Multi-Lag Hybrid Correlation Coefficient Estimator

The copolar correlation coefficient is one of the main polarimetric variables. It is used for recognition of the types of radar echoes and in separation of returns from rain and snow. The latter requires precise measurements of the correlation coefficient in areas with low and moderate signal-to-noise ratios. Correlation coefficient estimates are unusable when they become larger than one, which is common when the number of samples per dwell is small and in areas with signal-to-noise ratios lower than 15 dB. This can be caused by the mismeasurement of noise powers in the horizontal and vertical channels which becomes apparent, as increased number of invalid correlation coefficient estimates, in regions where signal-to-noise ratios are small (e.g., on the fringes of weather phenomena). Other cause is the inherent positive bias in the correlation coefficient estimator which is more pronounced when the number of samples per dwell is small. In that regard, a novel correlation coefficient estimator is presented herein which has a potential of being less biased for small number of samples. Mitigating the correlation coefficient bias will improve polarimetric recognition of echoes.