基于Walsh谱统计分布特征的水下入侵目标分类

提出一种利用目标回波Walsh统计分布特征识别水下入侵目标方法,文中探讨水下入侵目标回波的Walsh谱统计分布特征,提取其统计分布特征量,均值、标准差、偏度系数、峰度系数和波动系数构建特征空间,降低了直接用Walsh谱识别的特征维数。针对采集对应三种入侵目标的回波信号提取其Walsh谱统计分布特征量并进行分类识别实验,并给出了BP神经网络的分类结果。

Underwater Intruder classification Based on Walsh Spectrum Distribution

A underwater intruder classification method based on Walsh spectrum distribution of sound echo is developed in this paper.The primary distribution parameters of the Walsh spectrum such as mean,standard deviation,skewness coefficient,kurtosis coefficient,and fluctuation coefficient are extracted as features vector to recognize underwater target,which decrease features dimension of Walsh spectrum for target classification.The recognition and classification test of three kind of intruder target echo are made b...