A CODING ALGORITHM FOR THE COVARIANCE-MATRIX REPRESENTATION OF POLARIMETRIC RADAR DATA

In this paper, we present a coding algorithm for the covariance matrix representation of polarimetric radar data. The goal is to optimize st orage requirements and the throughput of computer applications for rad ar polarimetry, that rely on this kind of representation. The algorith m uses a 10-bit quantizer, with gain balance and a nonlinear compresso r expander and achieves a compression factor of 2.5. We first describe the algorithm, and explain the basic design choices; next we make som e theoretical considerations on the behavior of the quantizer with res pect to the input data distribution; finally, we characterize experime ntally the algorithm performance, using a data set acquired hy the NAS A JPL AIRSAR sensor.