Abstract
Abstract. The European COPERNICUS program provides an unprecedented breakthrough in the broad use and application of satellite remote sensing data. Maintained on a sustainable basis, the COPERNICUS system is operated on a free-and-open data policy. Its guaranteed availability in the long term attracts a broader community to remote sensing applications. In general, the increasing amount of satellite remote sensing data opens the door to the diverse and advanced analysis of this data for earth system science.However, the preparation of the data for dedicated processing is still inefficient as it requires time-consuming operator interaction based on advanced technical skills. Thus, the involved scientists have to spend significant parts of the available project budget rather on data preparation than on science. In addition, the analysis of the rich content of the remote sensing data requires new concepts for better extraction of promising structures and signals as an effective basis for further analysis.In this paper we propose approaches to improve the preparation of satellite remote sensing data by a geo-database. Thus the time needed and the errors possibly introduced by human interaction are minimized. In addition, it is recommended to improve data quality and the analysis of the data by incorporating Artificial Intelligence methods. A use case for data preparation and analysis is presented for earth surface deformation analysis in the Upper Rhine Valley, Germany, based on Persistent Scatterer Interferometric Synthetic Aperture Radar data. Finally, we give an outlook on our future research.
Highlights
During the last decades satellite remote sensing has become an important tool both in scientific earth observation and in data provision for informed decisions in politics and public administration
Access to the data is provided by means of the Data and Information Access Services (DIAS), which provide basic functionalities to download the data and to process them to some degree
Of a few techniques in AI that we can use in pattern analysis for remote sensing radar data are artificial neural networks (ANNs) and kernel methods such as support vector machines (SVM), which utilize kernels to complete nonlinear regression or pattern classification (Haworth et al 2014)
Summary
During the last decades satellite remote sensing has become an important tool both in scientific earth observation and in data provision for informed decisions in politics and public administration. For this purpose, the European Commission established the COPERNICUS® programme in 2014. A multitude of satellite remote sensing data are available free and open - on a long-term perspective This allows the full coverage of the earth’s surface with a high temporal resolution. Characteristic phenomena can be searched across different regions and for different time steps This refers, e.g., to the automatic selection of model components in data processing such as for the description of changes for interesting regions.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call