| Summary: | 碩士 === 中正理工學院 === 電機工程研究所 === 86 === Underwater acoustic signals are non-linear, time-varying, and with low signal-to-noise ratio. These properties make the signal analysis difficulty and complex. For resolving targets through the underwater acoustic signals, effective methods are proposed in this thesis to process underwater acoustic signals, Base on these methods, an signal acoustic recognition system is also designed.
Traditionally, the Fourier transform (FT) and Morlet wavelet transform (MWT) are the main tool for stationary and transient signals spectrum analysis, respectively. Here in, a modify power spectrum density (PSD) function is used to extract the critical features for stationary underwater acoustic signals, A multi-scaling MWT kernel is also proposed which can depict the underwater transient spectrum successfully.
To illustrate the effectiveness of these two novel design methods, some experiments are taken to perform by using simulation and recorded real underwater acoustic signals. Experimented results show that the proposed methods can detect and analyze both stationary and transient underwater acoustic signals successfully. An underwater acoustic signals analysis is also implemented on Matlab base personal computer to detect, analyze, and recognize targets by stationary signal features. It is hoped that an automatic underwater targets recognition system can be realized by methods discussed in this thesis in the future.
|
|---|
