Image Manipulation to Incorporate Uncertainty Into Image-Correlation Data Association Technique

Abstract

An image correlation technique for target data association is augmented to incorporate track and measurement uncertainties. In this development, the uncertainty of target representations is assumed to be represented by a Gaussian distribution. The new approach uses fuzzy logic to simplify the mapping of the uncertainty to a usable construct to adjust the image generation based on the uncertainty of the measurements. The result of the fuzzy logic process is used to restructure image representation of the attribute of the target. Large uncertainty measures have the effect of reducing the relative size of the attribute which, in turn, reduces the impact of the attribute of the track-measurement pair.