Comparative study of thermal contrast and contrast in thermal signal derivatives in pulse thermography

For detection of subsurface defects, flash thermography results are often evaluated by selecting the thermal image with the largest contrast between the sound and defective area. Another possibility is to calculate the logarithmic derivatives of the thermal signal, and evaluate the 1st or 2nd derivative images. In this paper, it is investigated how the contrast in these images depend on the defect size and defect depth. Analytical calculations and FEM simulations are presented, and based on these results equations for the contrast maxima are derived. Several advantages of the derivative images are shown; for example a defect with an aspect ratio around unity can be reliably detected in the derivative image, but not in the thermal image. To demonstrate these results experimental data are also presented.