Abstract:In light of the challenging conditions of exploration environments coupled with escalating exploration expenses, seismic data acquisition frequently entails the capturing of signals entangled amidst diverse noise interferences and instances of data loss. The unprocessed state of these seismic signals significantly jeopardizes the interpretative phase. Evidently, the integration of attention mechanisms and the utilization of generative adversarial networks (GANs) have emerged as prominent techniques within signal processing owing to their adeptness in discerning intricate global dependencies. Our research introduces a pioneering approach for reconstructing and denoising seismic signals, amalgamating the principles of self-attention and generative adversarial networks—hereafter referred to as SAGAN. Notably, the incorporation of the self-attention mechanism into the GAN framework facilitates an enhanced capacity for both the generator and discriminator to emulate meaningful spatial interactions. Subsequently, leveraging the feature map generated by the self-attention mechanism within the GAN structure enables the interpolation and denoising of seismic signals. Rigorous experimentation substantiates the efficacy of SAGAN in simultaneous signal interpolation and denoising. Initially, we benchmarked SAGAN against prominent methods such as UNet, CNN, and Wavelet for the concurrent interpolation and denoising of two-dimensional seismic signals manifesting varying levels of damage. Subsequently, this methodology was extended to encompass three-dimensional seismic data. Notably, performance metrics reveal SAGAN's superiority over comparative methods. Specifically, the quantitative tables exhibit SAGAN's pronounced advantage, with a 3.46% increase in PSNR value over UNet and an impressive 11.90% surge compared to Wavelet. Moreover, the RMSE values affirm SAGAN's robust performance, showcasing an 11.54% reduction in comparison to UNet and an impressive 29.27% decrement relative to Wavelet, hence unequivocally establishing the SAGAN method as a preeminent choice for seismic signal recovery.

Relative Attributes-Based Generative Adversarial Network for Desert Seismic Noise Suppression

The Application of Semisupervised Attentional Generative Adversarial Networks in Desert Seismic Data Denoising

A Novel Generative Adversarial Network for the Removal of Noise and Baseline Drift in Seismic Signals

Noise suppression of distributed fiber-optical acoustic sensing seismic data by attention-guided multiscale generative adversarial network

A Global and Multiscale Denoising Method Based on Generative Adversarial Network for DAS VSP Data

Attribute-Based Double Constraint Denoising Network for Seismic Data

Desert Seismic Low-Frequency Noise Attenuation Using Low-Rank Decomposition-Based Denoising Convolutional Neural Network

Unpaired training: Optimize the seismic data denoising model without paired training data

Seismic Shot Gather Denoising by Using a Supervised-Deep-Learning Method with Weak Dependence on Real Noise Data: A Solution to the Lack of Real Noise Data

Self-Attention Generative Adversarial Network Interpolating and Denoising Seismic Signals Simultaneously

MAE-GAN: A Novel Strategy for Simultaneous Super-resolution Reconstruction and Denoising of Post-stack Seismic Profile

Enhancing Seismic Data Denoising Through Multiscale Analysis Across Transformer and GAN

Self-Supervised Attenuation Method Based on Similarity Comparisons for 3D Seismic Random Noise

Seismic Noise Attenuation Using Unsupervised Sparse Feature Learning

Denoising of seismic data in desert environment based on a variational mode decomposition and a convolutional neural network

Adaptive Graded Denoising of Seismic Data Based on Noise Estimation and Local Similarity

Low-Frequency Noise Suppression Method Based on Improved DnCNN in Desert Seismic Data

Multiple Attention Mechanisms-Based Convolutional Neural Network for Desert Seismic Denoising

Unsupervised 3-D Seismic Erratic Noise Attenuation With Robust Tensor Deep Learning

Attention-Based Neural Network for Erratic Noise Attenuation From Seismic Data With a Shuffled Noise Training Data Generation Strategy

Self-Supervised Seismic Random Noise Attenuation With Spatial Attention From a Single Section