. However, such analysis can be skewed by the near-surface seasonal velocity variations, inaccuracy and repeatability in the acquisition parameters, and other inevitable noise. The common practice (cross-equalization) to address this problem uses the part of the data where changes are not expected to design a matching filter and then apply it to the whole data, including the reservoir area. Like cross-equalization, we train a recurrent neural network on parts of the data excluding the reservoir area and then infer the reservoir-related data. The recurrent neural network can learn the time dependency of the data, unlike the matching filter that processes the data based on the local information obtained in the filter window. We demonstrate the method of matching the data in various examples and compare it with the conventional matching filter. Specifically, we start by demonstrating the ability of the approach in matching two traces and then test the method on a pre-stack 2D synthetic data. Then, we verify the enhancements of the 4D signal by providing RTM images. We measure the repeatability using normalized root mean square and predictability metrics and show that in some cases, our proposed method performed better than the matching filter approach.

show abstract

Section: Introductionmentioning

confidence: 94%

Section: Introductionmentioning

confidence: 99%

Time-lapse data matching using a recurrent neural network approach

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Recently, studies applying machine learning to seismic noise attenuation have been performed to enhance the efficiency of the noise attenuation process. The major research topics of interest include defining or improving the network architecture (Saad & Chen, 2020;Sun et al, 2022;Jun & Cho, 2022) and attenuating certain types of noise (X. Zhao et al, 2019;Yuan et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

Loss functions in machine learning for seismic random noise attenuation

Jun,

Kim

2023

Geophysical Prospecting

Self Cite

View full text Add to dashboard Cite

Seismic random noise is one of the main factors that degrade the quality of seismic data. Therefore, seismic random noise attenuation should be performed appropriately through several stages during seismic data processing, and this requires sufficient experience and knowledge because the proper hyperparameters need to be determined based on the features of the noise in the target seismic data. Recently, machine learning‐based seismic noise attenuation has been widely studied because it provides suitable results by learning noise features from noisy data, unlike conventional physics‐based approaches. There are many important factors in the machine learning, and here, we focus on the loss functions of machine learning in terms of seismic random noise attenuation. The most widely used loss function is l2, but we train a model with various kinds of single and multiple loss functions and attenuate seismic random noise. We analyze the efficiency of loss functions by comparing the noise attenuated results of synthetic and field seismic data qualitatively and quantitatively. Our analysis indicates that the multiple loss function with l1 norm can be a proper choice for random noise suppression of seismic data.This article is protected by copyright. All rights reserved

show abstract

“…Walker et al, 2021) and 4) improving the repeatability of specific workflows (e.g. Jun & Cho, 2022). There has been, to date, no empirical consideration of the extent to which the existing publications and published work are reproducible.…”

Section: Introductionmentioning

confidence: 99%

How reproducible and reliable is geophysical research? A review of the availability and accessibility of data and software for research published in journals

Ireland¹,

Algarabel²,

Steventon³

et al. 2023

Preprint

View full text Add to dashboard Cite

Geophysical research frequently makes use of agreed methodologies, formally published software, and bespoke code to process and analyse data. The reliability and repeatability of these methods is vital in maintaining the integrity of research findings and thereby avoiding the dissemination of unreliable results. In recent years there has been an increased attention on aspects of reproducibility, which includes data availability, across scientific disciplines. This review considers aspects of reproducibility of geophysical studies relating to their publication in peer reviewed journals. Specifically, it considers: 1) the extent to which reproducibility in geophysics is the focus of published literature; 2) journal policies on the requirements for providing code, software, and data for submission and, 3) the availability of code, software and data associated for existing journal articles. The findings show that: 1) between 1991 and 2021 there were 27 articles with reproducibility in the title and 222 with reliability, with a year on year increases in both over the same period; 2) while 60% of journals have a definition of research data, only 20% of journals have a requirement for a data availability statement, and 3) despite ~86% of randomly sampled journal articles including a data availability statement, only 54% of articles have the original data accessible via data repositories or web servers. It is suggested that despite journals and authors working towards improving the availability of data, frequently these data are not easily accessible, therefore limiting possibility of reproducing studies.

show abstract

Repeatability enhancement of time-lapse seismic data via a convolutional autoencoder

Cited by 13 publications

References 37 publications

Time-lapse data matching using a recurrent neural network approach

Time-lapse data matching using a recurrent neural network approach

Loss functions in machine learning for seismic random noise attenuation

How reproducible and reliable is geophysical research? A review of the availability and accessibility of data and software for research published in journals

Contact Info

Product

Resources

About