基于独立成分分析的超低频电磁探测信号滤波

Filtering of Ultra-Low Electromagnetic Detection Signal Using Independent Component Analysis

将独立成分分析方法引入到超低频电磁探测曲线的滤波中。首先运用模拟数据来探讨运用独立成分分析方法进行滤波的可行性,然后运用实际的观测数据对基于独立成分分析的滤波方法进行了评价和比较。结果表明,运用独立成分分析方法进行探测曲线的噪声滤除是可行的,而且该方法可有效地滤除噪声、突出曲线的特征,有利于曲线的地质解译。同时,对相关问题进行了讨论。

The independent component analysis （ICA） was introduced for the filtering of the ultra-low electromagnetic （ULEM） detection curves. Simulated datasets were first used to validate the effectiveness of ICA in the filtering. The real datasets were then used to evaluate and compare the performance of the proposed filtering method. The results show that filtering of the ULEM signals based on ICA is feasible. Moreover, the proposed filtering method could effectively suppress the noise and enhance the valuable features of the curves, which facilitates geological interpretation of the resulting curves. Some issues are also discussed.