The Study On Noise Elimination Technology Of Magnetotelluric Sounding In Xiangshan Uranium Field

Xiangshan Uranium Field is the largest volcanic uranium field in China.After decades of exploration and exploitation, shallow geological characteristics and metallogenic regularity have been identified by people;But as the deep lead-zinc uranium deposits are exposed in recent years and there are many disputes to the deep geological structure and mineralization characteristics in this area. So China geological survey launched 3D geological mapping project in this area,intending to depict the 3D geological structure and find out its metallogenic regularities.This paper which is based on” 3D geological mapping project of Xiangshan Uranium lead-zinc multi-metal prospect area in jiangxi province ”, how to eliminate noises effectively and improve the MT data quality issues, provides technical reference for the smooth implementation of the project.This paper analyzes largely the features of MT data time series signal and the noise can be divided into impulse noise,harmonic noise,step noise,charging and discharging noise,square-wave noise,triangle noise etc according to the waveform of the time sequence of signal. Combined with the actual situation in the study area,we summarize and sum up the characteristic of interference and influence rule such as the daily work of mine mining equipment, the area quarry operations, wireless communications,high tension line,industrial stray current and humane activities. Comparing with the mutual reference technology application in multi-metal data processing at the same time,this study shows that,whether it is electric track or magnetic track,the data quality has been improved significantly when the direct correlation of mutual reference technology is better. For time domain noise using morphological filtering method which indicates that morphological filtering can effectively filter large scale disturbance and baseline drift of the suppress curve via the management of simulation data and real data,even the application of the combination of generalized morphological filter can preferably keep the structure characteristics of the target signal that highlights the local effective signal,therefore,morphological filtering is an effective and reliable way to the suppression of noise in measurement area.The paper finally conducted the de-noissing of observed magnetotelluric data of two lines through morphological filtering and completed the 2D NLCG inversion,and moreover, studied the effect of regularization factor(τ) to inversion results,obtaining the best fit for τ,then delineated the orientation and position of the main geological structure in line with the inversion results.