The Research Of Morphological Filtering Denoising And Improved Data Processing Method For EH4Electromagnetic Sounding System

Audio-frequency Magnetotellurics(AMT) sounding is a kind of electromagneticexploration method which uses natural alternating electromagnetic field as a fieldsource. Cagnird resistivity sounding which base on plane electromagnetic wave theoryanalysis the earth structure and electric property distribution by the characteristic ofapparent resistivity and phase curve. So one of the most basic and important contentsis the analysis of earth impedance. In1996, EH4electromagnetic sounding systemwhich is jointly developed by the EMI company and Geometrics company ofAmerican is an important audio-frequency magnetotelluric instrument. In recentlyyears EH4electromagnetic sounding system is used extensive in deep solid fuels andmineral resources prospecting, geologic hazard investigation, environmentalmonitoring for the characteristic of high work efficiency, abundance data collectionand portability equipment. This paper starts from the data processing of EH4electromagnetic sounding system,it researchs the basic method of data processing andde-noise method for the nesting IMAGEM software of EH4electromagnetic soundingsystem. And it researchs the improved method for the software.This article researchs the data processing of reading time series,FastFourier-Transformation into Frequency domain, computation of the Cross and Autopower-spectral, computation of the tensor impedance and apparent resistivity.Itanalysis the inadequacies of power spectral estimation method(Fourier spectralestimation method) and impedance tensor estimates(LS) by IMAGEM software, andput forward some improvements for them.The paper use the method of morphologicalfiltering to de-noise the time series.And it studys relatively the filtering effect ofdifferent kinds of strucrting elements in order to choose the best strucrtingelements.One of them is a filter with different size of same type strucrtingelements.The other is a filter with same size of different type strucrtingelements.Then the paper used the morphological filter to processing the measureddata and have achieved good results. It used the Hilbert marginal spectrum to estimatethe power spectrum and study relatively on the Hilbert marginal spectrum and Fourierspectrum of the test signal and measurement signal,and the superiority of the Hilbertmarginal spectrum is analyzed.This article discusses the principle and processing ofleast squares estimation method and Robust estimation, and it studys relatively on theprocessing results of improved method and IMAGEM processing on the measureddata.And the improved method is proved to be effective through the result.