PRECISE MATCHING OF 3-D TARGET MODELS TO MULTISENSOR DATA

This paper presents a three-dimensional (3-D) model-based ATR algorith m that operates simultaneously on imagery from three heterogeneous, ap proximately boresight aligned sensors, An iterative search matches mod els to range and optical imagery by repeatedly predicting detectable f eatures, measuring support for these features in the imagery, and adju sting the transformations relating the target to the sensors in order to improve the match, The result is a locally optimal and globally con sistent set of 3-D transformations that precisely relate the best matc hing target features to combined range, IR, and color images, Results show the multisensor algorithm recovers 3-D target pose more accuratel y than does a traditional single-sensor algorithm. Errors in registrat ion between images are also corrected during matching.