Application of wavelet filters for feature extraction in interferometric fringe patterns

The fast and reliable localization and classification of fault indicating fringe patterns in interferometric images is a major task in holographic non-destructive testing. For the purpose of feature extraction from gray value images, wavelet transformation has proved to be a suitable tool. In contrast to the Fourier transformation the local feature information will be preserved and furthermore the applied transforming wavelet can be adapted--under certain constraints--to the given problem.