This unit provides an introduction to dimensionality reduction methods related to hyperspectral data or imaging spectroscopy. First, feature engineering methods using selected parametric and nonparametric regression methods are presented. Second, feature selection or extraction, using filter, wrapper and embedded methods will be discussed. Finally, some alternative methods are presented.
This tutorial focuses on regression-based unmixing of urban land cover from simulated EnMAP imagery. The slide collection provides the theoretical foundation for the tutorial, including general introductions into urban land cover mapping and regression-based unmixing using synthetically mixed training data. The practical provides hands-on training for working with the EnMAP-Box, including a basic introduction into its functionalities and spectral library handling, as well as the regression-based unmixing processing chain.
The EnMAP-Box is a free and open source plug-in for QGIS. It is designed to process imaging spectroscopy data and particularly developed to handle data from the upcoming EnMAP satellite. The EnMAP-Box provides state-of-the-art applications for the processing of high dimensional spectral and temporal remote sensing data and a graphical user interface that enhances the GIS oriented visualization capabilities in QGIS by applications for visualization and exploration of imaging spectroscopy raster data and spectral libraries.