Data Processing in TNDT

Numerous mathematical techniques to enhance the signal-to-noise ratio in IR image sequences are discussed. It is shown that the most efficient method is to use temporal phenomena accompanied by the time evolution of pixel-based temperature functions. Standard image processing algorithms are summarized. Also, advanced processing techniques including pulse phase thermography, thermal wave lock-in technique, principal component analysis, polynomial fitting, correlation analysis, thermal tomography, neural networks and more exotic methods, are discussed. Some novel TNDT techniques, including vibrothermography, eddy current, and electromagnetic IR thermography are described. Particular attention is paid to the analysis of thermo-mechanical phenomena in solids.