一种基于预变形辅助前向加性高斯牛顿法的数字图像相关方法  

A digital image correlation method based on pre-deformation forward additive gauss-newton method

作　　者：叶敬晨 计宏伟[1] YE Jingchen;JI Hongwei(School of Mechanical Engineering,Tianjin Universityof Commerce,Tianjin 300134,China)

机构地区：[1]天津商业大学机械工程学院,天津300134

出　　处：《实验力学》2022年第5期621-628,共8页Journal of Experimental Mechanics

基　　金：国家自然科学基金面上项目(12172254)资助。

摘　　要：尽管数字图像相关(Digital Image Correlation, DIC)方法已经取得了显著进展,但提高DIC的测量精度仍然是关键问题之一,特别是在非均匀变形条件下。本文将变形分解为预变形和后变形两部分,进而将DIC位移求解过程分解为两个子过程:预变形估计和后变形迭代。借助预变形辅助策略将前向加性高斯牛顿(Forward Additive Gauss-Newton, FA-GN)算法推广为pFA-GN (Pre-deformation FA-GN, pFA-GN)算法。数值实验结果表明,本文提出的pFA-GN算法显著提高了DIC在非均匀变形条件下的测量精度。Although digital image correlation(DIC) has made significant progress, further improving the measurement accuracy of DIC is still one of the focuses, especially under non-uniform deformation. This paper decomposes the deformation into pre-deformation and post-deformation, and then the DIC displacement solution process will be decomposed into two sub-processes, pre-deformation estimation and post-deformation iteration. Forward Additive Gauss-Newton(FA-GN) algorithm is generalized as Pre-deformation FA-GN(pFA-GN) algorithm by means of pre-deformation assistance strategy. Numerical experiment results show that the proposed pFA-GN algorithm significantly improves the measurement accuracy of DIC under non-uniform deformation.

关 键 词：数字图像相关 预变形 非均匀变形 

分 类 号：O348.1[理学—固体力学]