Sparse and Low-Rank Matrix Decomposition for Automatic Target Detection in Hyperspectral Imagery

TL;DR

This paper introduces a method for automatic target detection in hyperspectral images by decomposing the image into low-rank background and sparse target components using a pre-learned spectral dictionary.

Contribution

It proposes a novel decomposition approach leveraging sparse and low-rank matrix techniques with a pre-learned target dictionary for hyperspectral target detection.

Findings

Effective separation of targets from background demonstrated on synthetic data.

Method performs well on real hyperspectral imagery.

Two strategies for target detection are evaluated.

Abstract

Given a target prior information, our goal is to propose a method for automatically separating targets of interests from the background in hyperspectral imagery. More precisely, we regard the given hyperspectral image (HSI) as being made up of the sum of low-rank background HSI and a sparse target HSI that contains the targets based on a pre-learned target dictionary constructed from some online spectral libraries. Based on the proposed method, two strategies are briefly outlined and evaluated to realize the target detection on both synthetic and real experiments.