基于稀疏表示的船体检测方法研究

为进一步强化航道安全，解决海事CCTV人工值守、非自动化问题，提出了基于稀疏表示的船体检测方法。利用稀疏表示实现对船体的检测时，首先构建样本特征矩阵，然后利用K-SVD算法对样本特征矩阵进行学习，得到冗余字典，最后对测试样本进行重构，根据马氏距离判断测试样本属性。通过与传统方法的试验比较，实验结果表明，该算法实时性好、检测准确率高，可以很好地对CCTV视频监控的船体进行检测与跟踪，解决CCTV人工值守、非自动化问题，节省大量人力资源。     

A video semantic detection method based on locality-sensitive discriminant sparse representation and weighted KNN

Video semantic detection has been one research hotspot in the field of human-computer interaction. In video features-oriented sparse representation, the features from the same category video could not achieve similar coding results. To address this, the Locality-Sensitive Discriminant Sparse Representation (LSDSR) is developed, in order that the video samples belonging to the same video category are encoded as similar sparse codes which make them have better category discrimination. In the LSDSR, a discriminative loss function based on sparse coefficients is imposed on the locality-sensitive sparse representation, which makes the optimized dictionary for sparse representation be discriminative. The LSDSR for video features enhances the power of semantic discrimination to optimize the dictionary and build the better discriminant sparse model. More so, to further improve the accuracy of video semantic detection after sparse representation, a weighted K-Nearest Neighbor (KNN) classification method with the loss function that integrates reconstruction error and discrimination for the sparse representation is adopted to detect video semantic concepts. The proposed methods are evaluated on the related video databases in comparison with existing sparse representation methods. The experimental results show that the proposed methods significantly enhance the power of discrimination of video features, and consequently improve the accuracy of video semantic concept detection.     

基于稀疏表示的小目标检测

现代战场作战环境复杂,智能化、网络化的干扰机是雷达探测的主要威胁。对于从副瓣进来的干扰,利用阵面空域自由度可以较容易地抑制干扰。但对于从主瓣方向进来的干扰,传统的反干扰方法失效,不能有效抑制干扰。干扰环境下,目标信噪比(SNR)较低,如果降低门限检测会增加很多虚警点迹。针对主瓣噪声干扰场景下小目标检测问题,提出了基于稀疏表示的检测技术,利用了目标可以稀疏表示,而噪声不能被稀疏表示的特性,达到了抑制噪声,降低虚警的效果。仿真实验验证了该方法的有效性。     

Near-field time-frequency localization method using sparse representation

This paper presents a novel near-field source localization method based on the time-frequency sparse model.Firstly,the method converts the time domain data of array output into time-frequency domain by time-frequency transform;then constructs sparse localization model by utilizing the specially selected time-frequency points,and finally the greedy algorithms are chosen to solve the sparse problem to localize the source.When the coherent sources exist,we propose an additional iterative selection procedure to improve the estimation performance.The proposed method is suitable for uncorrelated and coherent sources,moreover,the improved estimation accuracy and the robustness to low signal to noise ratio(SNR) are achieved.Simulations results verify the efficiency of the proposed algorithm.     

基于LSDANet的手机芯片屏蔽壳表面缺陷检测方法

为了解决手机芯片屏蔽壳表面白印缺陷微小、尺度各异等因素影响检测快速性和准确性的问题,本文提出一种基于长短连接通路和双注意力网络(long short link and double attention network, LSDANet)的手机芯片屏蔽壳表面缺陷检测方法。首先,通过构建基于编码和解码的语义分割模型和利用长短距离连接通路,提高网络模型对尺度各异缺陷的特征提取能力。其次,分别设计基于通道和空间的注意力机制,增大5—10 pixel尺寸的白印缺陷在空间和通道上的特征权重。最后,融合双注意力机制和长短距离连接通路分割模型,构建LSDANet缺陷检测网络,应用于手机芯片屏蔽壳表面缺陷检测。实验数据表明,LSDANet网络能够达到96.21%的平均像素精度、66.13%的平均交并比和39.03的每秒检测帧数,相比多种语义分割算法均具有更高的检测精度和速度。     

