Extention of the Grid Method to Large Displacements

摘要: The Grid Method is used from the beginning of nineties. This method based on well-known Moire effect: surface encoded with a periodic pattern. variations this pattern are analysed through phase either using global Fast Fourier Transform, or local approach, for example wavelet transform. efficient small displacements: typically, its resolution 1/100 grid step i.e. 5/100 pixels. classical subtraction performed at same geometrical point, which means that displacement assumption necessary. If large displacements achieved, image correlation technique great interest because it provides result in initial frame reference. For displacements, becomes difficult to implement, and sub-pixel interpolation larger windows have be used. More generally, typical spatial (8 64 pixels) appears than (from 4 17 pixels). Thus, latter could principle retrieve more localized phenomena. Even if basically, these two methods use different patterns (periodic random), possible decode information numerical strategies: transform random [1]; [2]. In second case, many maxima appear map, selection right one made by user himself, considering external knowledge. Nevertheless, solution widely spread among metallurgists community microscopic scale. particular marked very practical reasons. Then, situation following: large, but periodic. Instead technique, we propose here extend displacements. expected firstly an easier way extract secondly results better metrological qualities.