Hilbert transform pitfalls and solutions for ultrasonic NDE applications

Hilbert transform (HT) is a classical tool used to obtain complex analytical signal representation, which is useful for instantaneous frequency and envelop estimation of bandpass signals. However, noise has a significant adverse impact on the performance of HT. Furthermore, the narrowband signal condition in Bedrosian identity makes it problematic to analyze ultrasonic scattering signal using HT. In this investigation, two key issues related to Hilbert transform are addressed for enhanced instantaneous frequency (IF) estimation. First, in order to minimize the effect of the noise, ultrasonic signals are decomposed to multiple narrowbands and instantaneous frequencies within these bands are estimated. Second, a weighted estimated of IF based on envelop estimate of each narrowband is introduced. These methods are applied to various experimental ultrasonic data sets and utilized to examine microstructure scattering, effects of attenuation in large grained materials, and flaw detection in presence of high scattering noise. Simulation studies and experimental results support accuracy of the IF estimation. Enhanced IF estimation techniques provide tractable frequency estimation and makes it possible to quantify spectral shifts due to attenuation, scattering and dispersion effects.

[1]  J. Saniie,et al.  Analysis of order-statistic CFAR threshold estimators for improved ultrasonic flaw detection , 1992, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[2]  Alan S. Morris,et al.  Using short-time Fourier transform and wavelet packet filter banks for improved frequency measurement in a Doppler robot tracking system , 2002, IEEE Trans. Instrum. Meas..

[3]  Erdal Oruklu,et al.  Fast Chirplet Transform With FPGA-Based Implementation , 2008, IEEE Signal Processing Letters.

[4]  Leon Cohen,et al.  Time Frequency Analysis: Theory and Applications , 1994 .

[5]  Yu Zhang,et al.  Microembolic signal characterization using adaptive chirplet expansion. , 2005, IEEE transactions on ultrasonics, ferroelectrics, and frequency control.

[6]  E. Bedrosian A Product Theorem for Hilbert Transforms , 1963 .

[7]  Boualem Boashash,et al.  Estimating and interpreting the instantaneous frequency of a signal. I. Fundamentals , 1992, Proc. IEEE.

[8]  Nii O. Attoh-Okine,et al.  The Hilbert-Huang Transform in Engineering , 2005 .

[9]  M. Shen,et al.  Instantaneous Frequency Estimate of Nonstationary Phonocardiograph Signals Using Hilbert Spectrum , 2005, 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.

[10]  J. Saniie,et al.  Chirplet transform for ultrasonic signal analysis and nde applications , 2005, IEEE Ultrasonics Symposium, 2005..

[11]  D.A. Hutchins,et al.  The application of time-frequency analysis to the air-coupled ultrasonic testing of concrete , 2006, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[12]  Ramdas Kumaresan,et al.  Algorithm for decomposing an analytic signal into AM and positive FM components , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[13]  J. Saniie,et al.  A successive parameter estimation algorithm for chirplet signal decomposition , 2006, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[14]  J. Saniie,et al.  Application of Hilbert-Huang transform for ultrasonic nondestructive evaluation , 2008, 2008 IEEE Ultrasonics Symposium.

[15]  Ewa Hermanowicz,et al.  On bedrosian condition in application to chirp sounds , 2007, 2007 15th European Signal Processing Conference.

[16]  A.R. Messina,et al.  Identification of instantaneous attributes of torsional shaft signals using the Hilbert transform , 2004, IEEE Transactions on Power Systems.

[17]  Norden E. Huang,et al.  INTRODUCTION TO THE HILBERT–HUANG TRANSFORM AND ITS RELATED MATHEMATICAL PROBLEMS , 2005 .