The end goal of the project is to:
- provide answers on applicability of different Copernicus products for urban vegetation monitoring
- develop a software prototype for waterfront vegetation monitoring using Copernicus data but with higher tree density accuracy and higher temporal resolution than existing solutions
The main objectives are:
- To prepare a user-friendly report on accuracy and reliability of Copernicus data products using Riga as a study site (VHR and UAV data will ensure basis for detailed analysis).
- To develop and evaluate deep learning based land cover class fraction estimation for Sentinel-2 data
- To develop a software prototype based on end-users needs (Rigas Mezi will represent the end user in this project) including 4 submodules: 1) land cover class fraction (at least tree density) estimation; 2) water greenness estimation to analyse reed and algae distribution, 3) pixel based and 4) sector based change detection.
To provide a demonstration of this prototype in the form of produced data layers and popular science videos for Riga city and accuracy assessment using VHR historical data to analyse error sources.
Participating scientists
