This thesis takes the application of microwave sensor in military as researchbackground. At the beginning, the detecting theory of microwave sensor and themathematical construction approach for output signal are studied, the feature extractionmethod which is adapt to microwave sensor and two target sorters which are based on twokinds of neural network(FMM network and BP network) are developed.Then corresponding to the concrete engineering application, based on DSP, ahigh-speed signal processing system, which is developed with the purpose of the real-timeof signal processing is presented in this paper, and at last the double scanning recognitionsketch based on the one dimensional wave in time domain is determined for the concretesystem. Around the general developed process of DSP system, the hardware and so(?)twarealgorithm in DSP designing both side are carried on research. The system integrates signalacquisition circuit, DSP smallest system circuit and serial communication circuit and soon. DSP is in charge of the processing of signal and controlling other circuits. As for thesoftware, its major function is to recognize the detecting target, thus the extracted signalwill be firstly preproeessed, then the primary feature value of the signal will be computedout, and at last the template matching algorithm will be used to identify and recognize thetarget. There are several dominate advantageous in this system, such as real-time processing,high integration, small valume and low energy consume, ect. The thesis finally carry out theexperiments of MMW sensor's target recognition, and get a good result in the end. |