A performance assessment on the effectiveness of digital image registration methods
Steve T. Kacenjar, Bing Li, Alan Ostrow · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2011
Digital Image Correlation (DIC) of time-sequenced-imagery (TSI) is a very popular method in the study of medical, material deformation, and electronic packaging. Its use in processing the before-and-after images provides critical information about the scene deformation and structural differences between the imagery. Several correlation methods for implementing DIC have been developed and will be compared in this study. Each of these methods offer distinct trades offs with respect to processing complexity and lock-in accuracy. There are several factors that influence the effectiveness of these methods to provide robust operation and strongly localized correlation peaks. These factors include; camera positional stability during the time of image acquisitions, deformation of the object under study, and measurement noise. In addition, the signatures that are captured during DIC can often times be amplified through preprocessing and thus potentially enhancing DIC performance. This paper examines the impacts on two of these factors (measurement noise and image digital sharpening) using four popular correlation methods that are often implemented in DIC analyses.