基于融合能量代价函数的水声信号识别算法

Recognition arithmetic of underwater acoustic signals based on fusing energy cost function

提出了融合能量代价函数的概念及基于代价函数的小波包能量法,并将其应用于水声信号的识别。新算法以融合能量代价函数为标准,在整个小波库中构造最优小波包基,从小波包基上提取信号最有价值的特征值。由于从分类最佳的角度选择特征,所以与固定尺度小波包能量法相比,算法对分类特征模糊的信号有较好的识别效果。

The concept of fusing energy cost function and method based on the function are proposed, the method was used for the recognition of underwater acoustic signal. The new arithmetic constructed the best wavelet packet based on the wavelet library with the criterion of fusing energy cost function and contract the most valuable features of the signals. Compared with the ways of the fixed-scale energy method, the arithmetic has better recognition effect to the signals that have fuzzy features.