Controllable Source Electromagnetic Method Strong Interference Suppression Based On Machine Learning

Controlled-source electromagnetic method(CSEM)exploration is a very important method for people to obtain information in geology,metal and shale gas exploration.However,with the rapid development of cities,the interference of humanistic noise has become more and more serious.For example,high-voltage lines and some power stations have a huge impact on electromagnetic exploration,which has caused the difficulty of electromagnetic exploration to increase greatly,which greatly affects The resolution of controllable source electromagnetic prospecting makes it urgent to suppress noise.In order to improve the quality of CSEM data,this paper fully excavates the periodic characteristics of active periodic electromagnetic signals and proposes a new CSEM data processing method.(1)First,the signal to be measured is removed in the frequency domain by the fast Fourier transform(FFT)method of 50 Hz and its multiplier frequency.After removing the power frequency interference,the inverse Fourier transform restores the time domain data.(2)Immediately use the complementary set empirical mode decomposition(CEEMD)to remove the baseline wandering noise on the processed signal.(3)Finally,the Support Vector Machine(Support Vector Machine)in the machine learning algorithm is introduced to classify and filter the data,select high-quality signals,and complete signal-to-noise separation.In order to verify the effectiveness of the method,this article first carried out a simulation experiment of synthetic data.The processing results of synthetic data showed that the method can effectively and accurately separate the noise from the high-quality signal and greatly reducing the error rate.Later,a targeted field test was conducted in Huidong,Sichuan,and finally the method was applied to the processing of measured data in Huidong,Sichuan.From the results of the final apparent resistivity,it can be seen that the method described in this article can quickly and accurately isolate the human interference in the CSEM signal,screen out high-quality signals,and significantly improve the quality of CSEM data.