Research on Technique of Passive Sonar Target Recognition

Feature extraction of targets radiated-noise is the key technique of passive sonar target recognition system.In this paper,a feature extraction method of modulation continuou spectrum is studied,and then the classification experiment for three different classes of targets is done by exploiting neural network target classifier based on genetic-backpropagation algorithm,results of experiment show that the passive sonar target recognition system designed in the paper has higher correct classification rate.