.94 Ghz Radiometer Signal Detection Techniques

Based on specific research projects as a backdrop, wavelet analysis methodsof pretreatment and of feature extraction methods against radiometer actualoutput signal were researched. The following are the main points.1. On the exchange-power radiator of full power Principle, the technicalindicators of 3mm radiator for Metal targets on the ground were determined,signal analysis and noise analysis against radiometer actual output signalwere done.2. Radiometer actual output signal were pretreated, the effects ofde-noising of FFT and wavelet were compared. The Analysis of the thresholdmethod with soft threshold results was done. A new threshold for waveletthreshold methods was taken, the effects of different parameters in the newthreshold value method were compared.3. Feature Extraction Methods against the target signal were researched.To the wavelet decomposition scale space, the energy distribution of featureextraction, the amplitude distribution of feature extraction and thetime-frequency characteristics of feature extraction were done. To a waveletpacket Feature Extraction Method, the entropy of feature extraction to Waveletpacket best space was done.