Potential Field Inversion On Nodes For Stochastic Inversion Models

Instead of applying gravity inversion on a large number of small prisms which is computationally very expensive, it is proposed to perform inversion on a limited number of nodes which include all the available point density data (if any) plus points sampling the inversion space fairly. It is also possible to include sets of linear constraints in the inversion. If no noise is present, the inverted field reproduces exactly the added constraints and the gravity field. These methods can be applied to the magnetic potential as well as the joint inversion of gravity-magnetic field data.