Fabric Defect Detection Using Adaptive Wavelet Transform

A method of woven fabric defect detection using the wavelet transform adaptive to the fabric has been developed. With reference to the orthogonality constrains of Daubechies wavelet, by taking the minimization of the energy or the gray level of the pixels hi the output sub-images as the additional conditions and using the random algorithm method, two sets of wavelet filters adapted to the fabric texture were formed. The original images of normal fabric texture and the fabric texture with defects were decomposed into horizontal and vertical sub-images by using these filters and the feature indices of these sub-images were also extracted. By comparing the feature indices of the normal texture with that of the defective texture, the fabric defects can be successfully detected and located.     

采用部分灰度压缩扩阶共生矩阵的煤和煤矸石图像识别

提出一种基于部分灰度压缩扩阶共生矩阵的煤和煤矸石图像识别方法.首先,对煤和煤矸石0~255级灰度图像的前部分灰度信息作灰度级压缩和灰度矩阵扩阶处理,对剩余灰度级部分保持原灰度级不变;然后,根据灰度共生矩阵纹理特征分析理论,分别计算压缩扩阶后的煤和煤矸石灰度图像的能量、熵、惯性矩及相关性.最后,对煤和煤矸石各100张样本采集图像进行处理,并依据特征参数分类识别.结果表明:基于部分灰度压缩扩阶共生矩阵的特征参数能够很好地对煤和煤矸石图像进行有效识别,总的正确率达到93.5%.     

融合方向测度和灰度共生矩阵的纹理特征提取算法研究

为了解决提取图像纹理特征时所遇到的纹理方向抑制问题,提出一种融合方向测度和灰度共生矩阵的纹理特征提取算法。该算法通过灰度共生矩阵,提取图像的Haralick特征,其中包括对比度、相关性、能量、逆差矩等,然后利用方向测度引入权值因子,并将其与所提取的Haralick特征相融合,最后对融合后的各个分量进行高斯归一化处理,获取最终的纹理特征集。实验结果表明,与采用灰度共生矩阵方法相比,该算法可以有效的避免图像纹理方向的抑制,所提取的纹理特征具有更强的图像识别能力,对Brodatz标准纹理库分类的正确率也有一定的提高。     

基于多特征的旋转不变纹理图像检索

为了消除图像旋转对图像检索的影响,提出了一个基于非下采样轮廓波变换、灰度共生矩阵和新相似性度量的旋转不变纹理图像检索算法。非下采样轮廓波变换具有各向异性和平移不变性。灰度共生矩阵描述了图像的方向性、邻近空间关系和方差的变化范围。通过计算每个非下采样轮廓波变换尺度下的所有子带的平均能量和平均标准差,灰度共生矩阵的二阶矩角、惯性熵、惯性矩、反差分矩,惯性相关系数的均值、标准差,得到具有旋转不变性的纹理多特征。提出一种新的相似性度量以改进纹理图像检索性能。实验结果表明,对于1个含640幅图像的旋转纹理图像库,与基于双树复小波变换的方法相比,该文图像检索算法将旋转纹理图像检索准确率由73.28%提高至80.71%。     

基于灰度共生矩阵的高效积分图像计算(英文)

基于局部特征描述符的主要方法正在被用于纹理分类、目标检测和识别.灰度共生矩阵(GLCM)是一个体现纹理图像的很流行的方法,也被证明是一个非常强大的纹理分析工具.灰度共生矩阵所带来的子图像重叠非常严重,要花大量的时间去计算.本文提出一种高效灰度共生矩阵的计算方法,并通过实验结果来显示其效率.     

基于SVM的图像分类研究

支持向量机（SVM）是一种优秀的分类方法,灰度共生矩阵（GLCM）是一种很好的纹理分析方法;本文研究了图像的颜色、纹理等特征对利用支持向量机进行图像分类的影响,提出了一种使用灰度共生矩阵进行综合特征提取的SVM图像分类方法。实验结果表明SVM方法的可行性和有效性,并得出图像综合特征的提取有利于分类效果的提高。     

基于高斯-马尔可夫随机场模型的织物纹理特征参数提取方法

将高斯-马尔可夫随机场用于织物纹理的建模.在估计模型参数时,采用不完全 Cholesky共轭梯度法.获得特征参数后,通过纹理合成方法,仿真原始纹理图像.试验表明:仿真织物纹理图像与真实织物纹理图像具有较为一致的视觉效果,从而论证了高斯-马尔可夫随机场模型对织物纹理建模的有效性.     

