Automatic detecting and classifying defects during eddy current inspectionof riveted lap-joints

This article presents a novel method for automatic detection and classifica tion of cracks located in the second lower layer of the aircraft lap-joints during Eddy Current (EC) inspection. The cracks originating from the rivet holes were detected using a tailor-made deep penetrating EC probe. The pro posed method consists of three steps: pre-processing, feature extraction an d classification. The pre-processing, performed before the feature extracti on included median filtering, rotation and de-biasing of the EC patterns. T he rotation of the patterns was performed so that energy of the responses t o the rivets was maximized along the quadrature direction, while the defect responses were maximized in the in-phase direction in the impedance plane. Feature extraction was then performed using four different methods: discre te wavelet transform, Fourier descriptors, principal component analysis (PC A) and block mean values. The classification was performed using a standard multi-layer perceptron (MLP) neural network. AU the pre-processing methods showed similar classification performance on the used data set, but the PC A method compressed the data best. (C) 1999 Elsevier Science Ltd. All right s reserved.