Separation of Rain and Non-Rain events in Radar Imaging using Multiple Elevations and Texture Analysis

We present an algorithm for the characterization of rain and non-rain related echoes such as atmospheric anomalous propagation and ground clutter in weather radar imagery. The motivation behind the proposed work is the need for an accurate quality control algorithm which at the same time reduces human interaction. Similar to our previous work, the proposed algorithm uses textural information of radar reflectivity images. The proposed technique utilizes several radar elevations to build 3D volume blocks, in which each slice is constructed from several radar elevation scans. The proposed technique provides a better representation of the volume data set, suitable for 3D multiscale textural analysis.