Detection of Step Changes in a Long Sequence of SAR Images

Inhalt:
In this work, we focus on the detection of fast rate changes, like construction/destruction of buildings and other man-made infrastructures, which can be easily modelled through step pattern changes of the RCS (Radar Cross Section). Considering a sample dataset made of a long sequence of ENVISAT images, the achieved performance of the change detection algorithm are evaluated both in terms of detection/false alarm probability both through a comparison with other proposed approaches.