| ABSTRACT: | IMPROVING SIGNAL PROCESSING OF THE IMPACT-ECHO METHOD USING
CONTINUOUS WAVELET TRANSFORM
C. Chiang, C. Cheng
Chaoyang University Of Technology, Wufeng, Taichung, Taiwan
Fourier transform is frequently used in analyzing the frequency content of a transient signal. The drawback
of Fourier transform is that the time-domain characteristic cannot be shown simultaneously. Wavelet
transform is one of the time-frequency dual transforms that preserves both frequency content and time
information of a transient signal. The length of both time window and frequency window can be varied in
wavelet transform. As a result, the time resolution is better in high frequency band while the frequency
resolution is better in low frequency band. A typical problem in the processing o fimpact-echo signals arises
from peak detection based solely on frequency spectrum. Current study aims to improve the feature
identification of impact-echo signals obtained from buried objects in concrete. Rebars, steel tubes, and PVC
tubes of different diameters at various depths from the concrete surface are tested. Numerical simulations are
carried out based on models constructed using the finite element method. The received signals, both
experimental and simulated, are analyzed using both fast Fourier transform and continuous wavelet
transform. Multiple back-wall reflections can be easily detected from the wavelet coefficient plots (WCPs).
Low-frequency signals due to rebars at a greater distance is clearer shown in the WCP than those due to
rebars at a smaller distance from the side wall of concrete specimens. Such findings provide a more effective
way to pick up rebar signals using the impact-echo method.
|