Application of curvelet transform for denoising in electronic speckle pattern interferometry fringe patterns

Abstract

Denoising is the key problem in electronic speckle pattern interferometry (ESPI). In order to eliminate noise, to enhance the qualities of fringe patterns, we proposed a new filtering method which is based on curvelet transform for ESPIAnd its application to electronic speckle fringe filtering. Firstly, fringe pattern is performed on by curvelet transform;Secondly, the method which is combined with soft and hard threshold is applied to remove coefficient of noise;so as to separate signal from noise, threshold process is conducted by curvelet coefficient of square. Finally, the filtered image is obtained by reconstruction of coefficient of curvelet transform. This new method has applied on one computer-simulated fringe pattern and one experimentally obtained pattern of the filter, and compare to wavelet filtering and windowed Fourier filtering whose models are extensive used. The index that filtered results is evaluated with mean square error (MSE). Experimental results demonstrate that the method for filtering in ESPI has the advantages of high precision, fast filtering speed, simplicity in implement, and suitable for measurement of real time.