Fitting-correlation Analysis of Pulsed Thermographic Sequence Data

To detect the subsurface corroded defect of the aeroplane aluminum alloy with pulsed thermography, an improved approach was proposed to process the thermal image sequence. The acquired raw temperature data were preprocessed by applying polynomial fitting in logarithmic scale. Correlation coefficients relative to a specified pixel were calculated based on the reconstructed data. The mapping image of the computed values was used to increase the signal-to-noise ratio (SNR) and enhance the defect contrast. Besides angle cosine algorithm, the reference point location and optimizing truncation time window issues were also discussed. The experimental results show this technique is effective to emphasize the defect visibility compare to several conventional methods.