Source-independent full-waveform inversion using an amplitude-semblance objective function

Full-waveform inversion is a promising tool to produce accurate and high-resolution subsurface elastic models. Accurate estimation of the source wavelet is required for most full-waveform inversion algorithms to obtain reliable inversion results. We develop a novel source-independent waveform inversion method to avoid source wavelet estimation in full-waveform inversion. We define an amplitude-semblance objective function in full-waveform inversion to remove the source wavelet effect. Compared with other source-independent FWI methods, our new method does not require an optimized reference trace. Meanwhile, our approach can better handle the inconsistence of source wavelets for different sources. We verify our method using the Marmousi model. The results show that our source-independent full-waveform inversion can produce accurate velocity models even if the source signature is incorrect.