Statistical stacking and adaptive notch filter to remove high-level electromagnetic noise from MRS measurements

摘要

Traditionally, one of the major limitations for magnetic resonance sounding (MRS) measurement is that the weak signal generated by subsurface water molecules is prone to be disturbed by highlevel electromagnetic noise. In China, the power grid coverage is 94.6% and spiky noise and power-line harmonic noise are always present when utilizing MRS measurement in suburban areas or towns. In order to improve the performance of the MRS method, two new techniques, statistical stacking and adaptive notch filter, are introduced to remove spiky noise and power-line harmonic noise. Firstly, four stacking procedures are analysed to suppress the natural noise and spiky noise. It could be found that statistical stacking can be utilized in the areas with serious spiky noise and can improve the signal-to-noise ratio by a factor of 4 to 7. Moreover, the stacking number is less than other stacking procedures and the measurement time may decrease by nearly 50% in some suburban areas or towns. Secondly, there are a variety of filtering procedures available to suppress power-line harmonic noise, which are all based on analogue or digital notch filtering. But nearly all of them may cause distortion. An adaptive notch filter is applied here to remove power-line harmonic noise because harmonic frequencies are away from and (or) close to the Larmor frequency, even when the frequency offset between them is zero. From simulation results, it could be noted that the signal can be recovered after adaptive notch filtering because it is not irretrievably distorted but proportionally attenuated. Thus, the amplitude attenuation can accurately be compensated. The effectiveness of the two techniques applied to MRS measurements is demonstrated by field testing with the prototype of the MRS system developed by Jilin University, China. The results show that the statistical stacking and adaptive notch filter are effective methods to remove high-level electro-magnetic noise from MRS measurements.