摘要
将希尔伯特-黄变换(HHT)用于船舶声信号特征提取中,利用HHT对实录船舶辐射噪声进行特征提取后,利用神经网络进行分类.研究表明希尔伯特-黄变换方法对于信号的时频特性具有较高的分辨能力,适用于水声非平稳信号的分析.与传统时频分析方法相比具有很强的自适应特性和较好的时频聚集性,时频分辨力高于小波变换.结果表明对于船舶声信号识别,希尔伯特-黄变换方法是一种有效的特征提取方法.
The Hilbert-Huang Transform(HHT) was applied to extract features of ship noises.HHT has higher resolving power on time-frequency characteristic than methods based on traditional time-frequency analysis.HHT method is suitable for underwater acoustic non-stationary signal analysis,and it has good adaptive and time-frequency concentration characteristics.The results show that HHT is an effectively method to extract characteristics of ship noises.
出处
《哈尔滨理工大学学报》
CAS
2014年第3期69-73,共5页
Journal of Harbin University of Science and Technology
关键词
目标识别
特征提取
希尔伯特-黄变换
神经网络
target recognition
feature extraction
Hilbert-Huang transform
neural network