摘要
基于信号处理的声音特征提取和模式识别技术近些年发展较快 ,声信号具有精确性和稳定性等特点 ,借助计算机技术运用信号处理的方法对物体发出的声音进行分析 ,提取其振动特征量 ,可以为物体运转情况的预测提供精准的数据分析依据。声音模式识别在无人职守、故障检测、灾害预防等方面具有很广阔的应用前景。该文主要介绍了在声音模式识别的一般过程 ,其中包括声信号预处理、特征提取、模式分类过程的方法。基于信号处理的声音模式识别 ,主要是利用信号处理的方法手段 ,对信号进行分析 ,提取频域、时域、幅域特征 ,对这些特征进行统计分类 ,应用数学方法 ,设计合理的分类器 ,达到分类识别的目的。
The technologies of feature extraction and pattern classification based on signal processing are developing rapidly in recent years. The acoustic signal has higher accuracy and stability,so people can use signal processing method to analyse objects'sound by computer and extract its features.These features can offer reliable prediction for object's running.Voice pattern recognition has broad prospects of application in unmanned guard,defect detecting, disaster-preventing and so on.The paper introduces the process of voice pattern recognition,including the method for pre-processing,extraction of feature and classification of pattern. Based on the processing of signal, voice pattern recognition depends on the methods of signal processing. We extract signals' spectrum,dimensionality and time features, and classify them with mathematical method. Then we try to design a reasonable interpolator. This paper mainly researches the methods for signal processing.
出处
《计算机仿真》
CSCD
2004年第9期134-137,共4页
Computer Simulation
