An Efficient Approach for Ground Echoes Suppression Based on Textural Features and SVM

Abstract

The use of the Support Vector Machine (SVM) technique for the clutter identification in the context of meteorological data is presented. The clutter is due to ground echoes and anomalous propagation. The SVM is combined with textural approach which is based on the Grey Level Co-occurrence Matrix (GLCM) that is the most used in the textural analysis image. An incoherent radar site is considered for this study. The results reveal that over than 91.1% of ground echoes are identified and 90.3% of precipitations are preserved. In addition 95.99% of anomalous propagations are removed. The use of our approach is lasts than 1mn for the treatment of each image. We can then filter the radar image in real time