摘要
基于船舶辐射噪声信号具有非线性、非平稳的特征 ,提出采用提取船舶辐射噪声信号的非线性混沌特征量和多尺度小波能量特征 ,并将两者综合作为特征参数输入神经网络分类器进行船舶分类识别。实验结果表明 ,该方法能较好地区分不同类型的船舶。
In this paper, a method is presented to extract the nonlinear features and multi-scale wavelet packet energy spectra of ship radiated noise which is nonlinear and nonstationary. The nonlinear features and energy spectra are integrated as the inputs to the neutral network to class the different ship radiated noise signals. The results of the experiments show that the method can class different ships effectively.
出处
《中国海洋大学学报(自然科学版)》
CAS
CSCD
北大核心
2004年第6期1036-1040,共5页
Periodical of Ocean University of China
关键词
船舶辐射噪声
混沌特征
小波包分析
信号分类
ship radiated noise
chaos features
wavelet analysis
signal classification