一种稀疏信道非零抽头检测的新方法

为了提高稀疏信道估计时的精确度,需对其非零抽头位置进行检测。传统的匹配追踪系列算法在检测时须以信道的稀疏度为先验信息,或给出严格的迭代终止条件。针对这一情况,引入了一种新的方法——分层检测算法实现信道非零抽头位置检测。实验仿真表明,分层检测算法在SNR大于10 dB时检测信道非零抽头位置的性能是与匹配追踪算法相当的,而当信噪比较低时,其性能更优。在算法实现过程中同时分析了影响其性能的原因,并在此基础上对其进行了改进。     

基于图像稀疏表示的红外小目标检测算法

基于超完备字典的图像稀疏表示是一种新的图像表示理论,利用超完备字典的冗余性可以有效地捕捉图像的各种结构特征,从而实现图像的有效表示.针对红外小目标检测问题,提出了一种基于图像稀疏表示的检测方法,该方法采用二维高斯模型生成样本图像,继而构造超完备目标字典,然后依次提取测试图像的图像子块并计算其在超完备字典中的表示系数,背...     

基于局部线性嵌入稀疏表示的人群行为检测

针对人群行为检测中行为特征维数高,数据量大以及特征局部流形结构不稳定的问题,本文提出了一种新的人群行为检测方法。该方法首先利用改进变分光流算法初步提取人群运动特征,对视频序列中各帧图像分块获取低维行为特征,然后利用局部线性嵌入稀疏表示对其进行分类。基于局部线性嵌入(Locally linear embedding,LLE)的正则项能够有效的保留测试样本的局部流形结构,提高样本的判别能力。实验结果表明本文的算法能够有效地提高人群行为检测效率。     

利用改进的稀疏表示处理FBG传感信号

基于 压缩感知(CS)的正交匹配追 踪 ( OMP ) 算法,须以稀疏度 确定 为先验条件, 在 实际 应用 中稀疏度 不 易 确定 的情况下, 本文 提出了 稀疏度确 定方法和 二次正交匹配追踪 (TOMP)算法。 先 引入熵权法 采用 多指标融合并结合饱和值点法确定稀疏度 , 然后利用所提 方 法 对实验信号进行重构 。 实验 仿真结果表明: 与同类算法相比,本文所提 TOMP 算法增加 0.1s 运行时间降低了 12～ 22% 的重构误差,更好折中处理了重构误差和时间;与不同类算法相比,本文 所提方法重构的信号信噪比(SNR)最大可提升 22 dB ,且均方根误差(RMSE)降低 0.7,因此去噪效果更优。     

基于稀疏表达的微弱信号提取及检测方法

首先简要介绍了稀疏表达模型以及其常用的求解方法;然后讨论了一种基于稀疏表达的微弱信号提取及检测方法;在此基础上,为了减少时间与空间复杂度,开发了一种快速求解算法;最后给出信号的提取结果以及定性与定量分析,并用大量的实验结果证明讨论的基于稀疏表达的微弱信号提取及检测方法的顽健性与自适应能力。     

四阶累积量稀疏表示的DOA估计方法

针对现有稀疏重构DOA估计算法不能抑制噪声项以及在高斯色噪声背景下不再适用的问题,本文提出了基于四阶累积量稀疏重构的DOA估计方法。首先,利用接收数据的四阶累积量构建了稀疏表示模型,该模型抑制了噪声项;其次对四阶累计量矩阵进行奇异值分解,化简了稀疏表示模型,通过奇异值分解,不仅减小了数据规模,而且进一步抑制了噪声。对于稀疏表示模型的求解,先利用信号子空间与噪声子空间的正交特性选取权值矢量,然后利用加权l1范数法对模型求解实现DOA估计。理论分析和仿真实验表明本文算法在高斯白噪声和色噪声背景下均适用;能够处理非相干和相干信号,且在低信噪比条件下,对相干信号有更高的估计精度;较之同类的稀疏重构算法,本文算法具有较低的算法复杂度和更高的角度分辨力。      

