 |
|
|
Agent Air Force Research Labs: Dr. Tim Ross, Mr. Mark Minardi
Program Goals
The MSTAR program is developing a next-generation synthetic aperture radar (SAR)
automatic target recognition (ATR) system based on model-based vision
technology. This approach aims to provide a robust, scaleable ATR solution
for highly unconstrained imaging scenarios by modeling targets as well
as the associated uncertainties which arise due to target variation, articulation,
occlusion, and interaction with local scene objects. The MSTAR Program
has concentrated on recognizing GOB targets using high resolution 1ft SAR imagery,
although the techniques it is developing are readily extendible to a variety
of ATR scenarios.
The model-based approach utilizes offline-derived target
models in an online hypothesize-and-test manner that repeatedly compares predicted
statistics of target signatures with image data in an attempt to find the hypothesis
that maximizes the likelihood of the observed data. This processing involves
functions for Predicting Features, Extracting Features, Matching Features and
Searching over Hypotheses. In MSTAR, these functions are organized into
the so-called P-E-M-S Loop. ALPHATECH has had the leading role in the
design and development of MSTAR's PEMS loop.
ALPHATECH's MSTAR Programs
ALPHATECH is developing what are arguably the two most important components
of the MSTAR system, the Search and Match modules, as well as performing some
path-breaking work in ATR performance theory. Background information related
to these efforts may be found at:
Search Module
Points of Contact:
Dr. John Wissinger, MSTAR Program Manager, Search PI: (781)-273-3388,
x239
Mr. Robert Ristroph, Search Software Lead: (781)-273-3388, x339
Dr. Gil Ettinger, Match PI: (781)-273-3388, x334
Mr. Greg Klanderman, Match Software Lead: (781)-273-3388, x337
Dr. William Irving, PET PI: (781)-273-3388, x285