Wavelet Transform-Based Signal Denoising in Low-Field NMR

doi:10.11938/cjmr20182615

Abstract

Abstract: Low-field NMR often uses permanent magnet. The signals obtained contain high level of white Gaussian noises, and have low signal-to-noise ratio (SNR). In recent years, many denoising methods have been proposed for low-field NMR measurements. Most of these methods can remove noises without losing useful information contained in the original signals. Wavelet transform is the most popular denoising method among them. In this paper, we first introduced the theory of wavelet transform analysis, followed by review of three wavelet transform denoising methods for low-field NMR, namely the wavelet threshold method, the wavelet transform modulus maximum method and the correlation of wavelet coefficient method. Finally, we showed that four parameters could be calculated to evaluate the denoising performance.

Key words: low-field NMR, signal denoising, wavelet transform, wavelet threshold method, wavelet transform modulus maximum method, correlation of wavelet coefficient method

CLC Number:

O482.53

CHANG Xiao, SU Guan-qun, NIE Sheng-dong. Wavelet Transform-Based Signal Denoising in Low-Field NMR[J]. Chinese Journal of Magnetic Resonance, 2018, 35(3): 393-406.

References

[1] CARO D D, LIGUORI C, PIETROSANTO A, et al. Hazelnut oil classification by NMR techniques[J]. IEEE T Instrum Meas, 2017, 66(5):928-934.
[2] CARO D D, LIGUORI C, PIETROSANTO A, et al. Using a SVD-based algorithm for T₂ spectrum calculation in TD-NMR application to detect hidden defects in hazelnuts[C]//Instrumentation and Measurement Technology Conference. IEEE, 2017:1-6.
[3] KIRTIL E, OZTOP M H. ¹H nuclear magnetic resonance relaxometry and magnetic resonance imaging and applications in food science and processing[J]. Food Eng Rev, 2016, 8(1):1-22.
[4] MEMOLI A, ALBANESE D, ESTI M, et al. Effect of bug damage and mold contamination on fatty acids and sterols of hazelnut oil[J]. Eur Food Res Tech, 2017, 243(4):651-658.
[5] WANG X H, SUN P, ZHANG X, et al. Application of magnetic resonance technique to quality and safety evaluation of food[J]. Chinese J Magn Reson, 2017, 34(2):245-256. 王小花, 孙鹏, 张许, 等. 磁共振技术在食品质量与安全研究中的应用[J]. 波谱学杂志, 2017, 34(2):245-256.
[6] ZHOU F D, GAO X, CAI J B, et al. Determination of treated and control wood on fiber saturation point by low-field NMR technology[J]. Chinese J Magn Reson, 2017, 34(1):108-114. 周凡丁, 高鑫, 蔡家斌, 等. 利用低温NMR技术测定木材及其热处理材纤维饱和点[J]. 波谱学杂志, 2017, 34(1):108-114.
[7] GAUNKAR N G P, NLEBEDIM I C, BULU I, et al. Broadband analysis of response from magnetic cores used in inductive sensors for pulsed nuclear magnetic resonance applications[J]. IEEE T Magn, 2016, 52(7):1-4.
[8] WANG H M, NIE S D, WANG Y J. The research progress of de-noising methods in low-field NMR signal[J]. Chinese Journal of Medical Physics, 2013, (4):47-51. 王红敏, 聂生东, 王远军. 低场核磁共振信号降噪方法研究进展[J]. 中国医学物理学杂志, 2013, (4):47-51.
[9] VEER K, AGARWAL R. Wavelet and short-time Fourier transform comparison-based analysis of myoelectric signals[J]. J Appl Stat, 2015, 42(7):1591-1601.
[10] MALLAT S G. A theory for multiresolution signal decomposition:the wavelet representation[J]. IEEE T Pattern Anal, 1989, 11(7):674-693.
[11] YAN T F, YAN T L. Radio spectrum multi-domains analysis based on STFT[J]. Information Technology, 2010, 11:163-165. 尹天峰, 颜挺利. 基于STFT的无线电频谱信号的多域分析[J]. 信息技术, 2010, 11:163-165.
[12] DU Q Y, CHANG Y, QIAN S S, et al. Comparison of time-frequency analysis methods for radio frequency pulses used in magnetic resonance[J]. Chinese J Magn Reson, 2016, 33(4):646-654. 杜庆阳, 常严, 钱嵩松, 等. 针对磁共振射频脉冲的时频域分析方法比较研究[J]. 波谱学杂志, 2016, 33(4):646-654.
[13] DONOHO D L, JOHNSTONE I M. Adapting to unknown smoothness via wavelet shrinkage[J]. J Am Stat Assoc, 1995, 90(432):1200-1224.
[14] WITKIN A P. Scale-space filtering[C]//International Joint Conference on Artificial Intelligence. DBLP, 1983:1019-1022.
[15] DONOHO D L. De-noising by soft-thresholding[J]. IEEE T Inform Theory, 1995, 41(3):613-627.
[16] XIE Q M, XIAO L Z, LIAO G Z. Application of SURE algorithm to echo train de-noising in low field NMR logging[J]. Chinese Journal of Geophysics, 2010, 53(11):2776-2783. 谢庆明, 肖立志, 廖广志. SURE算法在核磁共振信号去噪中的实现[J]. 地球物理学报, 2010, 53(11):2776-2783.
