Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

1 Publication

Feature Extraction for Hyperspectral Imagery: The Evolution From Shallow to Deep: Overview and Toolbox

Rasti, B.; Hong, D.; Hang, R.; Ghamisi, P.; Kang, X.; Chanussot, J.; Benediktsson, J. A.

Hyperspectral images (HSIs) provide detailed spectral information through hundreds of (narrow) spectral channels (also known as dimensionality or bands), which can be used to accurately classify diverse materials of interest. The increased dimensionality of such data makes it possible to significantly improve data information content but provides a challenge to conventional techniques (the so-called curse of dimensionality) for accurate analysis of HSIs. Feature extraction (FE), a vibrant field of research in the hyperspectral community, evolved through decades of research to address this issue and extract informative features suitable for data representation and classification. The advances in FE were inspired by two fields of research—the popularization of image and signal processing along with machine (deep) learning—leading to two types of FE approaches: the shallow and deep techniques. This article outlines the advances in these approaches for HSI by providing a technical overview of state-of-the-art techniques, offering useful entry points for researchers at different levels (including students, researchers, and senior researchers) willing to explore novel investigations on this challenging topic. In more detail, this article provides a bird’s eye view of shallow [both supervised FE (SFE) and unsupervised FE (UFE)] and deep FE approaches, with a specific focus on hyperspectral FE and its application to HSI classification. Additionally, this article compares 15 advanced techniques with an emphasis on their methodological foundations and classification accuracies. Furthermore, to push this vibrant field of research forward, an impressive amount of code and libraries are shared on GitHub, which can be found in [131].

Related publications

Downloads

Permalink: https://www.hzdr.de/publications/Publ-31906
Publ.-Id: 31906