A joint inversion-segmentation approach to assisted seismic interpretation

Structural seismic interpretation and quantitative characterization are historically intertwined processes. The latter provides estimates of properties the subsurface which can be used to aid structural alongside original data a number other attributes. In this work, we redefine process as inverse problem tries jointly estimate (i.e., acoustic impedance) piece-wise segmented representation based on user-defined macro-classes. By inverting for quantities simultaneously, inversion is primed with prior knowledge about regions interest, whilst at same time it constrains belief actual measurements. As proposed functional separable in two quantities, these optimized an alternating fashion, where each subproblem solved using Primal-Dual algorithm. Subsequently, class input workflow aims extract perimeter detected shapes produce unique horizons. effectiveness method illustrated through numerical examples synthetic field datasets.