Towards Quantifying Clutter in Hyperspectral Infrared Images

O.O. Fadiran, P. Molnár, and L.M. Kaplan (USA)


Clutter complexity, hyperspectral imagery, factor analysis


We develop a method to quantify and characterize clutter in hyperspectral infrared (HSI) images in a framework similar to work done on single-band images. Hereby, all objects in a scene that may be mistaken for targets by an Automatic Target Recognizer (ATR) are considered clutter. A hyper spectral image contains a number of contiguous discrete bands within the spectrum. The aim is to obtain a measure of complexity for hyperspectral images, which will indi cate the inherent difficulty for an ATR to detect targets. We implemented 129 different image clutter metrics, and com puted them for a database of synthesized HSI images. A matched filter ATR was used to determine the amount of clutter in the images as a baseline. We developed a method to select a subset of the metrics that in combination cor related best with the amount of clutter in an image, and defined this as the clutter complexity measure (CCM). Our results show that the CCM obtained from a random sample of images is a good predictor of the amount of clutter for the entire database.

Important Links:

Go Back