受污染混沌信号的协同滤波降噪

发布时间：2018-04-10 18:36

本文选题：混沌信号 + 协同滤波　；参考：《物理学报》2017年21期

【摘要】：根据混沌吸引子的自相似分形特性,提出了一种利用协同滤波重构受污染混沌信号的降噪算法.所设计的降噪算法通过对相似片段的分组将一维混沌信号的降噪转化为一个二维联合滤波问题;然后,在二维变换域用阈值法衰减噪声;最后,通过反变换获得原始信号的估计.由于分组中的相似片段具有良好的相关性,与直接在一维变换域做阈值降噪相比,分组的二维变换能获得原信号更稀疏的表示,更好地抑制噪声.仿真结果表明,该算法对原始混沌信号的重构精度和信噪比的提升都优于小波阈值、局部曲线拟合等现有的混沌信号降噪方法,对相图的还原质量也更好.
[Abstract]:According to the self-similar fractal characteristics of chaotic attractors, a noise reduction algorithm using cooperative filtering to reconstruct contaminated chaotic signals is proposed.The proposed denoising algorithm converts the noise reduction of one-dimensional chaotic signals into a two-dimensional joint filtering problem by grouping similar segments. Then, the threshold method is used to attenuate the noise in the two-dimensional transform domain.The estimation of the original signal is obtained by inverse transformation.Because of the good correlation between the similar segments in the packet, compared with the threshold denoising in the one-dimensional transform domain, the two-dimensional transform of the packet can obtain the sparse representation of the original signal and suppress the noise better.Simulation results show that the proposed algorithm is superior to wavelet threshold in reconstruction accuracy and signal-to-noise ratio of the original chaotic signal. The existing noise reduction methods such as local curve fitting are also better for the restoration of phase diagram.
【作者单位】：华南理工大学电子与信息学院;广东技术师范学院电子与信息学院;
【基金】：国家自然科学基金(批准号:61372008) 广东省科技计划项目(批准号:20158010101006,2014A010103014)资助的课题~~
【分类号】：O415.5;TN911.4

【相似文献】