特征提取是目标识别与分类的重要前期工作，从信道畸变之中提取稳定特征对于目标识别来说尤为重要。本文首先基于人耳听觉的基本理论，提取了舰船辐射噪声的响度特征；接着基于射线理论，利用MATLAB仿真得到了信道的多途参数；最后将舰船辐射噪声通过多途信道作用，提取不同信道作用后的响度特征并进行对比分析。研究表明，舰船噪声在浅海多途信道中传播时，传播距离增大对其响度特征的影响较大也较复杂；接收深度的改变对响度特征的影响就相对小很多；声速梯度的改变也会对响度特征产生较大的影响。说明响度特征并不是一种稳定可靠的识别特征，这一点在声纳目标识别中值得注意。
Feature extraction is an important preliminary work for target recognition and classification. It is particularly important to extract stable features from channel distortion. Firstly, based on the basic theory of human hearing, this paper extracts the loudness feature of the ship’s radiation noise. Secondly based on the ray theory, the multi-path parameters of the channel are obtained by MATLAB simulation. Finally, the ship’s radiation noise is processed through the multi-channel, and the loudness feature of different channels is extracted and compared. The results show that when the ship noise propagates in shallow sea multi-path channel, its loudness characteristic is larger and more complex when the propagation distance is large. The influence of receiving depth on the loudness characteristic is much smaller, and the change of sound velocity gradient will also have a greater impact on the loudness characteristic. It shows that loudness feature is not a stable and reliable recognition feature, which deserves attention in sonar target recognition.
Stevens, S.S. (1957) Concerning the Form of the Loudness Function. The Journal of the Acoustical Society of America, 29, 603-606. https://doi.org/10.1121/1.1908979
[8]
Abarbanel, H.I. (1993) The Analysis of Observed Chaotic Data in Physical System. Review of Modern Physics, 65, 1133-1388. https://doi.org/10.1103/RevModPhys.65.1331
