Resolution analysis and enhancement in geophysical imaging and inversion

Abstract

Geophysical imaging and inversion will always be associated with some degree of blurring and resolution loss. In this doctoral thesis project, we have therefore assessed different methods of analysing and improving the resolution of seismic and electromagnetic images. As a first step, we investigated holistic migration as a method to achieve high-resolution images from severely undersampled seismic data, and it was found that such an approach is feasible if diffraction-separated data are used. Next, we investigated the feasibility of undersampling Controlled Source Electromagnetic (CSEM) data while retaining the resolution of the inversion. As there are major differences between the imaging process of seismic and CSEM data, we proposed subsampling based on quantities derived from well-known inversion theory. This study demonstrated that a typical CSEM survey is associated with significant data redundancy, and that the most important datapoints can be quantified through the so-called resolution matrices. Due to the nature of electromagnetic wave propagation, the inversions associated with electromagnetic sounding represent a blurred image of the true earth model. Thus, the final part of this doctoral thesis project included developing a point spread function inversion, which solves a deblurring problem to recover the high-resolution image.