鉴于Newton-Raphson (N-R)迭代法的数字图像相关方法(DIC)对初值敏感等问题，提出了一种使用SURF(speeded up robust features）特征匹配的数字图像相关方法。SURF算法能匹配出变形前后图像的特征点对，并获得点对的坐标值，使用与匹配点对所在的子区相对应的仿射变换来初始估计子区的变形参数，获得兴趣点的估计值。根据估算的初值进行迭代优化，得到使归一化最小平方距离相关函数(ZNSSD)最小化的兴趣点位移值。实验中，分别用该法及传统的基于尺度不变特征变换(SIFT)的初值估计方法对鲤鱼鳞片拉伸变形图样进行处理，结果表明所提的初值估计方法更加精确有效，并能够使后续的N-R迭代优化快速收敛。

Since the digital image correlation (DIC) method based on the Newton-Raphson (N-R) method is sensitive to initial values, a reliable initial guess method which utilizes an image feature matching algorithm of speeded up robust features (SURF) is presented. The SURF algorithm can match the feature points of the images before and after deformation, and get the coordinate positions. The deformation parameter of a point of interest is initially estimated from the affine transformation which corresponds to the subset region of matched feature points. By iterating and optimizing the estimated values, the displacement which makes the zero-mean normalized sum of squared difference (ZNSSD) minimized is obtained. In the experiment, both the presented method and scale-invariant feature transform (SIFT) algorithm are used in the deformation measurement of carp scales stretching. It is shown that the proposed initialization method is more sufficiently accurate and can enable the subsequent N-R method to converge quickly.