Image Correspondence Figure of Merit (FOM)

The overarching goal of this Phase II effort is to simplify the job of an analyst when choosing a reference image for image registration. Analysts are typically provided a group of images that overlap a target area. They are tasked with choosing the best among those images to serve as a reference image. However, the qualities that separate good images from mediocre images are subtle and difficult to gauge by eye alone. There are high stakes involved with target acquisition. Any improvements we can make to accuracy and reliability will have immediate and long term benefits. This Phase II effort is aimed at taking the guesswork out of the analyst's job to improve accuracy and reliability by developing a Figure of Merit (FOM). Our approach scans an image using a condition theoretic approach to isolate strong line and corner information, present in the image. It also examines the distribution of that information about the image. We then quantify the quality of the image with respect to registration potential. We also provide a Graphical User Interface (GUI) to streamline the analysts interaction with the FOM. The GUI provides extended functionality to intuitively access and process data on demand.