IECS remote sensing group


Automated identification of mires and peatlands using multi-temporal satellite data

Prime Contractor: Institute of Electronics and Computer Science
Subcontractor: SIA “SWH SETS”
Project number: 4000128570/19/NL/SC
Scientific manager: Ints Mednieks
Scientific manager of the subcontractor: Ilze Bērziņa
Time frame: from November 1, 2019 to October 31, 2021

MireClass project aims at creation of a Web application for monitoring and analysis of current state and dynamics of raised bogs and peatland inventory in an easy and convenient way using multispectral and Synthetic Aperture Radar data and Dynland unsupervised classification technology. The main objectives of the project are:

  • Develop a methodology for automatic and semi-automatic identification of selected mire habitats and peatlands by classification of satellite data using WiseClust technology and integrated expert knowledge;
  • Create specific tools for accurate automatic or semi-automatic classification of multi-temporal data into specialised subclasses of mires, which present interest to industry and public sector, using sample areas;
  • Develop a Web application providing tools for classification of mires and peatlands  and viewing the results;
  • Create a reliable 20m^2 per pixel binary (present or absent) map of raised bogs and peatlands for the whole territory of Latvia.