Automatic defect inspection of patterned thin film transistor-liquid crystal display (TFT-LCD) panels using one-dimensional Fourier reconstruction and wavelet decomposition

Large-sized flat-panel displays have become increasingly important for use in computer monitors and televisions. This paper has considered the problem of automatic visual inspection of micro-defects including pinholes, scratches and particles in patterned thin film transistor-liquid crystal display (TFT-LCD) panel surfaces. For large-sized TFT-LCD panel inspection, high-resolution line scan is demanded. We propose a global one-dimensional (1-D) Fourier-based image reconstruction scheme that directly works on the 1-D line images instead of the traditional two-dimensional area images. The proposed method fully uses the inherent geometric structure of a TFT-LCD panel. It first eliminates the frequency components that represent the periodic pattern of a TFT-LCD line image in the 1-D Fourier spectrum and then back-transforms the 1-D Fourier-domain image to the 1-D spatial domain image using the inverse Fourier transform. The Fourier reconstruction process can effectively remove the patterned background and distinctly preserve local anomalies in the resulting 1-D image. Wavelet decomposition is further applied to remove uneven illumination in the filtered image so that defects can be easily segmented with simple statistical control limits. Experimental results on a number of micro-defects embedded in TFT-LCD panels show that the proposed method can reliably detect various ill-defined defects without designing and measuring the quantitative features of individual defect types.     

One-dimensional-based automatic defect inspection of multiple patterned TFT-LCD panels using Fourier image reconstruction

Thin film transistor-liquid crystal display (TFT-LCD) has been used in a wide range of electronic devices. For large-sized and high-density TFT-LCD panel inspection, a high-resolution line scan is demanded. A TFT-LCD panel image at a fine resolution presents very complicated patterns with less regularity. The paper proposes a non-referential defect detection scheme that directly works on the one-dimensional (1D) line images using the Fourier image reconstruction. The 1D grey-level line image is first divided into small segments, each of the length of the repeated period for a given TFT-LCD panel. The divided segments are then combined as a two-dimensional (2D) image. The frequency components corresponding to the 1D background pattern can be easily identified in the 2D Fourier spectrum. By eliminating the frequency components in the 2D Fourier spectrum that represent the periodic structural pattern of the combined 2D image and then back-transforming the image using the inverse Fourier transform, the 2D Fourier reconstruction process can effectively remove the complicated background pattern and well preserve local anomalies. Experimental results on a number of micro-defects embedded in different patterned regions of TFT-LCD panels show that the proposed method can reliably detect various ill-defined defects.     

OLED显示屏表面缺陷自动检测方法

目的 为了解决OLED显示屏表面周期性纹理背景和缺陷边界模糊、对比度低的特征导致其表面缺陷检测困难的问题,开展OLED显示屏表面缺陷自动检测方法研究.方法 对OLED显示屏图像进行奇异值分解,选择前2个较大的奇异值重构图像纹理背景,对原图像和重构图像进行差分运算,获得残差图像.将残差图像像素随机赋予初始隶属度值,采用模糊C均值聚类法获得像素最终隶属度值.根据隶属度大小,将残差图像像素聚成2类,并从残差图像中准确地分割缺陷.结果 选取较大的2个奇异值可以有效地重构OLED显示屏的周期性纹理背景;模糊C均值聚类法分割缺陷获得的区域灰度一致性(U)平均值为0.9846.结论 基于奇异值分解的背景重构方法可以有效地检测OLED显示屏表面缺陷;与分水岭法和Otsu方法相比,模糊C均值聚类可以准确地分割模糊边界的缺陷区域.     

基于四元数奇异值分解的数字印刷质量评价

目的针对人眼视觉感受与印刷质量评价结果存在不一致的缺陷,提出全四元数及奇异值分解的数字印刷质量评价方法。方法首先,依据图像颜色空间属性,采用四元数作为印刷图像表示模型,其中四元数矩阵实部为图像亮度方差,虚部为红、绿、蓝颜色,建立印刷图像全四元数表示模型。然后,对四元数图像模型进行四元数奇异值分解,得到图像四元数特征向量。最后,印刷图像质量评价指标以图像奇异值特征向量线性相关程度进行衡量。结果印刷图像四元数模型可完整地表示数字印刷图像,有效地凸显人眼视觉敏感结构信息,在印刷质量评价中实现了图像信息并行处理。结论印刷图像四元数模型提高和改善了彩色印刷图像质量准确性,与人眼视觉特性一致性较好。     

