Thermal-wave image processing for characterization of subsurface flaws in materials

Infrared images resulting from back-scattered thermal waves in composite materials are corrupted by instrument noise and sample heat-spread function. This paper demonstrates that homomorphic deconvolution and "demultiplication" result in enhanced image quality for characterization of subsurface flaws in Kevlar and graphite composites. The choice of processing depends on material characteristics and extent of noise in the original image.<
>