TY  - RPRT
N2  - This document presents a new learning-by-example (LBE) technique for the computationally-efficient inversion of eddy current testing (ECT) data in non-destructive testing and evaluation (NDT-NDE) scenarios. More precisely, the developed approach exploits a uniform sampling strategy to build a training set of input/output (I/O) pairs and exploits such information to train a Support Vector Regressor (SVR). During the on-line testing phase, previously-unseen ECT data are given as input to the trained model in order to predict the position of a single narrow crack within a planar conductive structure. Some representative numerical results are shown, in order to preliminarily assess the capabilities of the developed approach when dealing with the presence of a non-negligible amount of noise on test data.
M1  - technical_report
Y1  - 2016///
UR  - http://www.eledia.org/students-reports/712/
TI  - Fast Inversion of Eddy Current Testing Data Through a Learning-by-Examples Approach for Robust Crack Localization
PB  - University of Trento
KW  - Eddy current testing
KW  -  inverse scattering
KW  -  nondestructive testing and evaluation
KW  -  statistical learning
KW  -  learning-by-examples
KW  -  support vector regression
AV  - public
A1  - Salucci, M.
A1  - Anselmi, N.
A1  - Oliveri, G.
A1  - Massa, A.
ID  - elediasc12712
ER  -