基于奇异熵的信号降噪技术研究

On Improving the Effectiveness of the New Noise Reduction Technique Based on Singularity Spectrum

基于信号奇异谱分析技术 ,提出了奇异熵的概念 ,分析了信号信噪比与奇异熵间的内在联系 ,并提出了一种根据奇异熵增量渐近特性来对信号奇异谱降噪阶次进行确定的有效方法。仿真结果表明 ,该方法可以使信号奇异谱降噪阶次得以合理确定 ,在极大程度上确保了降噪后的信号所含信息的完整性和获得最大信噪比改善 。

A new noise suppression technique based on singularity spectrum was proposed by Lu Z.M., but it did not explain how to determine the order of singularity spectrum which was important for effectiveness of noise suppression. The concept of singularity entropy was proposed for determining the order of singularity spectrum, which reflected the totality of information in a noise-containing signal. The variation of asymptotical tendency of increment of singularity spectrum was studied, and the method for determining the order was developed. Simulation results for noise-containing signal and Doppler signal show that the singularity spectrum method with singularity entropy can effectively suppress noise and maintain completeness of the signal.