An Adaptive Magnetization Vector Inversion Method Converging Neighborhood Information Constraints and Magnetic Moment Allocation

Magnetic inversion can effectively obtain the spatial distribution characteristics of underground targets. However, due to the inherent non-uniqueness characteristics of magnetic inversion and the complex geological environment, the conventional regularization method still has some limitations in improving the inversion accuracy, especially in the presence of significant remanent magnetization, which results in anomaly spreading in inversion results, sudden changes in magnetic moments, isolated magnetic anomaly cells, and magnetization cavity. Therefore, this paper proposes an adaptive magnetization vector inversion method that converges neighborhood information constraints and magnetic moment allocation. This method dynamically analyzes the magnetic moment state of each grid cell and its neighborhood during the iteration process, allowing for adaptive adjustment of constraint strength to improve the inversion accuracy. In addition, based on the statistical characteristics of magnetic moment modulus, the locally anomalous magnetic moment is identified and its spatial distribution is optimized by means of magnetic moment allocation strategy, thereby enhancing the physical consistency of the reconstructed targets. Simulation and experimental results show that the proposed method can significantly improve the accuracy of inversion.