Analýza post-požárových změn vegetace v Národním parku České Švýcarsko pomocí hyperspektrálních dat EnMAP

Abstract

The bachelor's thesis focuses on the use of hyperspectral data from the EnMAP satellite mission for the analysis of post-fire vegetation changes in Bohemian Switzerland National Park following the large wildfire of July 2022. The theoretical part describes the fundamental properties of hyperspectral data and their differences from multispectral data, provides an overview of current hyperspectral satellite missions, and addresses the possibilities of using hyperspectral data for vegetation analysis and assessment of its condition. An introduction to Bohemian Switzerland National Park and the progression of the wildfire are also included. The practical part covers the pre-processing of EnMAP imagery, the construction of a spectral library, and supervised classification of vegetation types using the Random Forest algorithm, with accuracy assessed by means of a confusion matrix. Post-fire vegetation changes were detected using the RdNBR index and a post-classification comparison of images from 2022 and 2024. The results allow the extent of damage to the vegetation cover to be quantified and the degree of its recovery approximately two years after the fire to be assessed.