Self-Supervised Learning for Seismic Data Reconstruction and Denoising

Meng, F.L.; Fan, Qinqin; Li, Yue

doi:10.1109/lgrs.2021.3068132

Cited by 21 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are already some works based on self-supervised learning in seismic processing and interpretation. Such as seismic denoising, some researchers achieve random noise suppression based on Noise2Void (Krull et al, 2019), a kind of self-supervised method (Birnie et al, 2021;Birnie and Alkhalifah, 2022), some researchers achieve denoising through self-supervised seismic reconstruction (Meng et al, 2022), and some researchers achieve it by adding additive signal-dependent noise to the original seismic data and learn to predict the original data (Wu et al, 2022). What's more, self-supervised learning can also be used to reconstruct seismic data with consecutively missing traces (Huang et al, 2022), reconstruct the low-frequency components of seismic data (Wang et al, 2020), predict facies and other properties (Zhangdong and Alkhalifah, 2020), and perform other seismic processing tasks, like velocity estimation, first arrival picking, and NMO (normal moveout) (Harsuko and Alkhalifah, 2022).…”

Section: Introductionmentioning

confidence: 99%

Deep learning based on self-supervised pre-training: Application on sandstone content prediction

Wang

Wang²,

Chen³

et al. 2023

Front. Earth Sci.

View full text Add to dashboard Cite

Deep learning has been widely used in various fields and showed promise in recent years. Therefore, deep learning is the future trend to realize seismic data’s intelligent and automatic interpretation. However, traditional deep learning only uses labeled data to train the model, and thus, does not utilize a large amount of unlabeled data. Self-supervised learning, widely used in Natural Language Processing (NLP) and computer vision, is an effective method of learning information from unlabeled data. Thus, a pretext task is designed with reference to Masked Autoencoders (MAE) to realize self-supervised pre-training of unlabeled seismic data. After pre-training, we fine-tune the model to the downstream task. Experiments show that the model can effectively extract information from unlabeled data through the pretext task, and the pre-trained model has better performance in downstream tasks.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep learning based on self-supervised pre-training: Application on sandstone content prediction

Wang

Wang²,

Chen³

et al. 2023

Front. Earth Sci.

View full text Add to dashboard Cite

show abstract

“…Therefore, many deep learning algorithms have been used in the field of seismic data noise reduction. Owing to the lack of abundant clean-noisy seismic data pairs, self-supervised learning is the focus of research [10]. Qiu et al [11] designed a stopping criterion to train denoising generative network whose label is the noisy seismic data.…”

Section: Introductionmentioning

confidence: 99%

Self-supervised deep learning for multi-profile seismic data denoising

Chuanchao,

Song,

Zhang

et al. 2023

International Conference on Computer, Artificial Intelligence, and Control Engineering (CAICE 2023)

View full text Add to dashboard Cite

Self-supervised deep learning has been widely exploited in seismic data denoising. However, most of the current methods take a single seismic profile to reconstruct two new similar profiles by sampling. Then, the two new similar profiles are conducted as input and label respectively to train the model. In fact, there obviously exists similarity and more structural information among adjacent profiles. Therefore, we want to use adjacent multiple profiles to denoise seismic data. But this ideal requires that the neural network model can process multiple profiles. In order to solve this problem, we introduce a supervised model which can input five continuous profiles and denoise the middle profile. In addition, a self-supervised training strategy is proposed for the supervised model to train with no clean profiles. The experimental results on synthetic noise show that our method can achieve a higher Signal-to-Noise Ratio (SNR). According to the experiment on real noise, the proposed algorithm also obtains cleaner and smoother denoising profiles.

show abstract

“…An extension of this method (named structured noise to void) aims to attenuate noise along extended blind masks (Broaddus et al., 2020), in which the noise is deemed coherent. In seismic studies, adaptations of these methods are used for the attenuation of random noise (Meng et al., 2021; Birnie et al., 2021), trace‐wise coherent noise (Birnie & Alkhalifah, 2022; S. Liu et al., 2022; Abedi et al., 2023) and deblending (Wang, Hu, et al., 2022).…”

Section: Introductionmentioning

confidence: 99%

Semi‐blind‐trace algorithm for self‐supervised attenuation of trace‐wise coherent noise

Abedi,

Pardo,

Alkhalifah

2023

Geophysical Prospecting

View full text Add to dashboard Cite

Trace‐wise noise is a type of noise often seen in seismic data, which is characterized by vertical coherency and horizontal incoherency. Using self‐supervised deep learning to attenuate this type of noise, the conventional blind‐trace deep learning trains a network to blindly reconstruct each trace in the data from its surrounding traces; it attenuates isolated trace‐wise noise but causes signal leakage in clean and noisy traces and reconstruction errors next to each noisy trace. To reduce signal leakage and improve denoising, we propose a new loss function and masking procedure in a semi‐blind‐trace deep learning framework. Our hybrid loss function has weighted active zones that cover masked and non‐masked traces. Therefore, the network is not blinded to clean traces during their reconstruction. During training, we dynamically change the masks' characteristics. The goal is to train the network to learn the characteristics of the signal instead of noise. The proposed algorithm enables the designed U‐net to detect and attenuate trace‐wise noise without having prior information about the noise. A new hyperparameter of our method is the relative weight between the masked and non‐masked traces' contribution to the loss function. Numerical experiments show that selecting a small value for this parameter is enough to significantly decrease signal leakage. The proposed algorithm is tested on synthetic and real off‐shore and land datasets with different noises. The results show the superb ability of the method to attenuate trace‐wise noise while preserving other events. An implementation of the proposed algorithm as a Python code is also made available.This article is protected by copyright. All rights reserved

show abstract

Self-Supervised Learning for Seismic Data Reconstruction and Denoising

Cited by 21 publications

References 15 publications

Deep learning based on self-supervised pre-training: Application on sandstone content prediction

Deep learning based on self-supervised pre-training: Application on sandstone content prediction

Self-supervised deep learning for multi-profile seismic data denoising

Semi‐blind‐trace algorithm for self‐supervised attenuation of trace‐wise coherent noise

Contact Info

Product

Resources

About