Optimized signal expansions for sparse representation

Traditional signal decompositions such as transforms, filterbanks, and wavelets generate signal expansions using the analysis-synthesis setting: the expansion coefficients are found by taking the inner product of the signal with the corresponding analysis vector. In this paper, we try to free ourselves from the analysis-synthesis paradigm by concentrating on the synthesis or reconstruction part of the signal expansion. Ignoring the analysis issue completely, we construct sets of synthesis vectors, which are denoted waveform dictionaries, for efficient signal representation. Within this framework, we present an algorithm for designing waveform dictionaries that allow sparse representations: the objective is to approximate a training signal using a small number of dictionary vectors. Our algorithm optimizes the dictionary vectors with respect to the average nonlinear approximation error, i.e., the error resulting when keeping a fixed number n of expansion coefficients but not necessarily the first n coefficients. Using signals from a Gaussian, autoregressive process with correlation factor 0.95, it is demonstrated that for established signal expansions like the Karhunen-Loeve transform, the lapped orthogonal transform, and the biorthogonal 7/9 wavelet, it is possible to improve the approximation capabilities by up to 30% by fine tuning of the expansion vectors     

轿车流线型曲面缺陷的图像识别与检测方法研究

提出一种基于灰度像素邻域模板匹配的轿车流线型曲面缺陷的图像识别与检测方法,采用图像处理方法进行轿车流线型曲面缺陷识别,提高轿车表面缺陷的智能修复能力。对采集的轿车表面图像进行降噪提纯处理,利用自适应权重均衡分割方法进行图像分割,采用Retinex信息增强方法进行曲面缺陷图像增强处理,结合灰度像素邻域模板匹配方法进行缺陷识别与检测。仿真结果表明,采用该方法进行轿车流线型曲面缺陷的图像识别,检测的准确匹配概率较高,对缺陷部位的定位较准,性能优越。     

Range-spread target detection using the time-frequency feature based on sparse representation

The wideband radar transmitting the linear frequency modulation signal often processes its echoes by the stretched processing. This paper deals with the range-spread target detection in white complex Gaussian noise. Here, we propose a new detection method for the range-spread target based on sparse representation, which selects the time-frequency feature to realise the target detection. It can be simply described as follows: first, the sketched signal is reconstructed from its noisy measurements by basis pursuit de-noising (BPDN); scatterers on the target are determined by its reconstruction and used to calculate the Wigner distribution; for the target embedded in noise, the time-frequency feature in its power-density spectrum is compared with the decision threshold. Meanwhile, the median absolute deviation (MAD) is adopted to estimate the noise variance. The mainly novelties can be concluded as follows: the Fourier matrix is selected to sparsely represent the sketched signal; the sparsity is used to improve the SNR of the received echoes; the Wigner transform is utilised to acquire the time-frequency feature of the range-spread target. Both the optimisation theory and time-frequency representation are introduced to solve the target detection problem. Experimental results on the raw data show that the proposed detector outperforms the conventional methods.     

基于稀疏表示与特征融合的人脸识别方法

针对人脸识别在有遮挡、表情、光照的变化或受到噪声污染时鲁棒性变差问题,提出一种基于稀疏表示与特征融合的人脸识别算法。首先采用低秩恢复算法得到训练样本和测试样本的干净人脸图像,提取干净人脸图像的LBP,HOG,Gabor三种特征向量;然后对部分训练样本进行SRC分类测试,根据SRC的识别结果与分类残差定义一个损失函数,再利用正则化最小二乘法计算出使损失函数最小的权重向量;最后根据该权重向量重构规则化残差进行分类。在ORL,Extended Yale B和AR数据库上进行实验,结果表明,该算法优于利用单一特征识别的方法,并且对光照、噪声、遮挡等因素产生的影响有较好的泛化性能。     

