Structure-guided impedance inversion based on MTV regularization

Abstract Seismic impedance inversion is a reliable technique for estimating subsurface rock attributes from poststack data. However, the efficacy of inversion results is often compromised by data noise, initial model, and model constraints. Traditional single trace inversions typically generate the inversion results with spatial discontinuities, especially when the noise level is high, due to the lack of geometric constraints. To enhance the imaging quality, we introduce the structural constraints into the modified total variation (MTV) regularization scheme, developing the Structure-Guided MTV (SGMTV) regularization inversion method. Numerical examples demonstrate that the proposed SGMTV inversion method can effectively reconstruct multitrace impedance parameters simultaneously, with enhanced structural features and less model noise.