An application of sparse inversion on the calculation of the inverse data space of geophysical data

Multiple reflections as observed in seismic reflection measurements often hide arrivals from the deeper target reflectors and need to be removed. The inverse data space provides a natural separation of primaries and surface-related multiples, as the surface multiples map onto the area around the origin while the primaries map elsewhere. However, the calculation of the inverse data is far from trivial as theory requires infinite time and offset recording. Furthermore regularization issues arise during inversion. We perform the inversion by minimizing the least-squares norm of the misfit function and by constraining the ℓ
₁
norm of the solution, being the inverse data space. In this way a sparse inversion approach is obtained. We show results on field data with an application to surface multiple removal.