Abstract: The ability to detect targets, either by human observer or by machine vision algorithms, is strongly affected by the clutter properties of the background. Therefore, the quantitative characterization of the clutter is a must in any target acquisition scenario. This paper will propose a new metric for clutter quantification based on blob analysis algorithm. The algorithm consists of comparison of various calculated properties of the blobs, found in the image, to the properties of the blob representing a target. The algorithm was tested on an experimental database of IR images measured during an air-born field test by a commercially available Focal Plane Array IR camera.!14
展开▼