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.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in seismic signal processing, due to the harsh exploration environment and high exploration costs, the collected seismic signals are often mixed with various noise interferences and there is data loss. The unprocessed seismic signals seriously affect the subsequent interpretation stage. Therefore, this paper aims to propose a new method, combining self - attention mechanism and generative adversarial network (GAN), to interpolate and denoise seismic signals simultaneously in order to improve the quality and integrity of the signals. Specifically, the goals of the paper include: 1. **Improve the reconstruction quality of seismic signals**: By introducing the self - attention mechanism, enhance the ability of the generator and discriminator to capture complex spatial dependencies, so as to more effectively recover the missing seismic signals. 2. **Achieve synchronous interpolation and denoising**: Develop a method that can perform interpolation and denoising simultaneously, avoiding the additional time and computational costs brought by training two independent networks separately. 3. **Improve the generalization ability of the method**: Ensure that the proposed method is not only applicable to 2D seismic signals, but also can handle 3D seismic signals and perform well on data with different degrees of damage. To verify the effectiveness of this method, the author conducted a large number of experiments and compared the proposed SAGAN model with existing methods such as UNet, CNN and Wavelet. The results show that SAGAN has significant advantages in performance indicators such as PSNR (Peak Signal - to - Noise Ratio) and RMSE (Root Mean Square Error). ### Formula Summary 1. **Calculation of the attention map in the self - attention mechanism**: \[ s_{i,j} = A(x_i)^T B(x_j) \] \[ \beta_{j,i} = \frac{\exp(s_{i,j})}{\sum_{i = 1}^{N} \exp(s_{i,j})} \] \[ o_j = d\left(\sum_{i = 1}^{N} \beta_{j,i} C(x_i)\right), \quad C(x_i) = W_C x_i, \quad d(x_i) = W_d x_i \] 2. **Loss functions of the generator and discriminator**: \[ L_G = -E_{z \sim p_z, y \sim p_{signal}}[D(G(z), y)] \] \[ L_D = -E_{(x,y) \sim p_{signal}}[\min(0, - 1+D(x, y))] - E_{z \sim p_z, y \sim p_{signal}}[\min(0, - 1 - D(G(z), y))] \] These formulas show the core calculation process of the SAGAN model, ensuring its efficiency and accuracy in processing complex seismic signals.

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

Accelerating CS-MRI Reconstruction with Fine-Tuning Wasserstein Generative Adversarial Network

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

STUGAN: An Integrated Swin Transformer-Based Generative Adversarial Networks for Seismic Data Reconstruction and Denoising

Enhancing Seismic Data Denoising Through Multiscale Analysis Across Transformer and GAN

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

SHBGAN: Hybrid Bilateral Attention GAN for Seismic Image Super-Resolution Reconstruction

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

SWDGAN: GAN-based Sampling and Whole Image Denoising Network for Compressed Sensing Image Reconstruction

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

Self-Attention Generative Adversarial Networks

Super-Resolution Reconstruction of Seismic Images Based on Deep Residual Channel Attention Mechanism

Distorted underwater image reconstruction for an autonomous underwater vehicle based on a self-attention generative adversarial network

Seismic Surface-Related Multiples Suppression Based on SAGAN

Structural dynamic response reconstruction using self-attention enhanced generative adversarial networks

Efficient Seismic Data Denoising via Deep Learning With Improved MCA-SCUNet

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

Self-adaptive denoising net: Self-supervised learning for seismic migration artifacts and random noise attenuation

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

Seismic data recovery using deep targeted denoising priors in an alternating optimization framework