A Near-Field Magnetization Vector Inversion Method Based on Direction Constraints

Magnetic anomaly data of underground magnetic targets can be acquired using accurate and reliable magnetic field measurement systems. Then magnetic inversion techniques can be applied to analyze and interpret the measured data, ultimately revealing the magnetization characteristics and spatial location of the subsurface targets. Compared with the traditional scalar magnetic susceptibility inversion, magnetization vector inversion (MVI) is especially suitable for dealing with magnetic anomalies with remanent magnetic influence. However, MVI involves the inversion of three components and thus has a stronger inversion non-uniqueness. Focusing on the inversion of near-field magnetic anomaly measurement data, this paper firstly analyzes the limitations of traditional depth weighting function in near-field magnetic inversion. Then for the non-uniqueness of MVI, a nearfield magnetic inversion optimization method based on direction constraints is proposed, taking into account that the direction of the magnetization vectors among the grid cells composing the magnetic target should be consistent. On the one hand, the method constructs a direction consistency constraints matrix, which constrains the magnetization directions of adjacent cells by quantifying the magnetization direction differences among grid cells. On the other hand, the main magnetization directions in the neighborhood of the grid cells are extracted by combining the local principal component analysis (PCA) method during the iterative solution process. The magnetization vectors of the grid cells are corrected by the adaptive weighted projection strategy, and the direction consistency among the grid cells is gradually enhanced. The inversion results on both simulated and measured data show that the proposed method can accurately invert targets even under complex geological conditions. This provides technical support for interpreting magnetic field measurement data.