A Multiscale Multispectral Approach to Digital Image Correlation for SHM Applications

Abstract

A novel technique for digital image correlation (DIC) targeting structural health monitoring (SHM) applications is presented. The method uses multispectral imaging enabling simultaneous data acquisition at variable fields of view (FOV). In general, for a given FOV, appropriate speckle sizes have to be a priori defined, limiting the capability of acquiring viable DIC data from variable working distances or with variable accuracy. Furthermore, straightforward application of multiscale DIC patterns that can be seen in the same wavelength of light can cause measurement errors, forcing the use of larger subset sizes and thus reducing the spatial resolution of the deformation measurements. To overcome such limitations, a patterning technique accounting for camera resolution, distance to target, metric-based optimization and variable wavelength is developed. Specifically, patterns are created for two length scales and applied onto a structure. A black on white speckle pattern for far field measurements in addition to an ultraviolet pattern for near field measurements. The patterns were shown to optimally perform at the a priori calculated working distances. The potential of this method for use in SHM applications is discussed.