[17] MA S B, KONG L, CHEN J J. An improved NMR signal de-noising algorithm based on wavelet transform[J]. Journal of Computational Information Systems, 2011, 7(13):4651-4659.
[18] XIAO L Z, XIE Q M, XIE R H, et al. Noise reduction for NMR logging with regularization-heursure algorithm[J]. Journal of Computational Information Systems, 2013, 56(11):3943-3952.
[19] PAN Q, MENG J L, ZHANG L, et al. Wavelet filtering method and its application[J]. Journal of Electronics & Information Technology, 2007, 29(1):236-242. 潘泉, 孟晋丽, 张磊, 等. 小波滤波方法及应用[J]. 电子与信息学报, 2007, 29(1):236-242.
[20] XIE R H, WU Y B, LIU K, et al. Using wavelet domain adaptive filtering to improve signal to noise ratio of nuclear magnetic resonance log data from tight gas sands[J]. Geophys Prospect, 2016, 64(3):689-699.
[21] MALLAT S, ZHONG S F. Characterization of signals from multiscale edges[J]. IEEE T Pattern Anal, 1992, 14(7):710-732.
[22] MALLAT S, HWANG W L. Singularity detection and processing with wavelets[J]. IEEE T Inform Theory, 1992, 38(2):617-643.
[23] ZHANG X F, XU D Z, QI Z F. Denoising algorithm based on modulus maximum wavelet field[J]. Journal of Data Acquisition and Processing, 2003, 18(3):315-318. 张小飞, 徐大专, 齐泽锋. 基于模极大值小波域的去噪算法研究[J]. 数据采集与处理, 2003, 18(3):315-318.
[24] WANG X W, DONG G B, XIE G H. A new de-noising method of NMR FID signals based on wavelet transform[J]. Nuclear Electronics & Detection Technology, 2008, 28(2):365-370. 王希武, 董光波, 谢桂海. 基于小波变换的核磁共振FID信号的去噪方法研究[J]. 核电子学与探测技术, 2008, 28(2):365-370.
[25] GAO B K, ZHOU Y H, SUN L C, et al. Modulus maxima wavelet denoising and alternating projection reconstruction of logging signals[J]. Control and Instruments in Chemical Industry, 2015, 42(9):993-996.
[26] SUN S Q, MENG Q Y, FANG X C, et al. Noise cancellation algorithm for surface magnetic resonance signals based on self-adaption and reconstruction of wavelet modulus maximum value[J]. Journal of Jilin University Engineering and Technology Edition, 2015, 45(5):1642-1651. 孙淑琴, 孟庆云, 方秀成, 等. 基于自适应和小波模极大值重构的地面核磁共振信号噪声压制方法[J]. 吉林大学学报(工), 2015, 45(5):1642-1651.
[27] SENDUR L, SELESNICK I W. Bivariate shrinkage functions for wavelet-based denoising exploiting interscale dependency[J]. IEEE T Signal Process, 2002, 50(11):2744-2756.
[28] SAEEDI J, MORADI M H, ABEDI A. Image denoising based on fuzzy and intra-scale dependency in wavelet transform domain[C]//International Conference on Pattern Recognition. IEEE Computer Society, 2010:2672-2675.
[29] PLACIDI G, ALECCI M, SOTGIU A. Post-processing noise removal algorithm for magnetic resonance imaging based on edge detection and wavelet analysis[J]. Physics in Medicine & Biology, 2003, 48(13):1987-1995.
[30] XU Y S, WEAVER J B, HEALY D M J, et al. Wavelet transform domain filters:a spatially selective noise filtration technique[J]. IEEE T Image Process, 1994, 3(6):747-758.
[31] ZHANG L, BAO P. Denoising by spatial correlation thresholding[J]. IEEE Transactions on Circuits & Systems for Video Technology, 2003, 13(6):535-538.
[32] BAO P, ZHANG D. Noise reduction for magnetic resonance images via adaptive multiscale products thresholding[J]. IEEE T Medical Imaging, 2003, 22(9):1089-1099.
[33] CHANG S G, YU B, VETTERLI M. Adaptive wavelet thresholding for image denoising and compression[J]. IEEE T Image Process, 2000, 9(9):1532-1546.
[34] PAN Q, ZHANG L, DAI G Z, et al. Two denoising methods by wavelet transform[J]. IEEE T Signal Process, 1999, 47(12):3401-3406.
[35] ZHENG C X, ZHANG Y M. Method for eliminating noise in nuclear magnetic resonance signals based on wavelet transformation[J]. Analytical Instrumentation, 2007, 26(1):43-46. 郑传行, 张一鸣. 核磁共振信号的小波变换消噪方法[J]. 分析仪器, 2007, 26(1):43-46.
[36] ZHENG C, ZHANG Y M. Low-field pulsed NMR signal denoising based on wavelet transform[C]//Signal Processing and Communications Applications, 2007. Siu 2007. IEEE, 2007:1-4.
[37] GE X M, FAN Y R, LI J T, et al. Noise reduction of nuclear magnetic resonance (NMR) transversal data using improved wavelet transform and exponentially weighted moving average (EWMA)[J]. J Magn Reson, 2015, 251:71-83.
[38] WANG Z, BOVIK A C, SHEIKH H R, et al. Image quality assessment:From error visibility to structural similarity[J]. IEEE T Image Process, 2004, 13(4):600-612.