欢迎访问《兵工学报》官方网站,今天是 分享到:

兵工学报 ›› 2018, Vol. 39 ›› Issue (11): 2202-2210.doi: 10.3969/j.issn.1000-1093.2018.11.014

• 论文 • 上一篇    下一篇

强剩磁条件下磁性目标三维正则化聚焦反演方法

李金朋1, 张英堂1, 范红波1, 李志宁1, 张光2   

  1. (1.陆军工程大学石家庄校区 车辆与电气工程系, 河北 石家庄 050003; 2.65185部队, 辽宁 铁岭 112611)
  • 收稿日期:2018-03-10 修回日期:2018-03-10 上线日期:2018-12-25
  • 通讯作者: 范红波(1982—),男,讲师 E-mail:ffhhbboo@163.com
  • 作者简介:李金朋(1991—),男,博士研究生。E-mail:18626648671@163.com
  • 基金资助:
    国家自然科学基金项目(51305454)

Three-dimensional Focusing Inversion of Magnetic Target in the Presence of Significant Remanence

LI Jin-peng1, ZHANG Ying-tang1, FAN Hong-bo1, LI Zhi-ning1, ZHANG Guang2   

  1. (1.Department of Vehicle and Electrical Engineering, Shijiazhuang Campus, Army Engineering University, Shijiazhuang 050003, Hebei, China; 2.Unit 65185 of PLA, Tieling 112611, Liaoning, China)
  • Received:2018-03-10 Revised:2018-03-10 Online:2018-12-25

摘要: 针对强剩磁条件下铁磁物质反演中存在磁化方向发生改变的问题,提出了强剩磁条件下磁性目标三维正则化聚焦反演方法。对于孤立磁源,首先估计其磁化方向,然后利用磁化方向估计值对磁性体进行反演;对于多目标磁源,利用弱敏感于磁化方向的磁总场模量数据进行反演。在迭代过程中,通过深度加权矩阵和最小支撑矩阵对经典Tikhonov正则化框架下的反演模型进行约束并得到目标函数,有效解决了反演解的多解性问题。对目标函数进行迭代奇异值分解,根据无偏风险估计准则自适应地确定正则化参数,实现了迭代过程的自动化。仿真和实验结果表明:在强剩磁条件下,该方法能够准确还原磁性异常体的轮廓形态,具有较好的模型分辨率。

关键词: 磁性目标, 强剩磁, 磁化方向, 磁总场模量, 奇异值分解, 无偏风险估计

Abstract: A three-dimensional focusing inversion approach of magnetic target is proposed for the effect of remanent magnetization on the true magnetization direction in the process of inversion. For an isolate magnetic source, its magnetization direction is estimated, and then the estimated magnetization direction is incorporated into an inversion algorithm. For multi-magnetic sources, the direct inversion of the total modulus data is made. Total modulus data depends weakly upon magnetization direction. In the iterative process, the depth weighting matrix and minimal support matrix are added in the inversion model of Tikhonov regularization theoretical framework to obtain the objective function and avoid the multiple solutions of inversion problem. Objective function is calculated by singular value decomposition, and the unbiased predictive risk estimator principle is used to adaptively determine the regularization parameter in the iterative process, thereby realizing the iterative process automation. Simulated and experimental results show that the proposed inversion method can reflect the outline form of magnetic anomaly and has good model resolution in the presence of significant remanence. Key

Key words: magnetictarget, significantremanence, magnetizationdirection, totalmodulusdata, singularvaluedecomposition, unbiasedpredictiveriskestimator

中图分类号: