Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
The main objective of this paper is to precisely classify surface and subsurface cracks and cavities using a feature-based giant magnetoresistive pulsed eddy-current (PEC) sensor. Based on the author's previous work involving amplitude spectral analysis combined with wavelet decomposition, the power spectral density analysis of the direct differential PEC response, and the reconstructed approximate and detail components of wavelet transform are performed to extract the defect feature. Principal component analysis (PCA) is designed to eliminate the dimensional method, which is identified with the ability to supply the low-dimensional feature. PCA confused linear discriminant analysis and the Bayesian classifier are both applied for defect classification. The experimental results reveal that cracks and cavities on the surface and subsurface can be classified satisfactorily by the proposed methods that have the potential for gauging automatic in situ inspection for PEC.