Integration of Time-frequency L4 matrix factorization of defect detection with Hilbert transform thermographic inspection

Abstract

The carbon fiber reinforced polymer (CFRP) has been widely used in aerospace, automobile and renewable engine. Since CFRPs are prone to cause impact-induced damage such as inner delamination, nondestructive testing methods have been applied to guarantee the safety. Optical pulsed thermography (OPT) is an effective nondestructive testing method. However, in the thermography processing, time and frequency information is not completely linked. In this paper, Hilbert transform is selected as well as integrating with L4 matrix factorization to mine both time–frequency information from thermal sequences. Compared with the pulsed phase thermography (PPT) which is mainly analyzed in frequency domain, the contrast is enhanced through the proposed method. In addition, time–frequency linked matrix factorization characteristics of defects at different depths can be interpreted as a basis for analysis of defects quality. The validity of the proposed method is verified by simulation and experimental results.