摘要
研究了响度、尖锐度、粗糙度、波动强度等主要的心理声学参数及其计算方法,利用主要的心理声学参数作为目标特征参数用于水下目标分类识别,并以正确识别率为准则对这些特征参数进行了修改。使用K-均值聚类方法对3类舰船噪声实测数据进行了目标分类识别仿真实验。实验结果表明,该方法提取的特征能够较好地反映信号本质,取得了较好的分类识别效果,特别是以修改后的心理声学参数为特征具有更高的识别率。
The calculation methods of the primary psychoacoustic parameters, such as loud- ness, sharpness, fluctuation, strength, and roughness are investigated. The primary psychoacoustic parameters are used as feature parameters to recognize underwater targets. The feature parameters are modified according to the recognition rate. The recognition and classification tests of three kinds of shipnoise data are made by K-means method. Simulation results show that the signal essence is reflected by extracted features and the classification result is obtained by the extracted features, especially for the modified psychoacoustic parameters.
出处
《数据采集与处理》
CSCD
北大核心
2006年第3期313-317,共5页
Journal of Data Acquisition and Processing
关键词
特征提取
目标识别
心理声学参数
K-均值算法
feature extraction
target recognition
psychoacoustic parameter
K-means method
