Comparison of optimisation strategies for the improvement of depth detection capability of Pulse-Compression Thermography

In Pulse-Compression Thermography, the impulse response of the Sample Under Test is retrieved pixelwise by applying a proper matched filter on the set of acquired thermal images obtained by stimulating the system with a heat source amplitude–modulated by a proper coded signal. Linear frequency-modulated chirp signals and binary codes are the most employed coded excitations, and to improve the detection capability of the technique, a non-linear frequency-modulated chirp signal can be used to deliver more energy to the sample in a frequency range of interest. In this work, we report the application of an exponential chirp to modulate the heating source and we compare it with a standard linear chirp excitation. To do a fair comparison, various windowing functions have been applied on the matched filters to reduce range side lobes, thus enhancing the retrieved impulse response quality. It is shown that the combined use of an exponential chirp and an appropriate matched filter obtained by exploiting the Reactance Transform window provides a faithful reconstruction of the sample impulse response and an enhanced signal-to-noise ratio with respect to the use of linear chirp. This has been demonstrated on a 3D-printed polymethyl methacrylate (PMMA) sample containing 16 flat-bottom holes of different depths.