A new method to deal with the effect of subset size for digital image correlation

A novel 2D digital image correlation technique is developed to deal with the effect of subset size and to improve the calculation accuracy. In this paper, a new method for efficient using pixels is discussed and the relations between pixels and the central pixel are investigated. A relatively large subset size is chosen in the method and all the pixels in a subset are not treated equally. Coefficients are set to each pixel rely on the importance of them to identify a subset from target images which are calculated by the normal distribution based on the distances between them and the central pixel. Compared with traditional digital image correlation, the method is a tradeoff between using a large subset size and a small one. Selection of subset size for traditional digital image correlation is replaced by selection of parameters for the equation in this method. The effect of subset size to correlation calculation is somewhat alleviated and the calculation accuracy is improved. Computer simulated, random speckle images are used to test the proposed technique and good results are reported.