An improved sparsity estimation variable step-size matching pursuit algorithm

To improve the reconstruction performance of the greedy algorithm for sparse signals, an improved greedy algorithm, called sparsity estimation variable step-size matching pursuit, is proposed. Compared with state-of-the-art greedy algorithms, the proposed algorithm incorporates the restricted isometry property and variable step-size, which is utilized for sparsity estimation and reduces the reconstruction time, respectively. Based on the sparsity estimation, the initial value including sparsity level and support set is computed at the beginning of the reconstruction, which provides preliminary sparsity information for signal reconstruction. Then, the residual and correlation are calculated according to the initial value and the support set is refined at the next iteration associated with variable step-size and backtracking. Finally, the correct support set is obtained when the halting condition is reached and the original signal is reconstructed accurately. The simulation results demonstrate that the proposed algorithm improves the recovery performance and considerably outperforms the existing algorithm in terms of the running time in sparse signal reconstruction.

Ultrasonic Nondestructive Signals Processing Based on Matching Pursuit with Gabor Dictionary

The success of ultrasonic nondestructive testing technology depends not only on the generation and measurement of the desired waveform, but also on the signal processing of the measured waves. The traditional time-domain methods have been partly successful in identifying small cracks, but not so successful in estimating crack size, especially in strong backscattering noise. Sparse signal representation can provide sparse information that represents the signal time-frequency signature, which can also be used in processing ultrasonic nondestructive signals. A novel ultrasonic nondestructive signal processing algorithm based on signal sparse representation is proposed. In order to suppress noise, matching pursuit algorithm with Gabor dictionary is selected as the signal decomposition method. Precise echoes information, such as crack location and size, can be estimated by quantitative analysis with Gabor atom. To verify the performance, the proposed algorithm is applied to computer simulation signal and experimental ultrasonic signals which represent multiple backscattered echoes from a thin metal plate with artificial holes. The results show that this algorithm not only has an excellent performance even when dealing with signals in the presence of strong noise, but also is successful in estimating crack location and size. Moreover, the algorithm can be applied to data compression of ultrasonic nondestructive signal.

Infrared small target detection using sparse representation

Sparse representation has recently been proved to be a powerful tool in image processing and object recognition.This paper proposes a novel small target detection algorithm based on this technique.By modelling a small target as a linear combination of certain target samples and then solving a sparse 0-minimization problem,the proposed apporach successfully improves and optimizes the small target representation with innovation.Furthermore,the sparsity concentration index（SCI） is creatively employed to evaluate the coefficients of each block representation and simpfy target identification.In the detection frame,target samples are firstly generated to constitute an over-complete dictionary matrix using Gaussian intensity model（GIM）,and then sparse model solvers are applied to finding sparse representation for each sub-image block.Finally,SCI lexicographical evalution of the entire image incorparates with a simple threshold locate target position.The effectiveness and robustness of the proposed algorithm are demonstrated by the exprimental results.

Power-line interference suppression of MT data based on frequency domain sparse decomposition

Power-line interference is one of the most common noises in magnetotelluric(MT)data.It usually causes distortion at the fundamental frequency and its odd harmonics,and may also affect other frequency bands.Although trap circuits are designed to suppress such noise in most of the modern acquisition devices,strong interferences are still found in MT data,and the power-line interference will fluctuate with the changing of load current.The fixed trap circuits often fail to deal with it.This paper proposes an alternative scheme for power-line interference removal based on frequency-domain sparse decomposition.Firstly,the fast Fourier transform of the acquired MT signal is performed.Subsequently,a redundant dictionary is designed to match with the power-line interference which is insensitive to the useful signal.Power-line interference is separated by using the dictionary and a signal reconstruction algorithm of compressive sensing called improved orthogonal matching pursuit(IOMP).Finally,the frequency domain data are switched back to the time domain by the inverse fast Fourier transform.Simulation experiments and real data examples from Lu-Zong ore district illustrate that this scheme can effectively suppress the power-line interference and significantly improve data quality.Compared with time domain sparse decomposition,this scheme takes less time consumption and acquires better results.

