A combined emd and tfpf algorithm for random noise reduction of full-wave magnetic resonance signals

Abstract

The invention relates to a method for reducing random noise of a full-wave magnetic resonance signal combined with an EMD and TFPF algorithm, and is a "blind" filtering method that does not need to design a filtering interval. First, using the decomposition characteristics of the EMD algorithm, the full-wave magnetic resonance signal is decomposed into different eigenmode components, and then the TFPF algorithm is used to encode the dominant mode component of the signal into the instantaneous frequency of the unit amplitude analysis signal, and the time-frequency distribution of the analysis signal is used. Random noise is suppressed along the characteristic of instantaneous frequency concentration. This method requires fewer filtering constraints, is simple to operate, does not need to design filtering intervals in the time-frequency domain, and has strong adaptability to full-wave magnetic resonance signals with low signal-to-noise ratios. Significantly improve the detection efficiency, a better noise reduction effect can be obtained in one measurement, the random noise can be effectively reduced without losing the signal component, the signal-to-noise ratio can be significantly enhanced, and the accuracy of the later inversion interpretation can be improved.

Description

Technical field: [0001] The invention relates to a method for reducing noise of a magnetic resonance sounding (Magnetic Resonance Sounding, MRS) signal, in particular to a data processing method for reducing random noise contained in a full-wave magnetic resonance signal based on the principle of a combined algorithm of EMD and TFPF. Background technique: [0002] Magnetic resonance groundwater detection technology is an emerging geophysical detection technology. Due to its advantages of direct detection and quantitative evaluation, it has received extensive attention in recent years and has achieved rapid development. At present, it has gradually developed from theoretical research to the development of a relatively mature instrument detection system, and has carried out field measurements in some areas, and successfully detected shallow groundwater resources. Significance. However, due to the relaxation of excited hydrogen protons in groundwater, the magnetic resonance si...

AI Technical Summary

Problems solved by technology

[0004] The noise filtering method for full-wave nuclear magnetic resonance signals based on independent component analysis of the above invention can solve the traditional underdetermined blind source problem and amplitude uncertainty problem in independent component analysis. Compared with traditional MRS signal denoising methods, it has the advantages of calculation It has the advantages of fast speed, high signal-to-noise ratio, and strong practicability, but this method can only filter out power frequency interference and a certain frequency interference, and cannot suppress noise interference with a large frequency band distribution such as random noise; and based on cross-correlation The NMR full-wave signal noise filtering method can suppress power frequency and its harmonic noise, random noise and spike noise at the same time, but it mainly uses the uncorrelated characteristics of noise frequency and signal frequency. The part cannot be removed; the ICA-based NMR groundwater detection signal noise elimination method does not require prior knowledge of the source signal and transmission channel, and does not need to lay a reference coil during the test, which is simple, convenient and efficient, but the method It requires at least three sets of data, and the calculation process is complicated, the amount of calculation is large, and it is difficult for non-technical personnel to control; the superposition method is currently the simplest and most commonly used random noise reduction method for magnetic resonance signals. In the environment, the groundwater magnetic resonance signal with a higher signal-to-noise ratio can be obtained by using fewer stacking times, but when the noise intensity is large, multiple stacking is required, which will increase the measurement time and reduce the working efficiency of the instrument. limited effect

Images