磁化
地质学
核磁共振
计算机科学
物理
磁场
量子力学
作者
Yiming Shu,Shuang Liu,Jiayong Yan,Zhenhua Zhou,Hongzhu Cai,Xiangyun Hu
出处
期刊:Geophysics
[Society of Exploration Geophysicists]
日期:2025-06-15
卷期号:: 1-75
标识
DOI:10.1190/geo2024-0555.1
摘要
The induced magnetization and the natural remanent magnetization comprise the total magnetization of the magnetic source. Magnetization vector provides comprehensive magnetization information on subsurface rocks, facilitating the interpretation of the structure and remanent magnetization of magnetic minerals. We propose the vector imaging framework to recover the magnetization vector information from surface or airborne magnetic data based on the Depth from Extreme Points (DEXP) imaging. We prove that the extreme value regions of the total field anomaly imaging will deviate from the true locations of the magnetic source in the presence of remanent magnetization. The DEXP imaging of the total magnitude anomaly (TMA) and normalized source strength (NSS) is utilized to mitigate the influence of remanent magnetization and oblique magnetization, yielding images approximating the source distribution. The source images are endowed with corresponding physical properties and transformed into three components of the magnetic source via the solution of the magnetic vector equation, assuming uniform magnetization directions for isolated sources. The focusing function accelerates model convergence, refining the total field anomaly fits through multiple iterations. The approach is investigated through applications to the synthetic examples and the aeromagnetic data from the Weilasito region (North China). The magnetic source distribution obtained from vector imaging aligns well with the geological data in the area and the magnetization directions are consistent with the results from various estimation methods. Compared with magnetization vector inversion, DEXP vector imaging provides a small number of possible magnetization direction options while achieving faster recovery of the magnetization vector distribution models.
科研通智能强力驱动
Strongly Powered by AbleSci AI