Using image processing to recognize radar wind profiler atmospheric echoes

Arguably the most difficult and error prone step in the wind profiler digital signal processing, signal identification is the process of identifying the atmospheric signal in the range doppler data matrix that is composed by processing radar return echoes. This data matrix might contain multiple types of signals (Atmosphere, clutter, RFI) as well as noise. The objective is to reject the interfering signals and identify the atmospheric return. Here we describe a prototype processing technique based on Matlab vine. Our implementation, named Tamarack, is currently applied on a beam by beam basis to heat map images of the stacked Doppler spectra model to reject interference and identify the atmospheric signal.