改善船舰噪声听觉辨识的特征注入法  

Feature introducing to improve listening identification of ship-radiated noise

作　　者：李训诰[1] 周利辉[1] 丁文强[1] 

机构地区：[1]海军潜艇学院,青岛266072

出　　处：《应用声学》2014年第5期397-404,共8页Journal of Applied Acoustics

摘　　要：为提高声纳员对舰船噪声的听测识别水平,本文提出了特征注入法及其实现方法。该方法通过对指定子带重构、子带加权重构和特征子带注入法三个核心环节的建模仿真,对船舰噪声进行重构,进而对重构后的舰船噪声进行听测识别和谱分析,并根据分析结果对重构方案进一步优化调整。结果表明,采用合理重构方案的特征注入法能有效改善人耳的听测辨识效果。In order to improve listening identification of ship-radiated noise,feature introducing and its implementation method were presented. Three parts were modeled and simulated, which included reconstruction with the selected subbands,weighting reconstruction and reconstruction with the introduction of characteristic ones. The ship-radiated noise was reconstructed and the processed noise was analyzed through listening identification and spectrum analysis. Furthermore,the reconstructing parameters were optimized according to the analyzing results. It＇s shown that feature introducing with reasonable reconstruction scheme can effectively improve the performance of listening identification.

关 键 词：听测 特征注入法 重构噪声 

分 类 号：TB566[交通运输工程—水声工程]