基于全变分模型稀疏信号降噪算法的研究  

Method for Locating Open Source Software Security Defects Based on Risk Trajectory

作　　者：王娜 吕东澔 张勇[1] WANG Na;LV Dong-hao;ZHANG Yong(Inner Mongolia University of Science and Technology college of Information Engineering,Baotou Inner Mongolia 014010,China)

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《计算机仿真》2023年第7期325-330,336,共7页Computer Simulation

基　　金：内蒙古自治区自然科学基金(2019BS06004);内蒙古自治区自然科学基金(2020LH06006);国家自然科学基金(61763038)。

摘　　要：针对稀疏信号的精确降噪问题,提出了一种基于非凸惩罚函数的全变分降噪算法。利用前向-后向分裂方法对由该惩罚项构造的代价函数进行迭代以获得该模型的最优解,从而得到准确的稀疏信号幅值。为了证明非凸惩罚函数在全变分降噪算法中对稀疏信号降噪的优势,对仿真信号及肌电实测信号分别进行降噪并分析,实验结果均证明降噪算法的可行性,其中仿真信号降噪结果与基于全变分模型的两种经典算法对比显示,所提算法比优化最小化降噪算法的均方差降低23.3%、平均绝对偏差降低22.4%、信噪比提高16.5%,比迭代裁剪降噪算法的均方差降低36.3%、平均绝对偏差降低34.9%、信噪比提高10.7%;证明所提算法能够削弱噪声对肌电信号的影响,验证了上述算法在实际应用中的有效性。Aiming at the problem of precise denoising of sparse singnals,a non-convex penalty total variation de-noising algorithm(NCP-TV)was proposed in this paper.The forward and backward splitting iterative method was used to obtain the optimal solution of the cost function constructed from the non-convex penalty function,and the ac-curate sparse signal amplitude was obtained.In order to prove the advantage of the non-convex penalty total variation algorithm in denoising sparse signals,the simulation signals and myoelectric measured signals were analyzed in this paper.The experimental results all prove the feasibility of the denoising algorithm,and the simulation signal denoising results were compared with the two classical algorithms based on total variation model.The comparison be-tween the NCP-TV and the optimization minimization denoising algorithm shows that the mean square error is reduced by 23.3%,the mean absolute deviation is reduced by 22.4%,and the signal-to-noise ratio is increased by 16.5%.Compared with iterative clipping noise reduction algorithm,the mean square error is reduced 36.3%,the mean abso-lute deviation is reduced by 34.9%,and the signal-to-noise ratio is increased by 10.7%.The results prove that the proposed algorithm can reduce the influence of noise on myoelectric signals,and verify the practical application value of the proposed algorithm.

关 键 词：全变分 稀疏信号 降噪 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]