Fractional Mapping of Savannah Vegetation Species using Drone and EnMap Hyperspectral Data

Karakizi, Christina, Okujeni, Akpona, Karantzalos, Konstantinos, Hostert, Patrick and Symeonakis, Elias ORCID: https://orcid.org/0000-0003-1724-2869 (2024) Fractional Mapping of Savannah Vegetation Species using Drone and EnMap Hyperspectral Data. In: 43rd EARSeL Symposium, 17 June 2024 - 20 June 2024, Manchester, United Kingdom. (Unpublished)

Preview

Accepted Version Available under License In Copyright. Download (380kB) | Preview

Abstract

In order to meet the UN Sustainable Development Goal for Land Degradation Neutrality (LDN) by 2030 it is imperative to accomplish an assessment of the state of savannahs as early and as accurately as possible. Monitoring the species composition of vegetation is also essential for land degradation assessments, while respective spatially distributed species composition maps are still unavailable. UAV technologies employing hyperspectral (HS) cameras are increasingly employed for mapping different woody species, but their high cost has so far impeded their use for continuous monitoring. The recent launch of the EnMap HS satellite sensor could significantly improve our ability to distinguish between different savannah woody species at larger scales. This works targets the accurate mapping of the fractional cover of savannah vegetation components, further analysing the woody component into different woody species. Experiments are performed over a South African (SA) savannah, in Kalahari region using EnMap HS data in combination with very high resolution multispectral drone data.