B-Spline Wavelet-Based Target Feature Extraction in Wideband Radar

A new approach to wideband radar target feature extraction is proposed. It uses one dimensional B-spline wavelet transform on the target impulse to reduce the amount of data which must be stored for each aspect angle. According to the multiresolution theory, the 3rd B-spline wavelet functions are constructed. Then, the coefficients of approximation signal extracted with the 3rd B-spline wavelet transform are acted as the target feature vectors. The performance of this novel approach is examined through experiments. The results show that it has great capabilities in extracting the target feature and reducing the amount of data for further processing greatly while strongly restrains the effect of additive Gauss white noises.