Study On The Echo Recognition Method Of Bottomed Shell Column With Ball Sphere Head

The bottomed target is very difficult to be recognized or detected, just because its echo signal has a small Signal-Noise Ratio(SNR) and a complicated structure, which are caused by the reverberation and add-echoes produced by the bottom. To prove the SNR, STFT time-frequency filter and Wigner-Ville across component filter are introduced in this thesis, both of them are based on the characteristics of Liner Frequency Modulated(LFM) signal and reverberation. From this point, a characteristics extraction method is given, and makes a recognition system for the echo signal of bottomed column shell target with a half sphere head with the Fuzzy Fusion Classifier(FFC) together.On the time-frequency domain, the distribution of LFM signal is a single straight bias, and the distribution of reverberation is a band along with the time axis, the STFT time-frequency filter is based on this characteristics. A LFM signal added with its frequency-shifted copy, will produced a across component in the result of its quadratic time-frequency distribution, which has a bigger SNR, the Wigner-Ville across component filter is based on this. A recognition system is realized with FFC and these characteristics extraction techniques.To test the methods in this thesis, an experience has been made between the November to the December of 2006.