Existing subset-based digital image correlation (DIC) must rely on user-selected key calculation parameters (i.e., subset size and shape function) to proceed with displacement/deformation analysis. However, the lack of clear guidelines for selecting these parameters leads to varying choices among users, thus introducing artificial uncertainty in DIC measurements. Previous theoretical analyses and experimental studies revealed that optimal calculation parameters must account for both local speckle pattern quality and deformation at each required point. That means these key parameters should vary at each calculation point, posing a significant challenge for optimal parameter selection. To tackle this challenge, a novel Smart-DIC is proposed to achieve user-independent, accurate and precise displacement field measurements. This method comprehensively considers the local speckle pattern quality and local deformation at each calculation point, and gives the explicit formulas to determine optimal calculation parameters. Based on these optimal calculation parameters, Smart-DIC outputs accurate and precise displacement measurements without the need for users’ inputs. To validate its metrological performance, numerical tests were processed using data from DIC Challenge 1.0 and real images of tensile tests of a commercial AL-based alloy. The experimental results underscore the capacity of Smart-DIC to efficiently achieve accurate and precise displacement measurements across various scenarios by automatically selecting optimal subset sizes and shape functions.

中文翻译：

---

智能 DIC：独立于用户的、准确、精确的 DIC 测量，自适应选择最佳计算参数


现有的基于子集的数字图像相关（DIC）必须依赖于用户选择的关键计算参数（即子集大小和形状函数）来进行位移/变形分析。然而，缺乏选择这些参数的明确指南导致用户之间的选择不同，从而在 DIC 测量中引入人为的不确定性。先前的理论分析和实验研究表明，最佳计算参数必须考虑每个所需点的局部散斑图案质量和变形。这意味着这些关键参数在每个计算点都应该有所不同，这对最佳参数选择提出了重大挑战。为了应对这一挑战，提出了一种新颖的 Smart-DIC，以实现独立于用户的、准确且精密的位移场测量。该方法综合考虑各计算点的局部散斑图案质量和局部变形，并给出明确的公式来确定最佳计算参数。基于这些最佳计算参数，Smart-DIC 无需用户输入即可输出精确的位移测量结果。为了验证其计量性能，使用 DIC Challenge 1.0 的数据和商用铝基合金拉伸测试的真实图像进行了数值测试。实验结果强调了 Smart-DIC 通过自动选择最佳子集大小和形状函数，在各种场景下有效实现精确位移测量的能力。