Dynamic compression reconstruction of pulsed thermography sequence for defect characteristic enhancement in composites

Abstract

In this study, a dynamic compression reconstruction (DCR) method for pulsed thermography sequence was proposed, which could effectively compress the original data length of pulsed thermography and significantly enhance the defect signal characteristics. The DCR method modulated the length of the pulsed thermography sequence to dynamically compress the image sequence with a newly defined F-function, thus reconstructing the DCR thermal image sequence. The length of the DCR thermal image sequence was compressed to 1/4 the length of the raw thermal image sequence. We assessed the DCR method using the pulsed simulation data of a homogeneous flat plate with circular defects in the subsurface, and found that the proposed DCR method could suppress temporal noise, reduce the edge blur of defects, and enhance defect recognition. To further verify the validity and practicability of this method, the pulsed thermography experimental data of the glass fiber reinforced plastic (GFRP) specimen was processed using the DCR method, and the experimental results were consistent with the numerical analysis results of the simulated data. The DCR thermal image of the GFRP specimen had lower spatial noise and higher signal-to-noise ratio (SNR) of the defects than the raw thermal image. Finally, the DCR method was compared with other pulse thermal signal processing methods, including TSR, PPT, and PCT, indicating that the pulsed thermal images processed by the DCR method could achieve the highest SNR of the defects.