Application of Atomic Sparse Decomposition to Feature Extraction of the Fault Signal in Small Current Grounding System

Applying the atomic sparse decomposition in the distribution network with harmonics and small current grounding to decompose the transient zero sequence current that appears after the single phase to ground fault occurred. Based on dictionary of Gabor atoms and matching pursuit algorithm, the method extracts the atomic components iteratively from the feature signals and translated them to damped sinusoidal components. Then we can obtain the parametrical and analytical representation of atomic components. The termination condition of decomposing iteration is determined by the threshold of the initial residual energy with the purpose of extract the features more effectively. Accordingly, the proposed method can extract the starting and ending moment of disturbances precisely as well as their magnitudes, frequencies and other features. The numerical examples demonstrate its effectiveness.

基于双动态头Sparse R-CNN的表面缺陷检测算法

为了减少缺陷检测中的冗余检测,提出基于双动态头Sparse R-CNN的缺陷检测算法,2个动态头的责任不同:第1个负责不同尺度和空间的特征提取,第2个负责匹配可学习的提议特征。为了更好地提取图像细节信息,改进特征金字塔(FPN)为特征金字塔网格(FPG),并且与第1个动态头相结合进行特征提取。其次,提出了交流注意力来改进检测阶段的多头自注意力模块,减少随着迭代注意力图相似导致建模能力下降的问题。最后,改进边框回归损失函数GIoU为Alpha-CIoU,加速收敛并提升检测的精度。实验结果表明:算法在晶圆和热轧钢2种表面缺陷数据集上都取得很好效果,平均精度分别为94.3%和88.1%。

Synthetic aperture radar imaging based on attributed scatter model using sparse recovery techniques

The sparse recovery algorithms formulate synthetic aperture radar （SAR） imaging problem in terms of sparse representation （SR） of a small number of strong scatters＇ positions among a much large number of potential scatters＇ positions, and provide an effective approach to improve the SAR image resolution. Based on the attributed scatter center model, several experiments were performed with different practical considerations to evaluate the performance of five representative SR techniques, namely, sparse Bayesian learning （SBL）, fast Bayesian matching pursuit （FBMP）, smoothed 10 norm method （SL0）, sparse reconstruction by separable approximation （SpaRSA）, fast iterative shrinkage-thresholding algorithm （FISTA）, and the parameter settings in five SR algorithms were discussed. In different situations, the performances of these algorithms were also discussed. Through the comparison of MSE and failure rate in each algorithm simulation, FBMP and SpaRSA are found suitable for dealing with problems in the SAR imaging based on attributed scattering center model. Although the SBL is time-consuming, it always get better performance when related to failure rate and high SNR.

Application of signal sparse decomposition in dynamic test

In dynamic test,sampling rate is high and noise is strong,so a signal sparse decomposition method based on Gabor dictionary is put forward.This method iteratively decomposes the signal with the matching pursuit(MP)algorithm and takes the coherence ratio of the threshold as a condition of iteration termination.Standard MP algorithm is time-consuming,thus an adaptive genetic algorithm is introduced to MP method,which makes computation speed accelerate effectively.Experimental results indicate that this method not only can effectively remove high-frequency noise but also can compress the signal greatly.

Pulse Signal Recovery Method Based on Sparse Representation

Pulse signal recovery is to extract useful amplitude and time information from the pulse signal contaminated by noise. It is a great challenge to precisely recover the pulse signal in loud background noise. The conventional approaches,which are mostly based on the distribution of the pulse energy spectrum,do not well determine the locations and shapes of the pulses. In this paper,we propose a time domain method to reconstruct pulse signals. In the proposed approach,a sparse representation model is established to deal with the issue of the pulse signal recovery under noise conditions. The corresponding problem based on the sparse optimization model is solved by a matching pursuit algorithm. Simulations and experiments validate the effectiveness of the proposed approach on pulse signal recovery.

Radar Imaging of Sidelobe Suppression Based on Sparse Regularization

Synthetic aperture radar based on the matched filter theory has the ability of obtaining two-di- mensional image of the scattering areas. Nevertheless, the resolution and sidelobe level of SAR imaging is limited by the antenna length and bandwidth of transmitted signal. However, for sparse signals (direct or indirect), sparse imaging methods can break through limitations of the conventional SAR methods. In this paper, we introduce the basic theory of sparse representation and reconstruction, and then analyze several common sparse imaging algorithms: the greed algorithm, the convex optimization algorithm. We apply some of these algorithms into SAR imaging using RadBasedata. The results show the presented method based on sparse construction theory outperforms the conventional SAR method based on MF theory.

