Application of HRRP even rank central moments features in satellite target recognition

Because high resolution range profile (HRRP) is sensitive to pose and translation, HRRP even rank central moments are proposed to be used as a kind of stable target feature in satellite target recognition. HRRP SNR is enhanced by wavelet denoising, then translation-invariant central moments are extracted. In order to reduce the feature vector dimension, even rank central moments are proposed to be used alone. Nearest neighbor fuzzy classifier (NNFC) that is very fit for combined features is introduced to process the central moments features vector. Experimental results with real satellite data show that good recognition performance is obtained.