The inversion of gravity data can generate subsurface density models by minimizing a Tikhonov parametric functional. The Tikhonov parametric functional has a misfit functional and a stabilizing functional which the latter functional governs the solution to be smooth or piecewise-constant. Piecewise-constant models can discern sharp geological interfaces. Therefore the models are highly demanded especially for mineral exploration. Some stabilizing functionals such as minimum support (MS) and minimum gradient support (MGS) stabilizing functionals produce piecewise-constant and compact models so that they are governed by a focusing parameter that must be selected properly. Choosing a suitable focusing parameter complicates the inversion process. In this paper, the zero order minimum entropy stabilizing functional is proposed for focusing inversion of gravity data based on a reweighted regularized conjugate gradient algorithm. The new algorithm avoids the selection of an optimal focusing parameter and generates piecewise-constant models. The effectiveness of the algorithm is validated by inversion of synthetic gravity data and field gravity data from the San Nicolas deposit in Mexico.
