Improvement of classifier and change detection tools in multi-temporal SAR images

Temporal SAR images appear to bring an efficient diversity and to be able to add to the informative content of several channels. With the ERS-1 SAR, temporal data are easily available in great number. The temporal data bring a fading diversity in order to improve the radiometric resolution while preserving the spatial resolution allowing a more efficient use of single-channel filters. Different statistical features are presented. These tools bring a more or less discriminating information according to the kind of data and the characteristic of the temporal set, allowing to distinguish various land surfaces. The methods described may be used for the discrimination of forested and nonforested areas. Another use of temporally complex data concerns interferometry.