Fault detection by interferometric fringe pattern analysis using windowed Fourier transform

Windowed Fourier transform (WFT), a tool of spatial-frequency analysis, is able to characterize the local frequency at any location in a fringe pattern. Changes in fringe frequency characterize defects in optical interferometric-based systems. Hence, the WFT is suitable for fault detection and condition monitoring in optical NDT. In this paper, a WFT-based algorithm is described and theoretically analysed for fault detection. This is followed by demonstrations using both simulated and real fringe patterns. Comparisons with the traditional Fourier transform approach and normalized cross correlation approaches are also carried out.