摘要
Mel倒谱系数分析依据人耳听觉特性,可以提取有利于船舶目标分类的特征。前向神经网络的反向传播算法对类别数目小但分类困难的模式识别问题有良好的分类效果。针对Mel倒谱系数分析提取的船舶目标分类识别特征,采用前向神经网络的反向传播算法,可以有效对船舶目标进行分类。
Mel-cepstrum analysis based on characteristics of human hearing is able to extract feature for ship target classification.Back-propagation algorithms of forward neural network is suitable for pattern recognition problem when the number of classes is small but classification is difficult.Aimed at ship targets features are extracted by Mel-cepstrum analysis,using back-propagation algorithms of forward neural network,ship targets are classificated.
出处
《传感器与微系统》
CSCD
北大核心
2011年第6期55-57,67,共4页
Transducer and Microsystem Technologies
基金
国防预研基金资助项目(4010501050102)
