 
Summary: Target recognition in hyperspectral images
Amir Z. Averbuch Michael V. Zheludev Valery Zheludev
School of Computer Science
Tel Aviv University, Tel Aviv 69978, Israel
Abstract
We present new algorithms that perform unmixing in hyperspectral images and then
recognize targets whose spectral signatures are given. The target can occupy sub or above
pixel. These algorithms combine ideas from algebra and probability theory. Experimental
results demonstrate the efficiency and the robustness of these algorithms on real hyperspectral
data.
1 Introduction
1.1 Data representation and extraction of spectral information
We assume that an hyperspectral signature of a sought after material is given. In many appli
cations, a fundamental processing task is to automatically identify pixels whose spectra have a
specified given spectral shape (signature). This problem raises the following issues: How the mea
sured spectrum of a ground material is related to a given "pure" the spectrum and how to compare
between them to determine if they are the same? As a result of spatial and spectral sampling,
airborne hyperspectral imaging sensors produce a 3D data structure referred to as a datacube.
The observed spectral radiance data, or the derived surface reflectance data, can be viewed
as a scattering of points in a Kdimensional Euclidean space RK
