数学形态学和小波变换的红外图像处理方法

为了从红外热图像中识别和跟踪目标,提出了一种采用数学形态学和小波变换相结合的红外图像处理新方法。该方法首先利用形态开、闭滤波器组对原始图像进行平滑处理;然后采用一种改进的数学形态学的分水岭算法对红外图像进行分割,同时利用考虑了图像纹理信息的小波阈值对分割的图像进行滤波处理;最后针对过分割问题提出了一种利用区域灰度中值对分割区域进行融合处理的算法。实验结果及算法性能评估结果表明,该方法能够较好地解决红外目标的识别问题,具有较好的实用性。     

多特征加权的纹理图像FCM分割算法

为了实现自然纹理的分割,根据自然纹理的不规则性和随机性强的特点,提出了一种由纹理的GMRF模型提取纹理的统计特性和样本方位特性的多特征融合的纹理图像分割算法。考虑到不同特征值对分类的不同影响,采取样本加权的模糊c均值分割算法控制各特征在聚类约束力上的可控制性。通过人工合成图像和自然纹理的分割实验证明,该方法具有较强的抗噪声能力,并且对于结构相似的粗纹理进行分割能有效地克服小岛区域和保持区域一致性。     

基于频域滤波的电缆线芯图像纹理消除研究

电缆作为电力传输的重要部分,其线芯质量关系到用电安全及传输效率。 在利用机器视觉技术测量电缆线芯结构参数 时,刀痕纹理严重影响了图像分割效果,降低了测量精度。 针对这一问题,提出了一种基于改进频域滤波的线芯截面图像刀痕 纹理消除方法:首先采用相对总变差模型分解线芯截面图像,得到纹理图像描述;对线芯截面纹理图像进行快速傅里叶变换,分 析刀痕纹理在频域中的能量主方向;研究改进型巴特沃斯陷波滤波器,设计椭圆形的邻域窄阻带以降低刀痕纹理对应的带状区 域能量,从而消除刀痕纹理。 实验结果表明,该方法的刀痕纹理消除率可达 92% ,提升了电缆线芯质量检测的效率。     

一种基于多通道Gabor滤波的纹理提取方法

研究了一种在多种纹理叠加的复杂轮胎X光图像中进行纹理提取的新方法.轮胎X光纹理在频谱中被看作是局部空间频率的辐射模型,并且不同纹理对应的主频率聚集区域基本能相互分开.基于Gabor滤波器对空间频率的带通滤波能力以及方向、形状、带宽和通带中心频率的可调节特性,构造了多通道Gabor滤波器组提取每种纹理对应的主要频率成分,提取的频率成分经过频-时域变换就实现了纹理提取.考虑到实际纹理灰度分布不是理想的正弦分布,改进了传统的Gabor滤波方式,使其保留了纹理的低频成分.实验结果表明,图像中叠加的多种纹理能被准确提取,纹理区域边界也没有失真,并且有效降低了纹理之间频谱混迭造成的影响.     

半全局立体匹配算法的改进研究

针对半全局立体匹配中单一 Census 精度不足且边缘区域视差效果较差的问题,提出一种基于颜色差信息 BTCensus 和加入分割约束的半全局匹配算法。首先该算法采用联合三通道 BT 算法和 Census 变换的计算方式实现代价计算,缓解重复纹理歧义并提高不连续区域配精度;同时将图像分割得到的场景信息作为约束对代价聚合函数进行改进,保证割块内部的纹理平滑,提高边界纹理区域的匹配精度;最后引入图像分割信息进行分步中值滤波,平滑视差图。实验表明,该算法在重复纹理区域、视差不连续区域和弱纹理区域都取得了较好的视差效果,可获得高匹配率的视差图。     

基于区域生长法的自适应图像分割的网眼织物瑕疵检测

为了解决网眼织物在织造过程中的实时检测问题,提出了一种基于频域滤波和区域生长法的瑕疵区域自适应分割方法。通过分析傅里叶变换后的功率谱获得网眼织物纱线分布特征后,通过设计滤波器自动滤除了与织物纹理背景相关的频率成分以削弱织物自身纱线分布结构对瑕疵检测精度的干扰。利用滤波后图像的灰度分布特点自适应地求取了分割参数,并使用区域生长法实现了对瑕疵区域的自动分割。利用对获得的二值图像进行形态学运算,得到了最终的瑕疵区域分割图像。实验结果表明,该算法能较准确地对瑕疵进行定位,瑕疵区域分割精度较高,自适应性和抗干扰能力较强,可以满足工业检测的要求。     