Coherence-based performance analysis of the generalized orthogonal matching pursuit algorithm

The performance guarantees of generalized orthogonal matching pursuit（ gOMP） are considered in the framework of mutual coherence. The gOMP algorithmis an extension of the well-known OMP greed algorithmfor compressed sensing. It identifies multiple N indices per iteration to reconstruct sparse signals.The gOMP with N≥2 can perfectly reconstruct any K-sparse signals frommeasurement y = Φx if K 〈1/N（1/μ-1） ＋1,where μ is coherence parameter of measurement matrix Φ. Furthermore,the performance of the gOMP in the case of y = Φx ＋ e with bounded noise ‖e‖2≤ε is analyzed and the sufficient condition ensuring identification of correct indices of sparse signals via the gOMP is derived,i. e.,K 〈1/N（1/μ-1）＋1-（2ε/Nμxmin） ,where x min denotes the minimummagnitude of the nonzero elements of x. Similarly,the sufficient condition in the case of G aussian noise is also given.

THE EXACT RECOVERY OF SPARSE SIGNALS VIA ORTHOGONAL MATCHING PURSUIT

This paper aims to investigate sufficient conditions for the recovery of sparse signals via the orthogonal matching pursuit （OMP） algorithm. In the noiseless case, we present a novel sufficient condition for the exact recovery of all k-sparse signals by the OMP algorithm, and demonstrate that this condition is sharp. In the noisy case, a sufficient condition for recovering the support of k-sparse signal is also presented. Generally, the computation for the restricted isometry constant （RIC） in these sufficient conditions is typically difficult, therefore we provide a new condition which is not only computable but also sufficient for the exact recovery of all k-sparse signals.

RECONSTRUCTION OF SPARSE POLYNOMIALS VIA QUASI-ORTHOGONAL MATCHING PURSUIT METHOD

In this paper,we propose a Quasi-Orthogonal Matching Pursuit(QOMP)algorithm for constructing a sparse approximation of functions in terms of expansion by orthonormal polynomials.For the two kinds of sampled data,data with noises and without noises,we apply the mutual coherence of measurement matrix to establish the convergence of the QOMP algorithm which can reconstruct s-sparse Legendre polynomials,Chebyshev polynomials and trigonometric polynomials in s step iterations.The results are also extended to general bounded orthogonal system including tensor product of these three univariate orthogonal polynomials.Finally,numerical experiments will be presented to verify the effectiveness of the QOMP method.

The Recovery Guarantee for Orthogonal Matching Pursuit Method to Reconstruct Sparse Polynomials

Orthogonal matching pursuit(OMP for short)algorithm is a popular method of sparse signal recovery in compressed sensing.This paper applies OMP to the sparse polynomial reconstruction problem.Distinguishing from classical research methods using mutual coherence or restricted isometry property of the measurement matrix,the recovery guarantee and the success probability of OMP are obtained directly by the greedy selection ratio and the probability theory.The results show that the failure probability of OMP given in this paper is exponential small with respect to the number of sampling points.In addition,the recovery guarantee of OMP obtained through classical methods is lager than that of ℓ_(1)-minimization whatever the sparsity of sparse polynomials is,while the recovery guarantee given in this paper is roughly the same as that of ℓ_(1)-minimization when the sparsity is less than 93.Finally,the numerical experiments verify the availability of the theoretical results.

一种新型的电能质量扰动信号分析的CDMSPSO-MP算法

针对匹配追踪(matching pursuit,MP)算法在检测电能质量扰动信号时存在的计算量大、重构信号质量不佳的问题,利用混沌动态多种群粒子群优化(chaos dynamic multi-swarm particle swarm optimization,CDMSPSO)算法对MP算法进行优化,提出了CDMSPSO-MP算法。首先,CDMSPSO算法使用Logistic映射替代伪随机数更新种群,提高信号重构时搜索时频原子的随机性;然后,将种群划分为多个小规模种群并设置相应的重组期,增加信号重构时频原子的多样性;最后,以扰动信号与原子内积的绝对值作为CDMSPSO算法的适应度函数,替代MP算法的遍历计算,提升信号的重构速度。实验结果表明,CDMSPSO-MP算法有效提高了计算速度,减少了无关时频原子作为扰动信号分量的计算,提高了重构信号的质量。

