Removing multiple types of noise of distributed acoustic sensing seismic data using attention-guided denoising convolutional neural network

Cong Wang, Xingguo Huang, Yue Li, Kristian Nørgaard JensenFrontiers in Earth Science 2023 · DOI

Abstract

Distributed optical fiber acoustic sensing (DAS) technology is increasingly vital for vertical seismic profile (VSP) exploration due to its high spatial resolution and resilience to extreme conditions. However, DAS seismic data is often compromised by complex noise types, including random, optical abnormal, fading, and horizontal noise. These overlapping noise profiles challenge traditional denoising methods. This paper introduces ADNet, an attention-guided denoising convolutional neural network designed to effectively improve the signal-to-noise ratio (SNR) of DAS data. By integrating a sparse block (SB), feature enhancement block (FEB), attention block (AB), and reconstruction block (RB), ADNet extracts global features and hidden noise information while preserving weak seismic signals. Results from synthetic and field data demonstrate that ADNet outperforms conventional methods and standard deep learning networks in both signal retrieval and noise suppression.

Introduction: The DAS Noise Challenge

Distributed Acoustic Sensing (DAS) has revolutionized seismic exploration by converting standard fiber-optic cables into dense arrays of high-resolution sensors. Despite these advantages, DAS data is notoriously susceptible to a "cocktail" of complex noise. Unlike traditional geophones, DAS sensors encounter fading noise and persistent horizontal bands that often overlap with actual seismic signals in the frequency domain.

Traditional methods like band-pass filtering or wavelet transforms often struggle to separate these signals without significant loss of data fidelity. Existing deep learning models, such as DnCNN, frequently fail to capture the global context required to identify long-range noise patterns or preserve the subtle reflected waves necessary for accurate subsurface imaging.

Methods: The ADNet Architecture

To address these limitations, the researchers developed ADNet, a 17-layer architecture optimized for seismic denoising. The network is structured into four functional blocks:

  1. Sparse Block (SB): Utilizing a combination of standard and dilated convolutions, the SB expands the network's receptive field. This allows the model to identify broad, sprawling noise structures, such as horizontal bands, without a massive increase in parameters.
  2. Feature Enhancement Block (FEB): This block fuses global and local features through concatenation, ensuring the network retains original context as it processes deeper layers.
  3. Attention Block (AB): The AB acts as the network's "brain," weighing different features to focus on hidden noise information buried within complex seismic backgrounds.
  4. Reconstruction Block (RB): ADNet utilizes residual learning. Instead of mapping noisy data directly to clean data, the network predicts the noise itself. The final denoised image is produced by subtracting the predicted noise from the original input.
The full ADNet architecture pipeline, illustrating the sequence of Sparse, Feature Enhancement, Attention, and Reconstruction blocks.

Training and Evaluation

Due to the lack of standard "clean" DAS datasets, the authors constructed a robust training set using elastic wave equations to create 5,000 pure seismic patches from 10 complex 2-D geological models. Realistic noise—including random, fading, and optical abnormal noise extracted from actual field records—was then injected into these patches. The model was evaluated against synthetic records and a field DAS record from the Tarim Basin in Xinjiang, China.

Results and Discussion

ADNet demonstrated superior performance compared to Variational Mode Decomposition (VMD), standard band-pass filtering, and the DnCNN model:

  • Signal-to-Noise Ratio (SNR): On synthetic data, ADNet achieved an SNR of 15.33 dB, significantly higher than DnCNN (11.95 dB) and traditional filtering (2.82 dB).
  • Structural Similarity (SSIM): ADNet maintained a higher SSIM of 0.786, indicating superior preservation of original seismic waveforms compared to DnCNN’s 0.623.
  • Field Performance: In real-world tests, ADNet effectively removed heavy horizontal noise, revealing critical "up-going" reflected waves that other methods missed.
Evaluation of denoising performance (SNR and SSIM) relative to training patch size, showing consistent improvement as the patch size increases.

Conclusion

ADNet provides a powerful solution for the simultaneous suppression of multiple DAS noise types. By combining dilated convolutions with an attention mechanism, the network achieves high-fidelity denoising while remaining computationally efficient. While "ringing noise" remains a challenge due to its similarity to weak reflected signals, future refinements in training set diversity and architecture are expected to further enhance the fidelity of DAS data processing.

Summary updated

Abstract

In recent years, distributed optical fiber acoustic sensing (DAS) technology has been increasingly used for vertical seismic profile (VSP) exploration. Even though this technology has the advantages of high spatial resolution, strong resistance to high temperature and pressure variations, long sensing distance, DAS seismic noise has expanded from random noise to optical abnormal noise, fading noise and horizontal noise, etc. This seriously affects the quality of the seismic data and brings huge challenges to subsequent imaging, inversion and interpretation. Moreover, the noise is more complex and more difficult to simultaneously suppress using traditional methods. Therefore, for the purpose of effectively improving the signal-to-noise ratio (SNR) of DAS seismic data, we introduce a denoising network named attention-guided denoising convolutional neural network (ADNet). The network is composed of four blocks, including a sparse block (SB), a feature enhancement block (FEB), an attention block (AB) and a reconstruction block (RB). The network uses different kinds of convolutions alternately to enlarge the receptive field size and extract global feature of the input. Meanwhile, the attention mechanism is introduced to extract the hidden noise information in the complex background. The network predicts the noise, and denoised data are obtained by subtracting the predicted results from the noisy inputs. In addition, we uniquely construct a large number of complex forward models for pure seismic data training set to enhance the network suitability. The combination design improves the denoising performance and reduces computational cost and memory consumption. The results obtained from both synthetic- and field data illustrate that the network has the ability to denoise the seismic images and retrieve weak effective signals better than conventional methods and common networks.
BibTeX
@inproceedings{wang2023removing,
  author = {Cong Wang and Xingguo Huang and Yue Li and Kristian Nørgaard Jensen},
  title = {{Removing multiple types of noise of distributed acoustic sensing seismic data using attention-guided denoising convolutional neural network}},
  booktitle = {Frontiers in Earth Science},
  year = {2023},
  doi = {10.3389/feart.2022.986470},
}
Figure