低纬度磁异常化极及分量换算的正则化方法

Regularization Method for Reduction to the Pole and Components Transformation of Magnetic Anomaly at Low Latitudes

本文在分析低纬度磁异常化极和分量换算算子放大高频噪声的频率特性基础上,采用Tikhonov正则化方法来稳定所讨论的转换算子。同时,提出了一种基于径向平均功率谱的磁异常噪声水平估计方法,进而利用后验策略的偏差准则求取正则化方法的正则参数。基于理论模型的试验结果表明,磁异常噪声水平估计方法精度较高,且低纬度磁异常正则化化极和分量换算的稳定性和精度相较常规算子有大幅提高。

Based on the analysis of frequency characteristics for the reduction to the pole operator and three components transformation operators for magnetic anomaly at low latitudes, we used Tikhonov regularization method to stabilize the operators discussed in this paper. Meanwhile, a noise level esti- mation method is proposed based on the radial average power spectrum of magnetic anomalies. Then, we used the discrepancy principle; a posterior strategy, to choose the regularization parameter. Test results based on theoretical model show that the proposed noise level estimation method has high pre- cision. Furthermore, compared it with the conventional operators, the regularization operators for the reduction at the poles and component transformation of magnetic anomaly at low magnetic latitudes exhibited greatly improved stability and accuracy.