舰船辐射噪声功率谱特征提取方法研究

基于分段最小二乘曲线拟合方法,提出了一种的提取舰船噪声信号功率谱中线谱特征的新方法,通过一种新的坐标变换,用最小二乘曲线拟合方法精确提取了连续谱特征,将连续谱特征和线谱特征组合得到了舰船噪声的功率谱特征。利用BP神经网络对海上实录的两类目标噪声进行了分类识别,实验证明了方法的有效性,识别率达到90%以上,分类效果很好。

A New Power Spectrum Feature Extractive Method of Ship Radiated Noise

A new method of line spectrum feature extraction in ship radiated noise power spectrum was proposed based on partial least-square curve fitting.A new coordinate transform and the least-square curve fitting were used to extract the continuous spectrum feature.The continuous spectrum feature and the line spectrum feature are combined as the power spectrum feature of ship radiated noise.The classification recognition experiment for two different classes of target noises in the sea was performed with BP neural network,and the results show that the method is effective with a recognition rate over 90%.