The Radio Plasma Imager (RPI) aboard the IMAGE spacecraft probes plasma at both far and near ranges by means of radio sounding. The RPI plasmagrams, similar in their concept to the ground-based and topside ionograms, contain not only a variety of signatures pertaining to the remote plasma structures and boundaries, but also a suite of the local plasma resonances stimulated by the RPI radio transmissions. Detection and interpretation of the resonance signatures is a valuable diagnostic tool providing the actual electron density and magnetic field strength at the spacecraft location, which are needed for the accurate processing of the remote sensing information on the plasmagrams. The high volume of the RPI sounding data demanded the development of automated techniques for routine interpretation of the plasmagrams. This paper discusses a new method for the detection and interpretation of the resonance signatures in the RPI plasmagrams that employs pattern recognition techniques to localize the signatures and identifies them in relation to model-based resonances.