基于Gammatone频率离散小波系数的水下目标鲁棒识别  

Robust Underwater Target Recognition Based on Gammatone Frequency Discrete Wavelet Coefficients( GFDWC)

作　　者：吴姚振[1] 杨益新[1] 田丰[1] 杨龙[1] 陶灿[1] 

机构地区：[1]西北工业大学航海学院,陕西西安710072

出　　处：《西北工业大学学报》2014年第6期906-911,共6页Journal of Northwestern Polytechnical University

基　　金：国家自然基金(11274253)资助

摘　　要：针对水下目标辐射噪声的复杂性和研究样本的局限性,提出了一种基于Gammatone频率离散小波系数的特征提取方法,结合人耳听觉感知机理,提取出了有效吻合人耳听觉特性的识别特征。该方法在2个方面改进了目前广泛采用的美尔倒谱系数:1用Gammatone滤波器代替三角滤波器,更好地模拟了人耳基底膜的滤波特性;2用离散小波变换替换离散余弦变换,使得识别特征具有优良的局部化特性。针对实录的水下动物叫声和舰船辐射噪声进行分类实验,表明所提出的特征提取方法在识别率和稳健性方面都有明显提高。In this paper, aiming at the complexity of underwater target-radiated noise and the limitations of study samples, we present a novel recognition feature approach based on GFDWC, which imitates the hearing perception mechanism of humans.Compared with Mel-frequency cepstral coefficients （ MFCC ） used extensively at present, this method improves the performance of underwater target recognition system in the following two aspects：（ 1） the triangular filters were improved based on gammatone function, which approximates well the filter response of basilar membrane;（2） taking localization characteristics into account, we replaced the discrete cosine transform （DCT） with discrete wavelet transform （ DWT） .The results of the recognition experiments of real underwater target-radiated noise and their analysis show preliminarily that the proposed method has good robustness and gives high recognition rates.

关 键 词：Gammatone滤波器 Gammatone频率离散小波系数 水下目标识别 特征提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TB566[自动化与计算机技术—计算机科学与技术]