Content-dependent on-the-fly visual information fusion for battlefield scenarios

摘要

We report on cooperative research program between Army Research Laboratory (ARL), Night Vision and Electronic Sensors Directorate (NVESD), and University of Maryland (UMD). The program aims to develop advanced on-the-fly atmospheric image processing techniques based on local information fusion from a single or multiple monochrome and color live video streams captured by imaging sensors in combat or reconnaissance situations. Local information fusion can be based on various local metrics including local image quality, local image-area motion, spatio-temporal characteristics of image content, etc. Tools developed in this program are used to identify and fuse critical information to enhance target identification and situational understanding in conditions of severe atmospheric turbulence.