Signal Detection And Recognition Of Fishery Detector

The work of this dissertation is one part of the project-fishery detector.It finished the initial work of signal collection,signal processing,estimation of DOA and recognition.In this paper, we collect signals from the receiver and stored them in computer.we use the primary principle to process the signals and estimate DOA,At last we have a recognition of the target.The work of this paper can be divided into five parts: signal collecting,signal detection,classification recognition,terminal display and experimental analysis.During the stage of data acquiring,we use C to programme the I/O procedure after matching the outer and inner circuit. We acquired the right order of data and realized the frame synchronization and bit synchronization.In the stage of the signal detection,we make use of the basic processing principle to remove the noise and estimate DOAIn recognition,we discuss two ways of feature extraction: one way is to extract the feature of LOFAR. Spectrum-analyze the T-F feature and reduce the dims by extracting the main member.The other way is power feature extraction on the basis of wavelet transformation. At last ,the features are passed into the neural networks(PNN)to recognize.The terminal display is to meet the demands of users.We display all functions in one interface.We develop an interface on VC++,which is convinient to users.Finally, we test the property of the whole system by two experiments.