脉冲噪声检测与滤除器

从信号的自相关角度出发,分析了信号受脉冲噪声污染的特点.根据信号受脉冲噪声污染的特点,提出了一种从幅度上检测脉冲噪声、从宽度上识别脉冲噪声的方法,然后对检测出的噪声点进行预测插值去噪处理.由于只对噪声污染点处理,未改变非污染信号点,从而极大地保护了信号的细节信息.该滤波器算法简单、能够自动适应信号变化和脉冲噪声特性.实验结果表明,与中值滤波器、均值滤波器相比,该滤波器能够在保护信号细节信息的同时,很好地去除脉冲噪声.

Detection and Attenuation Filter for Impulse Noise

Character of signal destroyed by impulsive noise is analyzed according to character of signal correlation. The methods of impulsive noise detection from amplitude and impulsive noise identification according to its width are proposed. After detection, the detected noises are deleted by predictive or interpolation filters. The detail information of the signal is greatly protected by processing the detected noises and keeping the other signals the same with their previousness.The proposed filter takes on the character of adaptiveness to signal variance and noise characteristics. Experiment results show that the proposed filter has advantages on the noise attenuation while preserving details of the signal comparing with median filter and mean filter.