A Robust DAS-to-Velocity Conversion Method Based on HOLp-OGS Regularized Inversion

The applications of Distributed Acoustic Sensing (DAS) have grown increasingly widespread in recent years. Unlike geophones, DAS measures the average strain (or strain rate) over a gauge length centered at each measurement point, resulting in the “gauge-length effect,” which attenuate high-frequency components. Consequently, a strain-to-velocity conversion is required for effective utilization and analysis. Additionally, DAS data typically exhibits low signal-to-noise ratio (SNR), further complicating accurate conversion process. To address these challenges, we propose a robust inversion framework that combines high-order Lp quasi-norm (HOLp) and overlapping group sparse (OGS) regularization for accurate and reliable DAS-to-velocity conversion. The HOLp quasi-norm enhances sharp transitions in the reconstructed velocity field, while the OGS regularization reduces global staircase artifacts, providing complementary benefits.

This combined approach improves noise resistance and enhances the fidelity of detail characterization, resulting in accurate and reliable strain-to-velocity conversion. Synthetic case studies demonstrate that, compared to the traditional norm constraint, the proposed method significantly improves conversion accuracy, even under low-SNR conditions, highlighting its effectiveness, robustness, and practical applicability in data conversion.