Abstract: It is difficult to accurately identify corrosion fringe in nondestructive testing (NDT), which will directly in fluence the effect of corrosion imaging. Array pulsed eddy current (APEC) is used to detect corrosion. With the purpose of overcoming the problem of low identification accuracy in conventional NDT which only extracts features in the time domain, on the basis that PEC has a broadband frequency, features are extracted in both time and frequency domains. A method based on principal component analysis (PCA) is proposed to enhance the accuracy of identification of corrosion fringe, the agreement between theory and experiment shows that the present method is correct.

Key words: Array pulsed eddy current, Corrosion, Feature extraction, Fringe identification, Principal component analysis