为使数字图像相关方法适用于大变形测量,提出一种变形初值估计方法。通过在变形前图像中第一个计算点附近选择三个或更多特征点并在变形后图像中选择其对应点,利用坐标对应关系获得该点可靠的变形估计值作为Newton-Rapshon方法的迭代初值。该方法可对被测物体存在任意刚体转动或大变形情况下的表面变形进行准确测量,克服了通常的基于整像素位移相关搜索再进行亚像素位移测量的数字图像相关方法在被测物面存在稍大转角或大变形情况下即不能使用的缺点。对存在相对刚体转动图像的位移场和聚丙烯泡沫塑料压缩实验的大变形进行了测量,结果充分显示本文方法的有效性。

In order to measure large deformation using digital image correlation method, a new initial guess technique is presented. The proposed technique obtains the reliable initial deformation guess of the first calculation point by manually selecting three or more points with distinct characteristics surrounding the reference subset center in the reference image and their corresponding locations in the deformed image. Then, reliable initial deformation guess can be determined according to the coordinate correspondence and is used as the iteration initial value of Newton-Rapshon method to further improve measurement accuracy. The proposed technique overcomes the shortcoming of the commonly used digital image correlation method which cannot work under the condition that the tested object has a rotation angle larger than seven degrees or large deformation. The measured displacement fields from image pairs with relative large rigid body rotation and large deformation compression test of polypropylene foam obviously demonstrate the validity of this technique.