Abstract：According to the change of physical properties,density inversion can delineate complex source distributions and simulate complex geological bodies.So it has become an important approach for solving geologic problems with gravity method.In this paper,2-D density non-linear inversion is developed using the RBF(Radical Basis Function) neural network method,which has the advantage of nonlinear mapping and better generalization capabilities.This method is tested on both synthetic and field data sets.The results show that the proposed method is effective and robust.In addition,the structures and parameters of the network,as well as noise effects on the inversion are discussed in this paper.
