 |
During the last few decades, most of the worldwide weather radar infrastructure has been upgraded to dual-polarization. Plans to demonstrate the utility of such technology for civil aircraft radars were undertaken by the EU Clean Sky Joint Technology Initiative (www.cleansky.eu). Dual polarization offers appealing advantages for avionic applications, such as the ability to correct X-band attenuation due precipitation and the possibility of implementing automated Hydrometeor Classification Algorithms (HCA). HCA can be useful to detect the presence of dangerous weather conditions related to the presence of hydrometeors like hail or graupel, typically associated with convection. Most of implemented HCAs rely on Fuzzy Logic (FL) methods. Supervised learning models based on a Support Vector Machine (SVM) are widely used for classifying remote sensing imageries and offer advantages in terms of computing time. The major problem of the SVM HCA is preforming the learning phase. This process has been implemented off-line using results obtained by a FL classifier. Effectiveness of the SVM HCA has been tested by using simulated scenarios relative to an intense convective event that occurred on 15 October 2012 in the Southern Mediterranean and with real data collected by C-band dual-polarization ground-based radars.