基于区域划分的多特征纹理图像分割

由于纹理图像的复杂性和多样性,仅依靠传统的单一特征实现纹理图像分割无法满足其对分割精度的要求。本文提出结合区域划分的多特征纹理图像分割方法。首先,依据像素灰度的空间相关性定义多个纹理特征;然后利用区域划分将图像域划分成不同子区域,待分割同质区域由这些子区域拟合而成;通过分别定义多个特征图像的同质区域之间的异质性势能函数和刻画各子区域邻域关系势能函数来定义全局势能函数,并构建非约束吉布斯概率分布,从而建立纹理分割模型;最后,采用M-H算法采样上述概率分布,从而获得最优图像分割结果。分别对模拟纹理图像、遥感图像、自然纹理图像和SAR海冰图像进行了分割实验,并与利用单一特征得到的分割结果进行对比分析,定性和定量的测试结果验证了算法的有效性。     

Gabor滤波器在自动视觉表面检测中的应用

提出一种基于Gabor滤波器的纹理表面缺陷检测方法。该方法基于Gabor滤波方案，计算纹理图像与某一Gabor滤波器卷积能量输出响应。选择Gabor滤波器的最优参数，采用简单的二值化技术将纹理差别转化为可检测的滤波器输出。纺织品、铣磨面等结构纹理和皮革、砂纸等随机纹理的实验结果证明了该方法在识别不可定量测量缺陷方面的有效性。     

利用自适应纹理分布的活动形状分割前列腺磁共振图像

基于前列腺磁共振图像性质,提出利用自适应纹理分布的活动形状图像分割方法来自动分割前列腺磁共振图像.该方法首先通过图像的分类与拟合确定感兴趣的腺体区域,同时估计若干形状参数用于分割过程中调整形状;然后融人多重纹理信息,建立纹理一致测度,将传统的活动形状按照自适应的纹理判别步骤细分为纹理分布形状与补充形状,提高活动形状的搜索匹配能力.在搜索匹配部分,利用已估计参数优化活动形状搜索的初始估计,并根据纹理分布形状和补充形状调整迭代过程.实验结果表明,该方法分割出来的前列腺轮廓与金标准的Hausdorff距离为6.00 pixel,分割精度为93％.该方法对活动形状的改进是有效的,利用自适应纹理分布的活动形状能够自动、准确地将前列腺从磁共振图像中分割出来.     

基于灰度差分和脊线检测算子的虹膜裂缝检测

虹膜裂缝是用于虹膜识别的特定形状虹膜纹理之一。为了解决在复杂背景下提取裂缝特征的问题,在对虹膜裂缝特征进行分析的基础上,提出了一种虹膜裂缝检测方法。首先用灰度差分模板得到裂缝纹理候选区域,并根据裂缝的先验知识筛选裂缝候选像素,在原图上分割成小区域;其次,采用脊线检测算子在小区域范围内对各种纹理进行检测;最后在二值化图像上计算该区域能量密度,排除非裂缝纹理,实现裂缝纹理的检测。该方法在图库中人工标定的1 201处裂缝中检测正确检出率达到94.09%。实验结果表明,本文方法在不需要手动去掉卷缩轮内区域的情况下,能够在多种虹膜纹理并存的复杂背景下检测出虹膜裂缝纹理。     

Automated Surface Inspection Using Gabor Filters

In this paper we present a machine vision system for the automatic inspection of defects in textured surfaces found in industry. The defects to be inspected are those that appear as local anomalies embedded in a homogeneous texture. The proposed method is based on a Gabor filtering scheme that computes the output response of energy from the convolution of a textured image with a specific Gabor filter. The best parameters of a Gabor filter are selected so that the energy of the homogeneous texture is zero, and any unpredictable defeats will generate significantly large energy values. A simple thresholding scheme then follows to discriminate between homogeneous regions and defective regions in the filtered image. This transforms texture differences into detectable filter output. The experiments on structural textures such as leather and sandpaper have shown the effectiveness of the proposed method.     

Creating machinable textures for CAD/CAM systems

Texture is applied on three-dimensionally modelled surfaces in computer graphics to enhance visual effect. This research focuses on the development of three-dimensional textured surfaces which are suitable for manufacturing. Three approaches for creating the three-dimensional texture are presented.The first approach is to process a design from either an artist's sketch or an image from a two-dimensional scan. The second approach uses a three-dimensional scanned texture. Both these approaches depend on the quality of the scanned image and are more tedious than the third approach, which is to convert texture using parameters and is a more direct approach. In the user interface design, two custom-made forms are developed to cater for both regular and irregular textures.The case studies have shown that the textures created are not only good for a visual effect, but are also machinable. The development work is incorporated into the ArtCAM system which is a specialised CAD/CAM system that is capable of generating three-dimensional shapes from two-dimensional artwork.     

