论文部分内容阅读
有效地从含有噪声的非平稳信号中提取特征是进行非平稳信号分类等研究的基础。应用流域算法,对含有高斯白噪声的非平稳信号的时频分布图进行分割,并根据能量占优的准则对其合并,提出了一种基于能量的特征提取方法。仿真结果表明该方法能有效地提取特征量,且对高斯白噪声具有很好的抗噪性能。
Effectively extracting features from non-stationary signals with noises is the basis for the research of non-stationary signal classification and so on. By using the algorithm of watershed, the time-frequency distribution of non-stationary signals containing Gaussian white noise is segmented, and then the energy-based feature extraction method is proposed according to the energy dominant criterion. The simulation results show that this method can effectively extract the features and has good anti-noise performance for Gaussian white noise.