Seismic data denoising is a crucial and challenging task for high-quality seismic exploration. Recent advancements in deep learning methods have demonstrated promising results in seismic denoising. However, the acquisition of ground truth data required for training remains unavailable, especially in field tests. We propose an unsupervised deep denoiser called the iterative diffusion denoising model (IDDM) based on a diffusion model to remove random noise. We present the diffusion process of IDDM according to the seismic noise model and the two-stage reverse process to iteratively train a deep restorer with the data pair created solely from the observed data. Hence, the IDDM is learned to approximate the reverse diffusion process of the seismic data, which leads to the effective seismic signal recovery and robustness to the variant noise level and complex distribution of the field seismic data. Moreover, the invariant features between adjacent states are introduced to the generative denoising model by the signal preserving module, enabling IDDM to gradually recover the effective seismic signals in high fidelity while thoroughly suppressing noise using only noisy data. The proposed approach shows excellent denoised results in synthetic and field data tests at low signal-to-noise ratios (SNRs), demonstrating its potential for practical applications in seismic data processing.

Seismic interpretation is crucial in seismic exploration to identify geological structures in the field. However, interpretation is often challenging due to inherent low-resolution (LR) limitations, and acquiring high-quality data is expensive with no guaranteed fidelity. To address these challenges, we propose a novel supervised deep learning-based method to enhance seismic data resolution by simultaneously focusing on dominant wavelet frequency and the spatial domain. First, we employ a deep learning module to estimate the location of low-frequency wavelets as semantic information, providing a prior that allows the model to process these signals and precisely enhance the dominant frequency. Subsequently, we use a feature wrapper to integrate the LR data with the semantic priors. We use them as conditions for a diffusion model to generate high-frequency features, considered priors with higher-frequency wavelets. Finally, we input the priors and the LR data into a feature fusion module (FFM) to generate the final output, doubling the sampling points and traces to achieve high-resolution (HR) data with high-frequency wavelets. The models are trained separately using cross-entropy, Kullback-Leibler divergence, and Charbonnier loss. The priors we use and the design of the generative diffusion model ensure high fidelity, preventing false seismic events and over-smoothing. Experiments on field data demonstrate the superiority of our method compared to three other approaches, highlighting its potential as a powerful seismic super-resolution tool in practical applications.

Recent advancements in deep learning (DL) methods have demonstrated promising results in distributed acoustic sensing-vertical seismic profiling (DAS-VSP) denoising. However, the satisfying noise-attenuating performance requires the training dataset to be comprehensive of noise types, which is difficult to achieve in real-life scenarios. To tackle this problem, we propose a universal DL denoiser called triplet masking denoiser (Tri-MD) based on a triple different data masking scheme in training and interfering. In the training stage, Tri-MD applies random masking to the data to force the denoising model to perform the signal recovery task rather than simply separating the signal from the noise. The characteristic allows Tri-MD to transform the complex noise suppression and signal preservation tasks. Thus, the noise-attenuating performance of our method does not suffer from degradation due to the scarcity of noise types in the training set. Moreover, we introduce a subnetwork for segmentation to estimate the signal location and create a semantic mask of the seismic effective signal, which helps the network extract the signal features to supplement the original features lost during the random masking process and improve the ability to recover the signal. Furthermore, we propose a novel interfering paradigm that utilizes an attention mask to ensure the test procedure is consistent with the training to enhance the robustness of Tri-MD. Noise reduction experiments on synthetic and two field DAS-VSP data demonstrate that our method can achieve excellent noise-attenuating performance in tests even when training is done on training datasets with scarce noise variety, proving that the generalization ability of Tri-MD is satisfying.

Seismic random noise attenuation is an essential procedure when processing seismic data. Due to various acquisition environments and complex geological conditions, random noise in seismic field data exhibits spatiotemporal levels, significantly increasing the difficulty of extracting seismic signals. The deep learning (DL) methods have shown excellent performances for seismic denoising. However, most existing discriminative DL methods cannot fit field data with noise levels and signal structures that differ from the training set. To tackle the unmatching challenge, we propose a self-supervised deep denoising model named noise estimation-based convolution neural network (NE-CNN), which contains a multiscale denoising module as a pretrained model and a dual-path noise estimation module that estimates the noise level of each data patch with a generalized Gaussian distribution and gray-level co-occurrence matrix (GLCM). With Stein’s unbiased risk estimate (SURE), we can fine-tune the pretrained denoising module according to the estimated noise levels in a self-supervised style solely on data to be processed, leading to a boost of robustness to nonstationary seismic noise. In synthetic and field data tests, NE-CNN performed satisfactorily compared with discriminative DL methods in denoising effects on seismic data with nonstationary noise.

Seismic random noise suppression presents two main challenges: achieving thorough noise suppression while simultaneously ensuring complete restoration of effective signal content. However, due to the complexity of random noise, existing denoising methods often only achieve an awkward balance between removing random noise and restoring effective signals. In this paper, we propose a novel random noise suppression model based on downsampling and super-resolution. By decoupling the denoising and signal restoration processes, our method reduces the difficulty of addressing these two challenges and mitigates the likelihood of suboptimal results. On the one hand, the high fitting-capacity Downsampling network uses non-linear transformations to separate random noise and effective signals while purifying the high-order features of effective signals. On the other hand, the Super-resolution network expands the low-dimensional seismic signal content containing the high-order features of the signal to restore the signal structure. Moreover, we propose a new adversarial loss by introducing the gradient between the generated data and the real data, which enhances the perceptual quality of the super-resolution results and recovers the content of effective signals better. Because both subnetworks are not affected by signal/noise features during processing, the model exhibits strong fitting and generalization abilities. The experimental evaluation on four different types of seismic data demonstrates the superiority of our method in suppressing random noise and restoring the content of effective signals.