基于黑客画像的网络攻击者识别方法

为能够准确、快速识别网络攻击者,提出一种基于黑客画像的网络攻击者识别方法。构建将稀疏自编码器和贝叶斯神经网络相结合的SAE-BNN模型,检测不同攻击类型的恶意流量;针对不同的恶意流量,通过提取黑客属性特征、流量特征、时间特征和相似性特征,与事先建立的黑客画像库中的黑客画像进行匹配。如果与某个黑客画像完全匹配,则由此确定该黑客的身份。当不能与黑客画像库中的任何黑客画像进行匹配时,将该黑客的特征作为标签,构建新的黑客画像,并更新画像库。实验结果表明,提出的异常流量识别方法在精度、召回率、F1值和准确率上均有提升。基于黑客画像的黑客识别算法与常规方法相比,极大提高了识别效率。

基于变分模态分解和稀疏表示的局部放电信号去噪算法

鉴于局部放电信号受各种噪声的干扰,文章提出一种基于变分模态分解和稀疏分解的局部放电信号去噪算法。以稀疏表示算法为核心,基于局部放电信号的特性构建其过完备字典,再采用匹配追踪算法在过完备字典中搜索出原信号的最佳匹配原子集合重构信号;为解决过完备字典维度过高而导致的搜索次数太多的问题,引进变分模态分解算法和峭度值筛选进行预处理和预重构;优化后的方法可以限制稀疏分解算法的搜索范围和字典参数,以减小计算复杂度。仿真验证以及对工程环境中实测信号的去噪结果表明:该方法具有更好的降噪效果,即使在极低信噪比的情况下,依旧能提取出有效的局部放电信号。

基于混合稀疏ICCP的联合抗差重力匹配定位方法

针对经典最近等值线迭代(ICCP)算法因重力异常测量误差导致匹配精度下降甚至失效的问题,提出联合抗差匹配算法以提高匹配精度及可靠性。首先,分析了匹配点集间的匹配残差在高斯噪声影响下呈非高斯分布,为抑制其影响,采用l_(p)范数代替l_(2)范数计算匹配残差,并利用匹配残差重调野值点以获得有效的匹配区域。在此基础上,提出混合稀疏ICCP算法,并利用其进行粗匹配,然后将粗匹配后的位置作为惯导系统(INS)指示位置,再使用经典ICCP算法进行精匹配,获得更高的定位精度。仿真结果表明,考虑重力异常测量误差的情况下,重力联合抗差匹配算法的误差最大值小于1 n mile,导航精度较传统ICCP算法提升60%以上,提升了算法的鲁棒性和匹配精度。

基于距离哈希的稀疏点集快速匹配算法研究

针对不同坐标系下部分重叠的稀疏坐标点集,提出一种基于距离哈希的同名点快速稳健匹配算法。将各点与其邻近点的距离关系映射成一个二进制码身份标签,通过身份标签相似度计算,找出两个点集中满足设定阈值的候选匹配点对,从而建立初始匹配关系。据此计算刚体变换矩阵对两组点集进行配准,确定两组点集之间的精确匹配关系。实验结果表明:该算法不仅速度快、准确率高、对于噪点和低重叠度具有稳健性,而且对两个点集之间的初始相对位置没有任何限制。

基于参数化正交匹配追踪算法的高分辨雷达距离像重构

雷达线性调频信号调频率误差会导致传统匹配滤波模型失配,显著降低一维距离像成像效果.针对调频率误差模型下的宽带雷达高分辨一维距离像稀疏成像问题,提出了一种基于参数化正交匹配追踪算法的高分辨雷达一维成像方法.该方法首先确定调频率参数取值范围,构建调频率参数可选集合;然后基于回波数据观测矩阵与调频率的解析关系,针对每一个可选值构建参数化的稀疏观测矩阵,并进行单步正交匹配追踪迭代;最后基于每一个参数可选值对应的单步残差能量逐步筛选出调频率参数.仿真和实测数据实验验证了所提方法的有效性.