Damage detection using the discrete wavelet transformation (DWT) of temperature field recorded on the surface of a plate structure is discussed. Defects in the form of voids and inclusions of various magnitude are considered. Computer-simulated experiments including the ones accounting for white noise and real thermal experiments are applied. Transient heat transfer problems are analysed. The efficiency of DWT in damage detection is studied by way of a number of examples. Haar, Daubechies 4, 6 and 8 wavelets in 1D and 2D DWT are used. The usefulness of the Lipschitz exponent evaluated with the use of DWT is discussed. The issue of denoising the thermal response signals using the soft threshold shrinkage procedure is also taken up.