一种快速超分辨铜带表面缺陷图像复原方法

在铜带表面缺陷检测系统中,针对仅硬件改善缺 陷图像精细特征信息,受制造水平、成本等因素制 约以及传统超分辨复原方法实时性不强等问题,提出一种基于粗糙集(RS) 与纹理特征预分类的快速超分辨率(SR)图像复原方法。本文方法利用RS属性约简原理,选择并优化对弱纹理缺陷目标 描述性较好统计特 征参数,并在匹配搜索时根据纹理特征对样本库进行预搜索分类,然后在分类得到的纹理内 容相近的样本 子集中对输入的低分辨率(LR)样本块精确匹配搜索。理论和实验结果表明:本文方法应用 于铜带缺陷在线 检测系统中,可使缺陷区域的高频信息增强、边缘和细节更加清晰,且算法实时性较好,在 兼顾图像复原质量和运行效率上具有优越性和可行性;并可用于其它金属表面的图像复原。     

一种改进的稀疏表示的手写数字识别

提出一种改进的稀疏表示的手写数字识别方法。首先将样本字符块训练成过完备字典,然后通过改进的基于L1/2正则化算法进行系数分解,最后通过求重构图像与原始图像的残差进行分类。实验采用标准的数字数据库MNIST进行手写数字识别,改进的稀疏表示方法较其他方法能够有较高识别率,达到98%以上,并且具有很好的噪声鲁棒性。     

基于改进YOLOv5s轻量化带钢表面缺陷检测方法

针对YOLOv5s模型参数量大、难以在嵌入式设备上部署的问题,设计了一种轻量化的YOLOv5s带钢表面缺陷检测方法。首先将主干网络中的部分卷积层替换为多分枝结构的RepGhost,增强了主干对特征信息的提取能力,推理时可以转化为单分支结构,保证了检测速度。其次提出了一种轻量级的FPN网络(GG-FPN),其中的G-Ghost用于削减C3模块中的冗余参数,而GSConv则利用大卷积核的深度可分离卷积和分支结构,保证精度和速度的双提升。实验表明,在NEU-DET数据集上,GG-FPN模型参数量较原FPN减少了24.7%,GFLOPs降低了20.6%。对 于整个模型,改进的算法mAP仅损失1.9%,参数量较YOLOv5s减少了37.5%,GFLOPs降低了33.1%,检测速度达到187 frame/s,更好地均衡了检测的速度与精度。     

基于小波分解表面波的涂层平板表面缺陷检测方法研究

激光照射在涂层平板表面产生的超声表面波传播时会产生色散现象,该现象会影响表面缺陷检测精度.针对表面波的色散问题,采用有限元仿真分析表面波在涂层平板表面的传播规律,利用Morse小波分解多频表面波信号,提取最大幅值单一频率的缺陷时域信号从而确定表面缺陷位置.在此基础上通过扫描缺陷附近检测点的方式,分析缺陷前沿入射波和缺陷...     

Image deblocking via sparse representation

Image compression based on block-based Discrete Cosine Transform (BDCT) inevitably produces annoying blocking artifacts because each block is transformed and quantized independently. This paper proposes a new deblocking method for BDCT compressed images based on sparse representation. To remove blocking artifacts, we obtain a general dictionary from a set of training images using the K-singular value decomposition (K-SVD) algorithm, which can effectively describe the content of an image. Then, an error threshold for orthogonal matching pursuit (OMP) is automatically estimated to use the dictionary for image deblocking by the compression factor of compressed image. Consequently, blocking artifacts are significantly reduced by the obtained dictionary and the estimated error threshold. Experimental results indicate that the proposed method is very effective in dealing with the image deblocking problem from compressed images.