希尔伯特-黄变换在海底大地电磁测深数据处理中的应用  

Application of Hilbert-Huang transformation to marine magnetotelluric sounding data processing

作　　者：于彩霞[1,2,3] 魏文博[1,2] 景建恩[1,2] 叶高峰[1,2] 张帆[1,2] 赵文轲[1,2] 毛星[1,2] 

机构地区：[1]地下信息探测技术与仪器教育部重点实验室(中国地质大学,北京),北京100083 [2]中国地质大学(北京)地球物理与信息技术学院,北京100083 [3]中国科学院地质与地球物理研究所,北京100029

出　　处：《地球物理学进展》2010年第3期1046-1056,共11页Progress in Geophysics

基　　金：国家863课题(2006AA09A201);斯伦贝谢-中国地质大学(北京)教育基金联合资助

摘　　要：海水运动产生的电磁场,是海底大地电磁探测数据中的主要噪声之一,影响了数据处理与解释的精度.为提高海底MT探测的应用效果,本文在海底MT信号处理中引入HHT时频分析方法.结合海底MT噪声的特点与海底MT信号的特性,对南海实测数据进行分析,通过对比去噪前后信号的Hilbert时-频谱和边际谱,表明利用经验模态分解及其多尺度滤波特性,能够有效压制海水运动产生的电磁噪声.经HHT去噪后的测深曲线在一定程度上有所改善,仍存在"飞点"现象;再结合Robust估计,明显改善海底大地电磁探测数据质量,得到满意的观测结果.Motion of ocean waves can generate an electromagnetic field, which is one of the main noises in marine magnetotelluric （MT） signals. It will affect the accuracy of signal processing and interpretation. In order to improve the effect of the application of marine MT exploration, this paper introduces the Hilbert-Huang transformation timefrequency method to marine MT signal processing. Based on the characters of noises and signals of marine MT from the South China Sea, we compare the Hilbert time-frequency spectrum and Hilbert Marginal spectrum of original signals and signals after noise suppression. The result shows that using the empirical mode decomposition method and multi-scale filter characteristics can effectively suppress the noises generated by the motion of ocean waves. The sounding curves after denoise through HHT is better than the original curves to some degree. However, there still exist the so called ＂flying spots＂. In such case, we resort to the Robust estimation method, and the corresponding result is much satisfying.

关 键 词：HHT 经验模态分解 Hilbert边际谱 海底大地电磁 噪声压制 

分 类 号：P738[天文地球—海洋地质] P631[天文地球—海洋科学]