基于自适应滤波的磁梯度数据三维相关成像

3D Correlation Imaging of Magnetic Gradient Data Based on Adaptive Filtering

摘要:

为了提升磁梯度数据相关成像法对不同深度的多磁性目标的成像效果，提出了一种自适应滤波方法。首先将待成像空间划分为三维规则网格，通过解析法计算位于不同网格的磁性体在观测平面的理论磁梯度，然后依据各磁性体的磁梯度数据频谱特征来决定观测数据的滤波阈值，最后将滤波后的观测数据用于相关成像。仿真模型实验表明，对于位置较浅磁性体，根据成像结果估计的磁性体中心位置与实际基本相符，滤波处理最多减小了约76%的误差，对于平均1.7m深的多个磁性体，估计的中心位置最大误差0.3m，相比于无滤波处理的结果误差减小了46%左右，信噪比不足时，对近似球体的模型仍然有较好的定位效果。说明自适应滤波在一定程度上能改善三维相关成像的分辨率，提升水下磁性目标定位的准确度。

Abstract:

In order to improve the imaging effect of magnetic gradient data correlation imaging method on multiple magnetic targets at different depths, an adaptive filtering method is proposed. Firstly, the space to be imaged is divided into three-dimensional regular grids, and the theoretical magnetic gradients of magnetic objects located on different grids in the observation plane are calculated using analytical methods. Then, the filtering threshold of the observation data is determined based on the spectral characteristics of the magnetic gradient data of each magnetic object. Finally, the filtered observation data is used for correlation imaging. The simulation model experiment shows that for shallow magnetic objects, the estimated center position of the magnetic objects based on the imaging results is basically consistent with the actual situation. The filtering processing reduces the error by up to 76%. For multiple magnetic objects with an average depth of 1.7 meters, the estimated center position has a maximum error of 0.3 meters, which is about 46% less than the error of the results without filtering processing. When the signal-to-noise ratio is insufficient, the model that approximates a sphere still has good localization performance. Adaptive filtering can to some extent improve the resolution of 3D correlation imaging and enhance the accuracy of underwater magnetic target localization.