Recognition Technology, And Space Targets Based On Rcs

Moving statues adjudgement, character extraction and target classify methods of space targets based on RCS sequences are discussed in this paper. Firstly, the concepts of RCS, typical moving statues of space targets and the theory of wavelet are introduced. Also the advantage of wavelet in non-stationary signal analyzing is emphasized. Secondly ,a realization of power spectrum estimation of signal on the basis of wavelet transform is given. This method is applied to the RCS time series, so as to adjudge moving statues of different space targets.Three character extraction methods are proposed in the next part respectively applying discrete wavelet energy feature, fractal-box-dimension feature and statistical feature. Distance measurement is introduced to evaluate the features. The computation results of distance measure show that the method based on discrete wavelet energy feature is the best. Finally, average distance classifier and BP neural network classifier are used to classify targets with features that have been extracted, and the result shows that the BP neural network classifier has higher identification rate than the average distance classifier.All experiments in this paper are based on the real data from some Monopulse Instrumentation Radar with high precision radar. The results of computer simulation by Matlab demonstrate that the methods are reliable and practical.