Parsimonious Seismic Tomography with Poisson Voronoi Projections: Methodology and Validation

• DOI: 10.1785/0220190141
• Publication Date: 2019-10-30T13:12:36Z
• Authors: Hongjian Fang; Robert D. van der...; Maarten V. de Hoop; Konik Kothari; Sidharth Gupta; Ivan Dokmanić
• Keywords: Smoothing; Regularization

Abstract Ill‐posed seismic inverse problems are often solved using Tikhonov‐type regularization, that is, incorporation of damping and smoothing to obtain stable results. This typically results in overly smooth models, poor amplitude resolution, a difficult choice between plausible models. Recognizing the average parameters can be better constrained than individual parameters, we propose tomography method stabilizes problem by projecting original high‐dimension model space onto random low‐dimension subspaces then infers high‐dimensional solution from combinations such subspaces. The formed functions constant Poisson Voronoi cells, which viewed as mean near certain location. low‐dimensional constrained, image reconstruction does not require explicit regularization. Moreover, recovered subsets whole dataset, increases efficiency offers opportunities mitigate uneven sampling space. final (high‐dimension) is obtained images different either solving another normal equation or simply averaging images. Importantly, uncertainty directly Synthetic tests show our outperforms conventional methods both terms geometry recovery. application southern California plate boundary region also validates robustness imaging geologically consistent features well strong along‐strike variations San Jacinto fault clearly seen methods.