In order to identify surface or subsurface features of an object, active thermal non-destructive testing (TNDT) techniques use the temperature surface profile that has been measured over the object. To improve the flaw detectability using depth analysis, this technique, however, calls for fresh processing and excitation approaches. In order to characterize carbon reinforced plastic material, this contribution aims to demonstrate the detection capabilities of the recently introduced orthonormal projection approach to pulse thermography for infrared imaging [1].