Contrast enhancement via texture region based histogram equalization

Abstract

This paper presents two novel contrast enhancement approaches using texture regions-based histogram equalization (HE). In HE-based contrast enhancement methods, the enhanced image often contains undesirable artefacts because an excessive number of pixels in the non-textured areas heavily bias the histogram. The novel idea presented in this paper is to suppress the impact of pixels in non-textured areas and to exploit texture features for the computation of histogram in the process of HE. The first algorithm named as Dominant Orientation-based Texture Histogram Equalization (DOTHE), constructs the histogram of the image using only those image patches having dominant orientation. DOTHE categories image patches into smooth, dominant or non-dominant orientation patches by using the image variance and singular value decomposition algorithm and utilizes only dominant orientation patches in the process of HE. The second method termed as Edge-based Texture Histogram Equalization, calculates significant edges in the image and constructs the histogram using the grey levels present in the neighbourhood of edges. The cumulative density function of the histogram formed from texture features is mapped on the entire dynamic range of the input image to produce the contrast-enhanced image. Subjective as well as objective performance assessment of proposed methods is conducted and compared with other existing HE methods. The performance assessment in terms of visual quality, contrast improvement index, entropy and measure of enhancement reveals that the proposed methods outperform the existing HE methods.     

Dominating direction based an efficient copy–move image tampering detection technique

Post-processing operations on a copy–move tampered image makes tampering detection a challenge. Most of the existing techniques are unable to address all the problems simultaneously. This paper presents a block-based technique, which addresses all these problems effectively. Here, at first, the approximate band coefficients are extracted using Contourlet Transform for further steps as it reduces the dimensionality as well as the effects of noise. Then Harris’ matrix is used to find the dominant direction of the block to counter the rotation. Finally, singular values of blocks are calculated using singular value decomposition, which works as a feature vector. The efficacy of the proposed algorithm has been verified through extensive simulations and comparisons with several other challenging copy–move image tampering detection techniques.     

基于VMD和SVD的柴油机气门间隙异常特征提取研究

为有效地从柴油机缸盖表面振动信号中提取气门间隙故障特征,提出一种基于变分模态分解(VMD)和奇异值分解(SVD)的特征提取新方法。采用VMD算法对缸盖振动信号进行分解,利用所得的模态分量构建特征矩阵;接着应用SVD理论将特征矩阵转变为表征频率特性的奇异值序列,探讨了稳定工况下的奇异值序列与不同气门间隙状态之间的关系;由于转速、负荷等工况的改变对信号特征层的影响与故障所引起的信号特征的改变可能非常相似,因此将奇异值序列作为特征参数,输入到随机森林分类器中,构建分类模型,对柴油机变工况下的气门间隙故障进行诊断。实验结果表明:该方法能有效识别气门间隙故障,突出故障敏感特征;与传统基于Hankel矩阵和小波包系数矩阵的SVD特征提取方法相比,该方法所提特征参数在柴油机变工况条件下具有更高的识别率。     

基于EMD的奇异值分解技术在滚动轴承故障诊断中的应用

针对滚动轴承故障振动信号的非平稳特征,提出了一种基于经验模态分解(EmpiricalModeDecomposition,简称EMD)和奇异值分解技术的滚动轴承故障诊断方法。该方法首先采用EMD方法将滚动轴承振动信号分解为多个平稳的内禀分量(IntrinsicModefunction,简称IMF)之和,并形成初始特征向量矩阵。然后对初始特征向量矩阵进行奇异值分解得到矩阵的奇异值,将其作为滚动轴承振动信号的故障特征向量,并输入神经网络来识别滚动轴承的工作状态和故障类型。实验分析结果表明,本文方法能有效地应用于滚动轴承故障诊断。     

Image texture classification using wavelet based curve fitting and probabilistic neural network

This article describes a new approach for image texture classification based on curve fitting of wavelet domain singular values and probabilistic neural networks. Image textures are wavelet packet transformed and singular value decomposition is then employed on subband coefficient matrices after introducing non‐linearity. Lower singular values are truncated based on energy distribution to effectively classify textures in the presence of noise. The selected singular values are fitted to the exponential curve. The model parameters are estimated using population‐sample analogues method and the parameters are used for performing classification. A modified form of probabilistic neural network (PNN) called weighted PNN (WPNN) is employed for performing the classification. Compared to probabilistic neural network, WPNN includes weighting factors between pattern layer and summation layer of the PNN. Performance of the approach is compared with model based and feature based methods in terms of signal to noise ratio and classification rate. Experimental results prove that the proposed approach gives better classification rate under noisy environment. © 2007 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 17, 266–275, 2007     

