Change Categorization in Short-Term SAR Time Series

Abstract:
Monitoring changes in time series data offers the opportunity to identify very frequently changing regions. As input data, SAR imagery has significant benefits against optical imagery, due to its independence against atmospheric effects and the acquisition time. The conventional way of time series analysis is given by the investigation of long-term image stacks, often covering a time span of weeks, months or even years. In the paper at hand, the categorization of changes detected in short-term image stacks is addressed. For this, several sub-aperture images are used, which are extracted from spaceborne SAR data, acquired with a long synthetic antenna.