中文摘要：

海底混响是一种与海底底质、海洋环境及发射信号参数等多种因素有关的随机过程,它的存在给水下目标探测识别带来很大困难。鉴于人耳听觉系统在听音辨物方面独特的优越性,利用Gammatone滤波器构建人耳听觉模型,提取听觉时-频特征和听觉谱特征,并讨论主动声呐掩埋目标回波和海底混响在这些特征下的可分离性,选取分离性好的特征构建特征空间,并利用径向基核函数支持向量机进行分类识别。实验结果表明,两类信号的特征具有良好的聚类性,能够获得较高的识别准确率,表明该方法能够有效地区分目标回波和混响。

英文摘要：

Bottom reverberation is a random process,related to seabed sediment,marine environment and transmitted signal parameters. The presence of reverberation causes great difficulties in underwater target detection and recognition. Based on the unique advantage of human listening ability on sound object discrimination,a Gammatone filter was used to construct an auditory model,the time-frequency features and auditory spectrum features were extracted,and the separation of active sonar embedded target echoes and bottom reverberation examined. The features providing good separation performance were selected to construct feature space,and the support vector machine,based on a radial basis kernel function,was used for classification and recognition. The experimental results show that the features of these two kinds of signals have good clustering performance,and can achieve high recognition accuracy. It is demonstrated that this method can effectively distinguish between target echoes and reverberation.