A coding algorithm for the covariance matrix representation of polarimetric radar data

A coding algorithm for the covariance matrix representation of polarimetric radar data is presented. The goal is to optimize storage requirements and the throughput of computer applications for radar polarimetry, which rely on this kind of representation. The algorithm uses a 10-b quantizer, with gain balance and a nonlinear compressor expander, and achieves a compression factor of 2.5. Some theoretical considerations on the behavior of the quantizer with respect to the input data distribution are discussed. The algorithm performance is characterized using a data set acquired by the NASA JPL AIRSAR sensor.<
>