Before you start

O.k. well now you got your hands on one of this fully polarimetric SAR data sets which will maximize your stress level and minimize you free disk space. For the free disk space I can't help you but in order to constrain the stress level you should do follow some easy rules before you start. Some of the advice sound incredible stupid, but believe me after some years in the field there only very few errors I haven't made ;o). First of all think about what you want. Make sure you understood the theory. Than think about the information in the data. Is the information you want really in the data. This seems logical but is a point that is quite often overseen. Always remember that a fully polarimetric data set only yields 5 degrees of freedom. You still know why? If not go back and read the basics. The [S]-matrix contains 4 Complex numbers, the off-diagonal terms are equal due to the reciprocity theorem in the backscatter case. That leaves with 3 amplitudes and 3 phases, since the absolute phase is ambiguous you have to spent one as a reference. That gives you 3 amplitudes and 2 phases. If your model (or whatever you need the data for) needs more input you have to make some assumptions or use additional data. Think about the resolution you need. If you want to make a classification for a map in a scale of 1:25000 a SIR-C/X-SAR data set is not what you are looking for. If you want to make a crop classification a data set from December might be not the best choice and so on. If you are not aware of this fact or neglect it you will most likely end up with a garbage in - garbage out result. O.k. If you are sure (and I mean sure as in sure-sure), that polarimetry can be the answer to your question or at least yield some possibilities to improve your approach you can start to work with the data.