Development of noise reduction algorithm for underwater signals

The underwater acoustic signal is affected by ocean interference and ambient noise disturbance during its propagation in ocean. Therefore, the signals reveal random process and time varying characteristics. In order to distinguish the weakened signal caused by the loss of long distance propagation, the received signal must be processed properly. During the research, a sophisticated procedure is developed for de-noising the signal automatically, the procedure consists of three parts: firstly, wavelet transformation of the underwater acoustic signals. Secondly, threshold of wavelet coefficients. Thirdly, inverse wavelet transformation of reconstructing the modified signals. Central to this research is the problem of de-noising which makes use of the genetic algorithms to find the optimal threshold value for the shrinking of wavelet coefficients. Two different noisy signals are used to examine the goodness of the varied designs for doing de-noise of signals. The first type is artificial signals. The second type is the actual underwater acoustic signal. In order to evaluate the effect of the proposed method, the criteria of mean-square error and signal to noise ratio are used to evaluate the result of the de-noising, for comparison purpose two existing de-noising methods are used in demonstration. The outcome of demonstration shows that the proposed method can achieve good performance on underwater signal de-noising.