The Acoustic Target Recognition Method Of Research

Smart weapons will play a more and more important role in future warefare. A distinct feature that make smart weapones different from traditional ones is that they can decide what kind of objects the invading enemy weapons are, and choose on itself the most effective measures to attack and destroy them. Of all the objects discrimination methods, the passive acoustic identification is a hot topic thoroughout the world. Giving off no signals, this method uses the acoustic singnal the objects being attacked gives off to achive its discirmination and thus have a fairly good concealment ability.The core problem of this thesis is to study the different acoustic signal identification methods. Considering that the acoustic signal received are subject to environmental, climatic and other factors and thus assumes the characters of non-stable signals, this paper uses the wavelet analysis to reduce noise and to extract charactrzations. At the same the time, comparasions between traditional noise-reducing and character-extracting methods and wavelet tranform are made to get some valuable conclusions. At the final chapter, test results are given to show the comparative advantage of the methods this paper proposes.A singal data format transform program, a de-noising, a character extraction and a neutral network classification programs are written to accompolish the tasks of different stages during the whole process.