瞬变电磁信号降噪算法

Improved ICA Solutions for TEM Signal Noise Suppression

研究地下异常体的有用信息,由于需要野外采集,所接收的瞬变电磁信号会叠加各种电磁干扰和噪声,而传统的降噪方法不能较好地对瞬变电磁接收的二次场衰减信号降噪,严重影响了利用信号对地下异常体特征的数据成图和地质状况解释。针对上述问题,提出了改进的基于独立成分分析的瞬变电磁接收信号降噪处理算法。由接收的瞬变电磁信号构成多维观测向量,利用独立成分分析方法从观测向量中分离出信号空间基向量,然后采用峰度判别准则对信号空间和噪声空间进行分离,保留二次场信号基向量,用二次场信号基向量张成信号子空间,实现降低噪声的目的。降噪后的瞬变电磁信号可以更好地反演出地下相关信息。通过对实测数据的验证,结果表明提出的算法应用于瞬变电磁法信号降噪是可行的,能有效地降低噪声并分离出二次场信号的有用信息,对提高瞬变电磁数据成图和地质状况解释质量具有较好的作用。

Useful information of underground anomalies bodies can be deduced using transient electromagnetic method. However, due to the need of outside field data acquisition, the received transient electromagnetic signals are usually superimposed electromagnetic interference and noise which seriously affect underground anomaly data mapping and geological conditions explanation using the transient electromagnetic signals. The traditional noise reduction methods for the received transient electromagnetic secondary field attenuation signal can not get good results. To this issue, a new algorithm was proposed based on improved ICA of noise reduction for the received transient electromag- netic signals. This method builds multi -dimensional observation vectors from the transient electromagnetic signals received, researches on basis vectors of signal spaces using independent component analysis, and then uses kurtosis criterion to separate the signal space and noise space in order to achieve the purpose of noise suppression. The noise processed transient electromagnetic signal can be better applied in the inversion calculation of underground informa- tion. By measured data verification, it was proved that the proposed algorithm has better effects being applied to the signal noise suppression of transient electromagnetic method. It can effectively reduce the noise as well as separate the useful information from the secondary field signal, and has better effects in improving the quality of transient electromagnetic data mapping and geological condition explanation.