Seismic data denoising based on attention dual dilated CNN
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.This study introduces an Attention Dual-Dilated Convolutional Neural Network (ADDC-Net) for seismic data denoising. ADDC-Net effectively suppresses random noise while preserving crucial seismic signals for better subsurface analysis.
Area Of Science
- Geophysics
- Signal Processing
- Machine Learning
Background
- Seismic data denoising is crucial for accurate seismic exploration and interpretation.
- Traditional noise suppression methods can degrade essential subsurface signal data.
- Effective random noise reduction remains a challenge in seismic data processing.
Purpose Of The Study
- To develop an advanced deep learning model for enhanced seismic data denoising.
- To address the limitations of existing methods in preserving signal integrity.
- To improve the characterization of subsurface structures through noise reduction.
Main Methods
- Introduction of the Attention Dual-Dilated Convolutional Neural Network (ADDC-Net).
- Utilizing expanded model width for complementary feature extraction.
- Incorporating dilated convolution to increase receptive field and attention mechanisms for signal preservation.
Main Results
- ADDC-Net demonstrated superior performance compared to DnCNN and DudeNet.
- Achieved significant improvements in Peak Signal-to-Noise Ratio (PSNR) by 2.8905 dB and 0.6410 dB, respectively.
- Showcased faster processing speeds than comparable convolutional networks.
Conclusions
- ADDC-Net offers a robust solution for random noise suppression in seismic data.
- The proposed network effectively preserves critical seismic signals, aiding subsurface interpretation.
- ADDC-Net represents a significant advancement in seismic data processing techniques.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
