By developing remote sensing based algorithm for land cover mapping in Latvia, the Institute for Environmental Solutions (IES) closes the third stage of the simulation of European Space Agency (ESA) satellite Sentinel-2 images for land cover/ land use (LC-LU) monitoring using hyperspectral remote sensing (SentiSimuLat project).
The aim of SentiSimuLat is to develop improved LC-LU classification algorithm based on Latvian user needs analysis. At the moment, Latvia performs land cover mapping as a part of the Corine Land Cover (CLC) initiative which is an adequate standardised solution for the European level, but its low resolution data and sporadic updates limits its applicability and uptake opportunities at the national level. Furthermore, there is no remote sensing based service for LC-LU in Latvia, which could provide valuable information of the country’s natural resources. Therefore, IES is developing an improved semi-automatic algorithm for LC-LU classification in Latvia that would also contribute to CLC mapping.
To gather the base of the remote sensing information for the algorithm, during the 2015 and 2016 seasons IES organised several remote sensing data acquisition missions in Cēsis, Burtnieki and Sigulda municipality pilot sites. Data were acquired by using the unique airborne remote sensing system ARSENAL which is able to simulate the performance of Sentinel-2 satellite. The validation of satellite and airborne remote sensing data was performed by using in-situ measurements.
The land cover mapping scheme will include the information, classes (e.g. grasslands, wetlands, agricultural lands, artificial objects, bare land areas etc.) and levels of details that are important for various sectors in Latvia, such as agriculture, forestry, regional development, nature protection and others. The first draft version of the algorithm has already been created and tested. The final stage of SentiSimuLat will see work being devoted to further improving the algorithm.
IES expects that the upcoming algorithm for land cover classification will provide more user-friendly interface, as well as automatic reading and writing of data sets. The algorithm will be developed to be used by people with medium level knowledge of geospatial data use. It includes specialists from Latvian organisations, like the Rural Support Service, the Latvian Rural Advisory and Training Centre, Ministry of Agriculture, Ministry of Environment and Regional Development, Latvian Environmental, Geology and Meteorology Centre and others.
The simulation of Sentinel-2 performance and development of remote sensing based algorithm for LC-LU mapping will contribute to the development of innovative earth observation products and services. Furthermore, it will improve the knowledge and skills of Latvian scientists and practitioners in acquiring, processing, analysing and applying earth observation data.
The SentiSimuLat project (Simulation of Sentinel-2 Images for Land Cover/ Land Use Monitoring Using Hyperspectral Airborne Remote Sensing) No. LV1-29 is funded by the Government of Latvia and ESA Contract under the PECS (Plan for European Cooperation States). The view expressed herein can in no way be taken to reflect the official opinion of ESA.
More about the SentiSimuLat project