| As China’s economy continues to rise in the new century,demand for various mineral resources has increased sharply,presenting new requirements and challenges in mineral resource exploration.Currently,most of China’s mineral resources are mined at depths of less than 500 meters.The resources located at this depth have been gradually explored and depleted after years of development.Deep resource exploration technology has made unprecedented progress with the launch of the national"deep land"strategic action.As an artificial source exploration method,controlled source electromagnetic method has been widely used in resource exploration.As the exploration targets become deeper and the geological environment becomes more complex,the technical requirements for this method have also increased,and 3D fine inversion is becoming more and more important.Currently,the mainstream inversion method is mostly smooth constrained inversion,and the inversion results are generally smooth.However,the resolution loss of the boundary of the anomalous body is large,which is basically reliable for reflecting the large geological structure background.For anomalous areas with relatively complex structures,especially sharp boundaries of block models,the resolution and accuracy are difficult to achieve fine exploration requirements,thereby affecting the geological interpretation work.This paper introduces the sparse transformation method in image signal processing aim at the insufficient resolution and accuracy of controllable source electromagnetic inversion.This study perform a sparse domain based inversion by constructing a new inversion objective function combined with the sparse transformation.Firstly,this paper introduces the research status of 3D forward and inverse of controllable source electromagnetic method at home and abroad,the development of sparse transform and its application in geophysics and derives the theoretical formula of 3D forward vector finite element method and inversion algorithm based on nonlinear conjugate gradient method in detail.Meanwhile,this paper introduces the basic theory and mathematical realization process of sparse transform and analyzes the characteristics of different sparse transform and the characteristics of multi-scale sparse domain regularization inversion.On the basis of previous studies,a new inversion objective function is constructed with the mixed norm of L2-L1.For the data fitting term,the L2 norm constraint is used,and for the model constraint term,the L1 norm constraint is used to ensure the stable convergence of the inversion while taking into account the resolution,and the advantages of both are fully exploited.In addition,based on the multi-scale and multi-directional sparse properties of the Shearlet transform and its ability to provide optimal sparse representation of target bodies,the spatial domain model parameters in inversion are converted to Shearlet sparse domain coefficients to extract the boundary features of target bodies.Then,the sparse transform based frequency-domain three-dimensional controlled source electromagnetic inversion theory is derived in detail,and the algorithm is implemented by programming.Finally,different models are designed to test the algorithm.The results show that The research results of this paper provide some experience for promoting the fine inversion of geophysical electromagnetic method. |
