Assessment of a lake surface elevation from optical satellite images

Abstract

In this paper, we focus on the implementation of an advanced geographic information system that enables the automatic extraction of annual levels of lake surface elevations. The implementation of the proposed system is based on the integration of the Open Access Hub platform which allows access to Copernicus open data, specifically the multispectral optical images of a pair of Sentinel 2 satellites. These make possible measurements with a five-day temporal and ten-meter spatial resolutions. The key building blocks of the proposed system are the subsystem for automatic data harvesting and preprocessing comprised of data decomposition techniques, image cropping and spatial alignment, digital terrain model or point definition of the altitudes of the terrain surrounding the monitored lake, cloud image recognition and a filtering algorithm, a land cover recognition algorithm, water index calculations and lake level segmentation, together with a system for mapping the recognized lake surface boundaries to altitudes. The results demonstrate that the proposed approach achieves sufficient precision for practical use.