雷达辐射源信号双谱二次特征提取方法

现有的双谱特征提取方法可以满足信号分类识别，但是出现了交叉项、平凡双谱以及特征维数过高等一些问题。针对以上问题，提出一种双谱二次特征提取方法，将双谱转化为灰度图像，以灰度值表示双谱幅度；再利用图像处理技术提取双谱二次特征，提取出能够表征辐射源信号双谱图像纹理信息的灰度共生矩阵特征集；将该特征集与Hu-不变矩特征集进行对比实验。仿真结果表明：该方法具有更好的分类识别性能，对于CW、LFM和NLFM信号的平均识别率均达90%以上。     

Fabric Defect Detection Using GMRF Model

It has been testified that the Gauss Markov random field model is most suitable for the characterization of fabric texture among a variety of available models because of its approximately constant character and the normality of the gray-level distribution found with typical fabric images. However, the general Gauss-Markov random field(GMRF) method for fabric defect detection is not always ideal in practice since in some cases, the estimated model parameters make the Markov error covariance not positively definite, which may render the method to fail thoroughly. In this paper, the use of the GMRF model for defect detection of fabric is discussed and an approach to this problem is proposed. Some detailed texture may be overlooked in this way, but good detection results can still be expected as far as fabric defect detection is concerned.     

应用图像分析技术自动识别织物的组织结构

通过对机织物图像的表面形态分析,建立3种基本组织(平纹、斜纹和缎纹)的表面纹理模型,运用傅里叶变换技术得到3种基本组织的频谱模型,并建立表面纹理模型和频谱图模型之间的对应关系,从而为运用图像分析和人工智能技术自动测量织物的结构参数、识别机织物类型奠定理论基础.实验结果证明,这种方法是准确可靠的.     

基于类别共生矩阵的纹理疵点检测方法

根据有规则纹理的特点,提出了基于类别的共生矩阵来描述纹理特征,从而很好地将正常纹理与疵点区分开.分析了传统的灰度共生矩阵在计算纹理特征时计算量大,且分辨能力差的缺点.为了克服灰度共生矩阵在计算量和分辨能力上的缺点,定义了类别共生矩阵.在类别共生矩阵的算法中,首先学习纹理的一些基本特征以确定类别共生矩阵的一些关键参数,如纹理的概率密度分布、纹理的主方向和周期,以及分类准则等重要参数,然后计算类别共生矩阵并提取白疵点增强、黑疵点增强和一致度等三个特征,最后采用异常点检测的方法即可很好地区分正常纹理和疵点.实验证明,该方法比已有的灰度共生矩阵计算量小,并具有更突出的分辨纹理和疵点的能力.     

基于FPGA的遥感图像纹理特征提取算法实现技术

针对遥感图像在轨云检测技术中图像纹理特征提取功能的实现需求,采用FPGA来实现纹理特征提取算法,以此提高遥感图像的纹理特征的提取效率。设计中充分利用FPGA片上RAM、ROM、寄存器等逻辑资源,发挥FPGA并行处理的优势,构建了16个灰度共生矩阵并行计算的结构,进一步计算得到4个纹理特征参数。利用Xilinx公司的System Generator工具实现设计并通过仿真。以Spartan-6 LX45 FPGA为目标芯片的验证实验表明:设计具有运算性能高、耗费逻辑资源少的特点,能够满足遥感图像在轨云检测的实时处理需求。     

基于Wold纹理模型和分形理论的织物疵点检测

应用基于Wold纹理模型,从织物纹理中提取非确定性随机结构.通过正常纹理和疵点纹理中随机结构的表面维分形特征曲线比较,以距离为判别标准定位疵点.几种典型织物疵点的分析表明,基于Wold纹理模型和分形理论的分析方法能准确地检测出织物表面的多种疵点.     