一种基于奇异值分解技术的模型定阶方法

摘要：模态参数识别中模型阶次的确定非常重要。基于奇异值分解技术,探讨模型阶次的确定方法,提出了利用奇异值的相对变化率来确定模型的阶次。该方法利用结构的量测脉冲响应信号构造Hankel矩阵,对其进行奇异值分解后计算奇异值的相对变化率,变化率最大的地方对应着模型的阶次。通过数值算例研究了噪声因素对模型阶次确定的影响,并利用模型实验数据验证了本方法的有效性。     

基于四元数矩阵奇异值分解的彩色图像分解

讨论了四元数矩阵的奇异值分解（QSVD）,四元数矩阵的奇异值仍是正实数,但两个酉矩阵是含有四元数的四元数矩阵。给出通过四元数矩阵的等价实矩阵求解QSVD的有效算法。最后应用QSVD进行彩色图像分解,并给出了在Fruits、Baboon等图像上的实验结果。QSVD使许多基于SVD的图像处理方法可以推广到彩色图像处理上而不用再将彩色图像分解成三个通道图像进行处理。     

一种新的差分奇异值比谱及其在轮对轴承故障诊断中的应用

因滚动体和保持架的随机滑动,轴承故障信号多为伪循环平稳信号。针对这种情况,提出了应用周期截断矩阵的奇异值分解的轮对轴承故障诊断方法。研究了轴承故障伪循环平稳信号的奇异值分布,结合奇异值能量差分和奇异值比,提出了一种新的能量差分奇异值比谱作为周期截断矩阵的嵌入维度计算方法;利用能量差分奇异值比谱计算嵌入维度并利用轮对轴承振动信号构造周期截断矩阵,对矩阵进行奇异值分解,并提出利用差分能量谱确定奇异值有效秩阶次并重构矩阵从而分离出周期信号;对该信号做包络分析以实现轮对轴承的故障诊断。应用轮对实验台的复合故障轴承振动数据对该方法进行验证,结果表明,所提方法能够有效提取轴承外圈、滚动体及保持架的特征频率的基频及其倍频,与传统应用Hankel矩阵进行奇异值分解降噪方法相比,该方法抗干扰能力显著,能够分离同频带的不同故障周期信号,且得到的包络谱谱线清晰,谐波丰富,使故障诊断的可靠性得到了显著提高。     

基于二次样条小波细节信号峰值的有效奇异值确定

研究了Hankel矩阵方式下交流分量、直流分量和噪声的奇异值分布特点,指出当矩阵阶数足够大时,代表这三个分量的奇异值在总奇异值向量中将被作为单独的坐标被分离开,而其衔接点处存在突变。提出利用二次样条小波对奇异值向量进行变换,并根据其细节信号的峰值位置来确定有效奇异值。研究结果表明,当细节信号的最大峰值位于第二个坐标时,表明原始信号中存在较大直流分量,此时根据第二最大峰值可确定全部有效奇异值,否则就根据最大峰值位置确定有效奇异值。信号处理实例证实了此方法的准确性。利用此法有效地分离出了轴承振动信号中的调制特征,进而据此准确地判断了轴承滚道的损伤情况。     

基于小波和奇异值分解的图像边缘检测

针对传统图像边缘检测抑制噪声能力弱的问题,给出了一种小波变换和局部梯度场内奇异值分解相结合的边缘检测方法.首先在图像预处理阶段,为了提取准确的边缘特征,文中利用小波变换的时频局部化特性,对图像进行小波变换.该文对用小波求取的梯度场使用局部梯度奇异值分解的方法;利用奇异值的特性和良好的稳定性,使提取的边缘特征更加突出并且能够达到抑制噪声的目的.实验证明该文方法既能在无噪声影响的图像中提取出清晰完整的单边缘,又能在有噪声干扰的情况下提取出理想的边缘.     

