Thermal Wave Imaging (TWI) is a non-destructive evaluation (NDE) technology which uses time-dependent heat flow as a tool for investigating material properties and subsurface features of solid objects. This method has been successfully used in NDE for various materials and also it has been used for metal loss measurement. In this dissertation, we use 'two-sides' (through thickness) experimental setup and use Peak Slope Time (PST) and Half Maximal Time (HMT) to calculate thickness, we combined the temporal information and 2D wavelet denoising technology to reduce noise level in original noisy sequence, and then use polynomial curve fitting to calculate the corresponding PST and HMT, we found that HMT is a much better method than PST, and also we got very good results both from simulated and experimental image sequences.; In the sonic IR imaging NDE technique, a short ultrasonic pulse (typically a fraction of a second) is applied to the target to excite the defects to heat up, while an infrared camera images the time evolution of the heating effect to identify the defective areas in the target. But this heating is typically a fraction of a degree, so the image is quite noisy which makes the detection really difficult sometimes. In this dissertation, we use wavelet domain thresholding and morphological operators to improve image quality and enhance the detectability of small defects.
展开▼
