Image feature based quality assessment of speckle patterns for digital image correlation measurement

A novel method is proposed to assess the quality of speckle patterns for deformation measurement using the digital image correlation (DIC) technique. Different from existing methods that focused on the characteristics of individual speckle or frequency of grayscale values, our approach explores the usage of the image features essential for image registration, which is the basis of DIC. An indicator called density and evenness of features (DEF) is defined, combining the distribution density and evenness of image features. Numerical and real experiments demonstrate that the DEF is sensitive to the quality of various speckle patterns. It can detect the subtle difference between good speckle images leading to small gap in measurement accuracy. The DEF can also provide reliable guidelines to design high-quality transferable speckle patterns, as it shows clear relation to the main controllable parameters in generation of speckle patterns, including speckle duty ratio, image contrast, speckle radius and its dispersion, as well as speckle edge sharpness.

Automated summary

A novel method is proposed to assess the quality of speckle patterns for deformation measurement. This method utilizes image features essential for image registration and provides sensitive detection of subtle differences between speckle patterns. It also offers reliable guidelines for designing high-quality transferable speckle patterns.