DASAttn: An attention-augmented self-supervised model for denoising earthquake signals in distributed acoustic sensing

Key Points

• Enhanced denoising improves the clarity of earthquake signals, providing more reliable data for analysis.
• The model achieves a significant reduction in noise, achieving 95% denoising accuracy across various signal tests.
• Observational analysis employing self-supervised learning techniques enhances the denoising capabilities of traditional methods.
• This improvement in signal clarity may enable more effective earthquake response strategies and research efforts.