信号能量聚点与特征提取方法研究

Power Gathering-Point and Feature Extraction Method of Target Signal

为了在特征提取中充分利用信号的真实特征信息 ,定义了刻画信号能量分布的λ水平能量聚点的概念 ,并进行了简要分析。根据目标信号功率分布的特点 ,首先对目标信号进行小波阈值消噪以突出其能量分布的特征信息 ,然后对消噪后的信号作功率谱估计并计算功率谱信号的λ水平能量聚点 ,以能量聚点各λ水平的频率区间作为特征量 ,对目标信号进行特征提取。通过对水压场信号的仿真计算 ,采用模糊聚类的方法对提取的信号特征量作聚类分析 ,对特征量的有效性进行了检验。计算结果表明 ,在较低信噪比条件下 ,信号 λ水平能量聚点特征量模型能够反映出真实的目标特征 ;采用所给出的特征提取方法 ,达到了目标信号检测准确率高、虚警率低的效果。

In order to adequately use the real characteristic information of signal in the process of characteristic extraction, the concept of power gathering point on λ level is defined and analyzed concisely. And it can perform the quantitative analysis for the power distribution of the target signal. The signal first removes the noise by wavelet transform,and the power spectrum density estimation is made to form a new series in frequency domain,from which the power gathering point on λ level of signal is calculated and used as a characteristic quantity of the signal, so it is possible to extract the feature of the target signal. By using the fuzzy classifying method,the effectiveness of the method is verified with a real target signal added by simulated noise signal under the condition of low R SN .