基于GLCM的人群行为识别研究

随着社会经济的发展和城市化进程的加快,如何对公共场所人群行为进行监控,防止由于人群密度过高导致的重大伤亡事件发生是很有必要的。结合混合高斯模型和灰度共生矩阵提出一种监控人群行为的算法,将纹理特征通过灰度共生矩阵计算特征量,机器学习所有特征量来判断人群行为,并通过标准数据集和自行拍摄数据集验证算法的有效性。采用人群的整体特征来表征不同人群,通过机器学习来区分不同人群的行为情况,可用于安防监控、资源管理等领域。     

Data Analytical Techniques for Surface Texture Characterization of Knitted Fabrics Based on Laser Triangulation Scanner

The paper presents a study of surface texture character-ization of knitted fabrics during simulated wear based onlaser-scanned profiles.The three-dimensional profilewas obtained by scanning the fabric surface using the la-ser trlangulation technique.The analytical techniquesused to derive quantitative parameters Included statisticalanalysis,fractal analysis,and Fourler analysis.     

基于统计与频谱模型特征融合的纹理图像分割

纹理分析方法主要包括统计法、结构法和频谱法。由于不同纹理分析方法的侧重点和适用对象不一样,传统的单一特征分析方法存在一定的局限性。结合统计法和频谱法对纹理进行分析,利用灰度共生矩阵得到纹理的统计特征,应用Gabor变换得到多尺度、多方向的纹理特征,提出一种依据纹理宏、微特性加权的新的特征融合的方法,最后进行K均值聚类得到分割结果。实验结果表明,与传统应用单一纹理分析方法相比,该方法在保持边缘准确性和区域一致性上有一定程度的提高。     

表面纹理的参数评定研究

针对表面纹理的参数评定问题,提出了表面纹理可以从纹理图案和纹理粗糙度两方面入手加以评定的观点．重点探讨了运用三维粗糙度的几项特征参数定量地评定表面纹理问题．介绍了特征参数的计算原理、测量方法,并且对所测的数据进行了统计分析,拟定了部分纹理的参数标准,验证了采用特征参数评定表面纹理的可行性     

RGB and LBP-texture deep nonlinearly fusion features for fabric retrieval

Fabric retrieval is very challenging since problems like viewpoint variations, illumination changes, blots, and poor image qualities are usually encountered in fabric images. In this work, a novel deep feature nonlinear fusion network(DFNFN) is proposed to nonlinearly fuse features learned from RGB and texture images for improving fabric retrieval. Texture images are obtained by using local binary pattern texture(LBP-Texture) features to describe RGB fabric images. The DFNFN firstly applies two feature learning branches to deal with RGB images and the corresponding LBP-Texture images simultaneously. Each branch contains the same convolutional neural network(CNN) architecture but independently learning parameters. Then, a nonlinear fusion module(NFM) is designed to concatenate the features produced by the two branches and nonlinearly fuse the concatenated features via a convolutional layer followed with a rectified linear unit(ReLU). The NFM is flexible since it can be embedded in different depths of the DFNFN to find the best fusion position. Consequently, DFNFN can optimally fuse features learned from RGB and LBP-Texture images to boost the retrieval accuracy. Extensive experiments on the Fabric 1.0 dataset show that the proposed method is superior to many state-of-the-art methods.     

基于图像分解的织物组织自动识别

利用维纳滤波从水平和垂直两个方向分解织物图像,经过合理的二值化,分别获得包含经纱和纬纱特征信息的子图像,从而计算经纬纱密度等参数.综合子图像的部分信息,可以确定织物组织点的位置,再利用每个组织点区域内经纬向纹理特征,判断织物组织类型.经过对平纹、斜纹、缎纹3种基本组织布样的实验,表明该方法可以很好地识别织物组织类型.     

Influencing Factors of Fabric-Skin Adhesion Based on Gray Correlation Analysis Method

Fabric-skin adhesion was objectively described by the indices of the maximum adhesion force F_(max) , the maximum separation distance L_(max) ,and adhesion work W as well as the adhesion force-separation distance curve. Firstly,gray correlation analysis method was adopted to investigate the correlation levels between adhesion indices,and secondly the relative importance of fabric structural parameters to fabric-skin adhesion,as well as the correlation levels between skin adhesion, water absorption, and wicking properties of the fabric. The results prove that W exhibits clear correlations with both F_(max) and L_(max) , yet the relevance between F_(max) and L_(max) is weak. Fabric adhesion indices are most associated with fabric mass and least with fabric thickness,whereas fabric wicking and water absorption present closest correlation with fabric thickness. Therefore, it is concluded that the relevance between fabric wicking, water absorption, and skin-adhesion properties are rather comprehensive than straight.