APPLICATION OF SENTINEL-1 RADAR DATA FOR MAPPING HARD-TO-REACH NORTHERN TERRITORIES

Abstract

The new European space satellites Sentinel-1A and 1B with C-band radars on board, launched in 2014 and 2016 respectively, provide regular radar data on the Earth’s surface with high temporal resolution. These new non-commercial data provides extensive opportunities for research of remote Arctic territories, poorly supplied with optical images due to cloud conditions. Difficulties in recognizing objects on radar images can be compensated for by the possibility of using multiple repeated surveys, which make it possible to identify areas of the terrain which are similar in character of changes. In the study, four Sentinel-1A images of the largest from the New Siberian islands – Kotelny – were used, which were acquired during the summer period from July 3 to August 20, 2015. After preprocessing aimed at improving the visual properties and coregistration of the multitemporal images, an automated clustering of the multitemporal image set was carried out. Clustering results were analyzed on comparison with additional sources of spatial information. Both specialized software for Sentinel-1 radar data processing - SNAP, and the GIS software complex ArcGIS were used. The latter provided the creation of the spatial data base for comparing the results of radar data processing and cartographic sources. The map of the territory zoning was obtained as clustering results which is based on the changes in the normalized radar cross section (sigma nought) over the summer period, and the approximate correspondence of the areas to the main types of the relief and landscapes of the island was established.

Keywords

• satellite radar data
• Sentinel-1
• data processing
• image interpretation
• Kotelny Island