摘要
提出了一种应用复值小波变换进行湖底回波特征提取的方法:采用线性相位的复db小波,对复解析信号进行多尺度的复值小波变换,然后提取合适尺度上的幅度信息作为目标识别的特征矢量.结合实测数据的分析表明,利用复值小波变换提取的幅度特征是一种有效、稳健的特征,能获得较高的正确识别率.复值小波变换也可以采用Mallat快速算法,因此这种方法得到特征矢量维数少,使用时实时性能好,便于工程实现.
A feature extraction method of underwater echo using complex valued wavelet transform （WT） is proposed： the complex I./Q orthogonal signal is processed with linear phase complex filter banks, and then the magnitude information in the proper scale is taken as the feature vector. Since both I./Q channels simultaneously are processed, the magnitude feature vector extracted from the complex wavelet coefficients can acquire more information about the targets. It is also shown by the feature extraction and recognition experiment of underwater echoes that the magnitude feature with CWT is a kind of robust, effective feature, and good recognition result can be obtained.
出处
《武汉理工大学学报（交通科学与工程版）》
2010年第2期318-322,共5页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
第42批中国博士后科学基金项目(批准号:20070420181)
首批中国博士后科学基金项目(批准号:200801487)资助
关键词
复值小波变换
线性相位
复值小波
湖底回波的特征提取
complex valued wavelet transform
linear phase
complex valued wavelet function
feature extraction of underwater echoes
