Dynamic land use monitoring by fusion of satellite data – DynLand
“Funded by the Government of Latvia through an ESA Contract under the PECS (Plan for European Cooperating States)”
Prime Contractor: Institute of Electronics and Computer Science
Subcontractor: Latvian State Forest Research Institute “Silava”
Project number: 4000118771/16/NL/SC
Scientific manager: Dr.sc.comp. Ints Mednieks
Scientific manager of the subcontractor: Mg.biol. Juris Zariņš
Time frame: from December 1, 2016 to November 30, 2018
DynLand project aims at creation of dedicated software tools for dynamic land use monitoring exploiting satellite data that can be integrated in GIS software environment commonly used in governmental services and research institutions dealing with land use.
The objectives of the project are:
- develop specialized methods for processing satellite data (implementing fusion of multispectral images and Synthetic Aperture Radar (SAR) data) to obtain land use maps for rural areas and detecting changes between the current and earlier status;
- implement prototypes of software tools for producing the updated land use maps that can be integrated into the processing environment used in organizations related with land use monitoring;
- demonstrate that developed software can perform on the basis of combined multispectral and SAR data from different satellite missions (e.g. Sentinel-1, Sentinel-2, TerraSAR-X, WorldView-2);
- test produced software on the basis of available field data from the National Forest Inventory of Latvia and information layers that should be updated, e.g. cadastral parcel map and field blocks from the GIS system of the Rural Support Service of Latvia.
Project activities:
11.01.2017. Project banner prepared and presented to participants of the ESA information day (http://www.izm.gov.lv/images/zinatne/EKA/Projekti_rollup/DynLand_baneris.pdf ).
03.03.2017. Meeting of project teams from EDI and SILAVA. Preparation of the methodology report, characteristics of field data to be used and purchase of the high resolution satellite data discussed.
13.03.2017. Methodology report presented to ESA describing intended use of remote sensing data for reaching project objectives.
25.05.2017. Meeting of project teams from EDI and SILAVA. Preparation of the Software Requirements Specification and first Progress report, characteristics of field data and other supplementary data available for the design of land cover classifiers discussed.
26.05.2017. Software Requirements Specification and the first Progress report presented to ESA.
16.06.2017. High resolution satellite data (WorldView, TerraSAR-X) procured.
20.07.2017. An algorithm for unsupervised clustering of multispectral image pixels developed. Optimization of it for processing large images continues.
08.09.2017. A program for interactive mapping of clusters to defined categories developed.
01.02.2018. Telephone meeting held with ESA for presentation of the project status.
21.03.2018. The result of the land cover classification for the first sample image was prepared and handed over to partners “Silava” for independent accuracy testing.
19.06.2018. The first evaluation of the land cover classification results for the sample image was performed in “Silava”. The accuracy of the classification planned for the project has been reached.
17.08.2018. Development of software prototype for detection of overgrown areas finished.
17.08.2018. Pabeigta programmas prototipa izstrāde aizaugušu teritoriju detektēšanai.
01.10.2018. Presentation of project progress to ESA representatives at ESA PECS Annual Review meeting.
06.11.2018. Project execution extended to 31.03.2019.
06.01.2019. Development of software prototype for detailed classification of bogs has been completed.
15.03.2019. Completed testing and correction of the developed software.
26.04.2019. All Project closing documentation sent to ESA.
Participating scientists
