A novel feature extraction method for radar target classification using fusion of early-time and late-time regions

摘要

This paper proposes a feature vector fusion of early-time and late-time regions, which improves the performance of radar target classification. For verifying the performance of the proposed method, we use the calculated radar cross section (RCS) of four full-scale targets and measured the RCS of three scale model targets. Then, we extract a feature vector from a waveform structure in the early-time region. The resonance frequencies are extracted using an evolutionary programming (EP)-based CLEAN algorithm in the late-time region. The extracted feature vectors are passed through the feature fusion process and then used as inputs for a neural network classifier. The results show that the proposed method exhibits better performance than those that use either early-time or late-time features.