I work a lot with synthetic aperture radar (SAR) data. SAR imagery requires some demanding image processing, not least to ameliorate image speckle and correct for topographic distortions.
Polarimetric SAR (polSAR) utilizes four transmit-receive configurations (HH,HV,VH and VV) allowing us to synthesize all possible polarisation configurations (including linear waves at 45 degrees and circular polarisation). Polarisation synthesis and other polSAR operations are not trivial and are computationally intensive. Helpfully, a number of companies offer PolSAR software (PCI, SARscape). So too does ESA (PolSARPro).
An alternative is Radartools (RAT). This program was developed at the Technical University of Berlin and is maintained by Andreas Reigber and Maxim Neumann (contributions were made by a number of others). The former works in DLRs microwaves and radar institute, the latter at JPLs Radar group. These guys know what they are doing.
Radartools is freeware, running on an IDL Virtual Machine platform. The software includes tools for processing amplitude SAR, PolSAR, InSAR and PolInSAR data. Data from all the major spaceborne platforms can be ingested. Output to ENVI, Generic Binary and PolSARPro formats are possible.
RAT offers possibly the most appealing visualisation of SAR data I have seen. I don’t know what they have done, or how they have done it, but opening a SAR image in RAT results in a visually pleasing display of image data with suppressed speckle. The zoom functions are a weakness with RAT, and the GUI lacks ‘bells and whistles’ (including contrast stretching). But RAT competes with the best in many ways.
For me though it is the PolSAR functionality that makes RAT great (the amplitude SAR functions are entirely satisfactory). Decompositions include the well known Pauli basis, Freeman-Durden and Sphere-Diplane-Helix (Krogager) decompositions and the Moriyama (for urban area analysis) and TVSM (an Eigenvector type approach). Alpha Entropy and Anisotropy can be calculated and Wishart classifications can be performed. Filter options include the relatively simple Boxcar and more advanced Simulated Annealing. Watch out for memory issues with large datasets: I’ve encountered a few problems. The workflow in RAT is not always clear so some trial and error is needed. However, RAT offers an excellent range of PolSAR tools.
RAT doesn’t offer terrain correction or radiometric correction of terrain induced effects. I export RAT products to SARscape for correction. NEXT may be another option for freeware users. Nevertheless, RAT is a great tool with many great features: and it is entirely free. The mail list is a great place to solicit help and Messrs Neumann and Reigber kindly monitor it and respond to cries for help. So next time you need some SAR software swing by Berlios and try out RAT.
NB. I did not mention the RAT InSAR functionality because I have not used it.