Application of image analysis techniques and numerical modelling to analyse transient thermographic data

A quantitative image analysis study of the defect imaging response of pulsed video thermography (PVT) has been presented. A dedicated numerical finite difference model has been developed that truly computes the complex 3-dimensional heat flow around a circular subsurface defect. The accuracy and validity of the numerical computations have been checked and found to agree with the simple 1-dimensional thermal wave theory in the appropriate limit. The numerical analysis presented has further supported the authors´ earlier analysis of the defect imaging process of PVT based on a simple finite defect first order thermal wave theory. The discrepancies between the experimental PVT data and the model prediction at long times (after contrast peak) are thought to be contributed by the actual geometric shapes and sizes of the defects. Corrosion products could have accumulated around the edge and interior of the defects within the mild steel test piece. This effectively shortened the thermal diffusion path of the long wavelength longtime components across the defect, hence diminishing the image contrast