Entropy-Based Technique for Denoising of Acoustic Emission Signals

The acoustic emission (AE) method has been successfully used in recent years to monitor the condition of industrial and civil infrastructures. In AE, time-of-arrival (ToA) estimation is considered a key parameter for the accurate localization of a growing defect. This paper describes an entropy-based filtering approach for the ToA estimation of noisy signals and compares its performance to that of the commonly adopted Akaike Information Criterion (AIC). The proposed method consists in coarsening the input data using the Crutchfield-Packard algorithm and calculating the local (instantaneous) entropy. In the present study, we demonstrate that the local entropy of the background noise component differs from the useful (informative) signal. As a result, the approach permits filtering the noise component by selecting a proper threshold value. The proposed method has been tested on experimental data aimed at localizing a source of AE in a square 1 x 1m aluminum plate. The entropy approach allows an overreaching precision in the final localization of the targets compared to the classical AIC.