Progress in materials science has led to novel high-performant materials, like fiber reinforced plastics (or composites), which are nowadays more and more used in aerospace and automotive industry. However, the composed nature of these materials makes them also sensitive to certain damage features. To assess the structural health, engineers have devised various non-destructive test approaches. One emerging technique concerns optical infrared thermography, in which the material is optically stimulated, and the thermal response is captured with an infrared camera. There are several thermographic implementations described in literature, but unfortunately there is no consensus on the optimal approach. In this project, we will investigate optical infrared thermography and identify both its opportunities and limitations for NDT of composite materials. This involves a critical evaluation of different excitation approaches as well as various post-processing techniques to enhance the detectability of (deep) defects. The experimental study is supported by numerical modelling. Further, human interpretation of the thermographic data may yield inaccurate (or even wrong) results. Therefore, an algorithm will be implemented to efficiently extract and size defects in an objective way. This study is not limited to coupon samples, but will explicitly lay the bridge towards real composite parts with industrial complexity.