A Method Based on Wavelet-transform in Contour Window for Filtering Electronic Speckle Pattern Interferometry Fringes

The speckle fringes in Electronic Speckle Pattern of Interferometry (ESPI) have to be filtered before interpretation due to a lot of speckle noise mixed in them. The typical filtering method based on average filtering and median filtering within contour window have the difficulty of filter parameter selection, so the problem of damage stripe information often occurs. To solve the problem, a method based on wavelet-transform in contour window for filtering noise is put forward. This method includes three steps. Firstly, Wavelet-transform is carried out within the contour window. Then a judgment whether there are fringe boundaries within the contour window is proceeded. If there are no fringe boundaries within the contour window, then low-pass filtering is applied. If there are fringe boundaries in the contour window, for all pixels on the fringe boundaries, the high frequency information of wavelet-transform sub-bands will be retained. Then an average filtering is performed for these pixels on the boundaries, and a low-pass filtering is performed for other pixels in the contour window. This method can provide more protection for the stripe boundary.