A clutter removal strategy for weather radars, based on neural network approaches and using polarisation diversity as feature space

Meteorological radar is an important tool capable of providing high resolution measurements of precipitation. In this article, we suggest the use of polarisation diversity, using the knowledge of scattering properties of the targets, as a method of obtaining a good discrimination between clutter and precipitation. Although, this requires a more sophisticated radar than the conventional incoherent one, it provides more information on targets and their identification/quantification as well. Basically, most (linear) polarisation diversity systems work on switching polarisation at each transmission of the microwave pulses and receiving a backscattering signal in both polarisation channels. Also, some systems include the phase information, making Doppler computations also possible.