摘要
对被动声纳信号进行了分析和特征提取，基于听觉系统识别声音信号的原理，提出了一种新颖的平稳恒Ｑ 特征，并给出了距离指数和识别指数２ 种更为合理的评价方法。实验结果表明本文提出的平稳恒Ｑ特征较原有几种典型特征，具有良好的分类正确性．
An analysis of passive sonar signal and extraction of their features is described.By researching the principle of hearing system in identification of sound signal, a new steady constant Q feature is proposed. Two more reasonable evaluation methods of distance and identification index are given. Results of the experiment show that the steady constant Q feature method has higher correct classification rate.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
1999年第6期16-20,共5页
Journal of Southeast University：Natural Science Edition
关键词
被动声纳信号
分类
特征提取
谱估计
小波变换
passive sonar signal classification
feature extraction
spectrum estimation
wavelet transform
speech identification
artificial neural network