一种抗数字印刷/拍摄的数字水印算法研究及验证

目的数字印刷和手机拍摄是检验含有数字水印的印刷产品版权有效性的主要途径,解决在数字印刷和手机拍摄场景下水印信息难以提取这一问题。方法提出一种基于离散小波变换和奇异值分解的抗数字印刷/手机拍摄的数字水印算法。首先对宿主图像进行二级离散小波变换,对低频子带进行奇异值分解,将Arnold置乱后的水印嵌入到奇异值分解的对角矩阵中,最后进行小波逆变换得到嵌入水印图像。结果含水印信息的峰值信噪比(PSNR)达到了39.67,在不同攻击下的归一化相关系数(NC)均在0.89以上,实际的印刷拍摄测试中NC值也均达到0.94,证明该水印算法在保证不可见性的同时对于印刷产品实际应用中的常见攻击具有很好的抵抗能力。结论对不同品牌的智能手机、印刷设备以及不同种类的纸张采集到的含水印图像进行测试,均能够在印刷和拍摄场景下实现水印提取,证明了该算法的有效性。     

Fuzzy singular value shrinkage for Poisson image denoising

Poisson noise is a fundamental problem in various imaging applications, such as low-light photography, computed tomography and fluorescence microscopy. To remove Poisson noise, an adaptive iterative singular value shrinkage algorithm based on variance-stabilizing transformation (VST) and fuzzy logic classification is proposed in this paper. Since Poisson noise is signal dependent, we use the VST to convert it into signal independent Gaussian noise. The transformed image is divided into a number of blocks, and the similarity of these blocks is well judged according to the similarity criterion of an approximate Kullback–Leibler (KL) distance, and they are arranged to form a low-rank matrix. Then, the proposed algorithm uses singular value decomposition and adaptive soft-thresholding contraction operator to reduce noise, because large singular values point to the position of interesting information, and small singular values point to the position of the noise. In addition, to better preserve the structural information of the image, an adaptive iterative regularization technique based on fuzzy logic classification is proposed. Finally, a potentially noise-free image is obtained by unbiased inverse VST. Experimental results show that the proposed algorithm is competitive with several popular Poisson denoising techniques in both visual and objective metrics.     

SVD-based digital image watermarking using adaptive generated watermark

In this paper, an evolutionary-based algorithm for digital image watermarking is proposed. Robustness and imperceptibility are two important properties of digital image watermarking which conflict with each other. A singular value decomposition (SVD)-based intelligent algorithm is applied to make a trade-off between robustness and imperceptibility. In the proposed algorithm, first, the original image is divided into 8×8 blocks. Then some special blocks are transformed into discrete cosine transform (DCT) domain. SVD decomposition of DCT coefficients of special blocks is the next stage of the method. Finally, a binary watermark image is embedded into the singular values by a quantisation method. The major property of the proposed method is the generation of a binary watermark image using genetic algorithm to solve the optimisation problem between robustness and imperceptibility. So the watermark could be variable and image adaptive. The simulation results show improvement in robustness of the proposed method against several attacks such as noise contamination, filtering, joint photographic experts group (JPEG) compression and some geometrical attacks like cropping, rotation and rescaling in comparison with a recently similar existing work.     

小型液晶屏盒内缺陷检测中纹理的消除

提出了一种在小型液晶屏盒内缺陷检测中消除纹理影响的方法,并成功制作了国内首台小型液晶屏盒内缺陷自动光学检测仪.本方法的基本原理是:采集标准液晶屏图像,利用其图像数据建立补偿矩阵,在检测的预处理过程中使用补偿矩阵消除纹理.这种方法有效解决了液晶屏纹理影响检测结果的难题.理论分析及现场实验结果表明,该方法达到消除液晶屏纹理的同时完整保留缺陷的目的,使液晶屏盒内缺陷自动检测的准确性大大提高.     

基于奇异值分解的数字图像的特征提取

特征提取是图像识别过程中的重要的一部分,本文利用奇异值分解这种有效的代数特征提取方法,获得图像的有效特征描述,把同一类图像集中的各幅图像的信息矩阵的奇异值向量作为矩阵中的一列构成图像特征矩阵,求出图像类间离散度矩阵和图像类内离散度矩阵,比较它们的相似程度,最终获得有用和有效的特征。     

基于Hilbert谱奇异值的滚动轴承故障诊断

提出一种基于Hilbert谱奇异值的故障特征提取方法,将其与支持向量机结合应用于轴承故障诊断。利用小波阈值降噪的方法对拾取的轴承故障振动信号进行滤波降噪,然后利用经验模式分解将降噪信号分解为若干个IMF分量之和,对每个IMF分量进行Hilbert变换得到振动信号的Hilbert谱,对Hilbert谱进行奇异值分解得到反映轴承状态特征的奇异值序列,再利用奇异值作为特征向量,应用支持向量机进行轴承故障诊断,并对不同核函数的诊断结果进行了分析比较。对正常轴承、内圈故障、外圈故障、滚动体故障的实际信号的诊断验证了该方法可的有效性。