Colored visualization of multitemporal SAR data for change detection: issues and methods

Inhalt:
This paper proposes a method to generate a colorful product for stacks of SAR images, making it easy to see areas that are subject to change. Two modes of representation are proposed: one intended for Sentinel changes, on a large temporal scale; the other for TerraSAR-X images, allowing you to see point-time events. The importance of fine coregistration and temporal calibration is emphasized. Lastly, filtering methods, in particular by deep learning, based on the temporal redundancy of the data, are proposed.