Developing a multi-scale adaptive generalized morphological filtering algorithm for denoising ocean wave-induced interference in marine electromagnetic data

Abstract

As one of the marine electromagnetic detection method, marine magnetotellurics(MT) is an important geophysical exploration method for detecting underground electrical structures. However, since the nature magnetotelluric field is used as the field source, MT data is easily contaminated by various noise, especially the magnetic field generated by the motion of seawater waves, which leads to alterations in both apparent resistivity and phase characteristics within the frequency range impacted by the interference. Building on the principle of mathematical morphology, we propose an innovative approach that integrates multi-scale morphological filtering with adaptive generalized morphological filtering, and applied it to the marine electromagnetic field for the first time. Multi-scale morphological filtering considers the variation characteristics of noise at different time scales, and adaptive morphological filtering considers the non-stationary characteristics of wave-induced interference and ground electromagnetic signals. In addition, the size of structural elements is selected by introducing the evaluation index of mutual relation number, which realize the adaptive selection of structural elements. This approach can effectively suppress the wave-induced magnetic noise in marine MT signals. The results derived from applications of the method to both simulated datasets with varying levels of ocean wave-induced disturbances and field datasets demonstrate its ability to effectively mitigate magnetic noise stemming from wave-induced effects. As a result, the observed apparent resistivity and phase trends display a high degree of coherence and continuity. This provides new insights into noise suppression in the marine electromagnetic field.