Premium
Surface nuclear magnetic resonance signals recovery by integration of a non‐linear decomposition method with statistical analysis
Author(s) -
Ghanati Reza,
Hafizi Mohammad Kazem,
Fallahsafari Mahdi
Publication year - 2016
Publication title -
geophysical prospecting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.735
H-Index - 79
eISSN - 1365-2478
pISSN - 0016-8025
DOI - 10.1111/1365-2478.12296
Subject(s) - hilbert–huang transform , noise (video) , stochastic resonance , signal (programming language) , signal to noise ratio (imaging) , gaussian noise , physics , nuclear magnetic resonance , computational physics , computer science , algorithm , acoustics , mathematics , white noise , statistics , optics , artificial intelligence , image (mathematics) , programming language
ABSTRACT Presence of noise in the acquisition of surface nuclear magnetic resonance data is inevitable. There are various types of noise, including Gaussian noise, spiky events, and harmonic noise that affect the signal quality of surface nuclear magnetic resonance measurements. In this paper, we describe an application of a two‐step noise suppression approach based on a non‐linear adaptive decomposition technique called complete ensemble empirical mode decomposition in conjunction with a statistical optimization process for enhancing the signal‐to‐noise ratio of the surface nuclear magnetic resonance signal. The filtering procedure starts with applying the complete ensemble empirical mode decomposition method to decompose the noisy surface nuclear magnetic resonance signal into a finite number of intrinsic mode functions. Afterwards, a threshold region based on de‐trended fluctuation analysis is defined to identify the noisy intrinsic mode functions, and then the no‐noise intrinsic mode functions are used to recover the partially de‐noised signal. In the second stage, we applied a statistical method based on the variance criterion to the signal obtained from the initial phase to mitigate the remaining noise. To demonstrate the functionality of the proposed strategy, the method was evaluated on an added‐noise synthetic surface nuclear magnetic resonance signal and on field data. The results show that the proposed procedure allows us to improve the signal‐to‐noise ratio significantly and, consequently, extract the signal parameters (i.e., T 2 * and V 0 ) from noisy surface nuclear magnetic resonance data efficiently.