Study On Signal Denoising Method In Spontaneous Potential Curve Based On Independent Component Analysis

Logging in Hydrogeology, Spontaneous Potential (SP) logging is one of the most common and the most important method that has been proposed and put into use as early as in the 1930s.This method is simple and high practical value. It can not only for determine lithology, divide interface of aquifer, but also to determine the salinity of formation water, separate layer of salt-water and fresh-water. However, the curve of the SP logging is influenced by many factors, such as stratigraphic configuration, potential electrode, Counter weight erosion,and external electromagnetism environment etc. so the results of the SP curve cannot truly reflect the real status of strata.If the interference for SP curve cannot been eliminated correctly and effectively, enhance the SNR (Signal to Noise Ratio) will reduce the effect of its, in extreme cases it curve cannot been accurately interpreted. So how to enhance the SNR that become a struggling target more and more.At present, the traditional signal de-noising method has the Fourier transform, adaptive wavelet device, wavelet de-noising, etc. But the frequency band of SP signal is broader, showing the non-stationary random distribution, spectrum overlap, and traditional de-noising method is difficult to overall balance. Therefore, using a more efficient way to eliminate the noise of SP signal and enhance the SNR that become very important.Independent component analysis is a new technology for signal processing, which applies to handle signals whose prior information is not clear. According to characteristics of the obtained source signal, the source signal is restored or separated from the observed signal in the light of principle of statistical independence, thus helping to achieve signal enhancement and analysis. According to SP logging signal of noise, this paper realizes the ICA theory, and apply it to de-noising, and detailedly analyse the characteristics of SP logging signal noise and explain the ICA theory. Using the algorithm based on negative entropy de-noising that compared with traditional method. This method has a wide engineering application in the future.