A comparison of target decomposition theorems in SAR interferometry applications

Theorems performing decompositions of the scattering matrix have been introduced to recognize observed targets and distinguish some of their characteristics, most of all the geometrical ones. For this reason, their main applications in the field of remote sensing have been suggested to be just the target identification and the classification of land coverages. A controversial aspect deals with their use with distributed targets that seem to be hardly representable by means of the scattering matrix and require for this scope higher order matrices. Some hints on these limits may be retrieved by considering interferometry, i. e., by studying interferometric quantities after performing the decompositions: those targets better resembling "ideal point scatterers" would present a degree of coherence considerably different from that of distributed targets. More generally, the correlation properties of various decomposition theorems will be presented and discussed, and the usefulness of the information so derived will be estimated.