Algorithm Development for WFOV Mission Data Processing

ABSTRACT: Etegent is proposing development against two algorithm areas from the solicitation, jitter suppression and clutter suppression.For the jitter suppression Etegent is proposing to apply their subpixel-resolution, scene-based cross correlation technique to measure in-scene jitter, combined with an image-filter based technique to reproject the images to a common alignment with minimal pixel spreading.Both techniques have performed well in past efforts and are highly optimized in the Matlab environment, but will require expansion to a GPU implementation and may require additional capability enhancement.For clutter suppression Etegent will apply a principal components background suppression (PCBS) algorithm.This too will be expanded to a GPU implementation, along with a number of candidate approximations to render it suitable for a real-time processing budget.Of course, these proposed techniques are a starting point; actual algorithm development will be guided by the needs of the furnished data sets and may change during the work.; BENEFIT: The research conducted under this Phase I effort will lay the groundwork for the fuller integration and testing of a set of real-time processing algorithms at the Space and Missile Systems Centers Falcon Shield algorithm research facility.These algorithms will be well vetted and should readily lead to real-time operation on off-the-shelf workstation class machines with advanced GPU capabilities.These algorithms will address two significant portions of the real-time processing chain, taking nearly raw sensor data and providing a cleaned up data stream with jitter effects mitigated and static background scene content removed.Etegent also sees potential for broader application within the intelligence community (e.g. NASIC, NGA) and in their own commercial software for nondestructive inspection data management (NLign).