Compared with the other geophysical approaches,magnetic resonance sounding(MRS)technique is direct and nondestructive in subsurface water exploration and it has been developed quickly in recent years.The magnetic resonance sounding(MRS)oscillating signals are recorded with high sampling rate,so the comprehensive information in the signals provides more potential in signal processing and inversion interpretation.However,MRS measurement always suffers bad signal-to-noise ratio(SNR).For the random noise interference,due to its stochastic properties,it always distributes arbitrarily.Normally,random noise appears randomly in signal recordings and difficult to remove.As long as the average magnitude of random noise is larger than that of signal,the accuracy of parameter estimation will be reduced.The general objective of this study is to solve the problem of the oscillating MRS signals contaminated with random noise.We propose to use segment time-frequency peak filtering method,the MRS oscillating signal is divided into several segments,and each segment of the signal is transformed into the instantaneous frequency of analytic signal.Then a significant energy concentration is produced around the IF on the time-frequency plane of analytic signal,and the unbiased estimation of underlying MRS signal is achieved by taking the peak of Wigner-Ville distribution of analytic signal.The validity of theoretical considerations is demonstrated by numerical simulations,it shows that the desired MRS oscillating signal can be recovered in noise level down to a SNR of-5 dB.
展开▼
