Dimensionality reduction of imaging spectroscopy data

This unit introduces dimensionality reduction methods related to hyperspectral data or imaging spectroscopy. Following an overview of feature engineering methods that use selected parametric and nonparametric regression methods, feature selection or extraction using filter, wrapper, and embedded methods are discussed. Lastly, alternative methods are presented.

This unit was originally published in March 2020. The revised version is accessible since February 2023. It is recommended to study the unit “Principles of imaging spectroscopy” before starting this unit.

How to cite this slide collection: K. Berger, C. Atzberger, M. Danner, M. Wocher, T. Kuester, S. Foerster (2020). Dimensionality reduction of imaging spectroscopy data – Solutions to deal with the high dimensionality of hyperspectral data. HYPERedu, EnMAP education initiative, German Centre for Geosciences GFZ; originally published March 2020, revised February 2023.

Please help us to further improve the hyperspectral resources and send us your feedback to hyperedu@eo-college.org