EMSI has its roots in Air-to-Air and Ground-to-Air Inverse Synthetic Aperture Radar (ISAR) imaging of aircraft. As part of this work, we developed the novel process of Distortion Invariant Matching. We developed this process in the late 90s to to fully automate radar image processing applications that previously required extensive human operator intervention. In particular, we developed algorithms and code to automatically register image templates to radar imagery. This automation reduced the labor intensive image processing time from two hours to about two seconds! The process we developed is at a Technology Readiness Level (TRL) 9 and is still in use by the Air Force.
Our Distortion Invariant Matching (DIM) algorithm is illustrated below. It optimally registers two feature sets subject to a parameterized transformation. In radar feature matching it is typical to use the transformations of rotation, translation, and crossrange scale factor. We use this in our classification process to compare ISAR data of two vessels or vehicles.