基于线性模板和模糊规则的虹膜坑洞纹理检测

基于虹膜表面特定形状纹理的虹膜识别方法是虹膜识别的一种有效手段。然而,由于虹膜纹理复杂多样,形态各异,因此如何快速、准确地检测出虹膜表面各种不同类型的特征纹理仍然是目前亟待解决的一个难题。针对坑洞纹理的检测,提出一种基于定长线性模板和模糊规则的虹膜坑洞纹理的检测方法,首先使用一个定长的线性模板提取出虹膜表面所有纹理的边缘。然后,根据坑洞纹理与其他类型的纹理在形状上存在的不同,定义相应的模糊规则,从各种不同类型的纹理中识别出坑洞纹理。该算法对图库中人工标定的坑洞纹理检测的正确率为86.06%;对色素斑、裂缝以及眼睑的排除率分别为84.62%,71.56%,100%。实验结果表明,该方法能够从背景复杂的可见光虹膜图像中识别出坑洞纹理,为虹膜的特征提取提供了一个新的思路。     

Automatic measurement of volume percentage stroma in endometrial images using texture segmentation

The popularity of digital microscopy and tissue microarrays allow the use of high-throughput imaging for pathology research. To coordinate with this new technique, it is essential to automate the process of extracting information from such high amount of images. In this paper, we present a new model called the Subspace Mumford-Shah model for texture segmentation of microscopic endometrial images. The model incorporates subspace clustering techniques into a Mumford-Shah model to solve texture segmentation problems. The method first uses a supervised procedure to determine several optimal subspaces. These subspaces are then embedded into a Mumford-Shah objective function so that each segment of the optimal partition is homogeneous in its own subspace. The method outperforms a widely used method in bioimaging community called k-means segmentation since it can separate textures which are less separated in the full feature space, which confirm the usefulness of subspace clustering in texture segmentation. Experimental results also show that the proposed method is well performed on diagnosing premalignant endometrial disease and is very practical for segmenting image set sharing similar properties.     

基于形态学与支持向量机的虹膜坑洞纹理检测

坑洞纹理是虹膜表面上一种重要的特征纹理。对可见光虹膜图像而言,如何能不受睫毛、眼睑、光斑及光照不均匀等干扰因素的影响,并能快速、准确地提取出该特征纹理仍然是目前一个亟待解决的难题。提出一种基于形态学和支持向量机(SVM)的可见光虹膜坑洞纹理检测方法。首先使用灰度形态学和二值形态学相结合的方法提取出所有目标纹理;然后使用区域生长方法定位所有目标纹理并计算各个目标纹理的特征向量;最后再使用SVM和定义约束条件的方法提取出最终的坑洞纹理。通过实验证明:该方法能较好地克服光斑等干扰的影响,对坑洞状纹理的检出率高于其他同类方法。     

基于BAB策略与SVM的虹膜块状纹理检测

块状纹理是虹膜表面上一种重要的特征纹理。对于可见光虹膜图像而言,由于受眼睑、光斑及光照不均匀等因素影响,如何准确地提取出该特征纹理仍然是目前一个亟待解决的难题。提出一种基于分枝定界(BAB)策略和支持向量机(SVM)的可见光虹膜块状纹理检测方法。首先使用线性模板来提取虹膜上所有目标纹理;然后,采用BAB策略定位所有目标纹理并计算每个目标纹理的特征向量;最后,使用SVM实现对块状纹理的检测。通过实验证明,该方法对于虹膜块状纹理的检出率和检测时间均优于同类方法。     

基于双线性模板和分块策略的虹膜色素斑检测

本文色素斑是一种重要的虹膜特征纹理。然而,由于该纹理颜色深浅不一,形状各异,因此如何从具有复杂背景的可见光虹膜图像中检测出该纹理仍然是目前亟待解决的一个难题。提出一种基于双线性模板和分块策略的可见光虹膜色素斑检测方法,首先使用两个不同长度的线性模板来提取虹膜上所有不同类型的特征纹理;然后,采用分块策略得到一个能描述该纹理灰度分布特征的特征向量;最后,使用BP神经网络实现对色素斑纹理的检测。该算法对图库中色素斑纹理检测的正确率达到81.25%,对非色素斑纹理的正确识别率达到83.97%,总的正确检测率达到83.23%。实验结果表明,该方法能够从背景复杂的可见光虹膜图像中检测出色素斑纹理,为虹膜的特征提取提供了一个新的思路。     

Directional Fractal Signature Analysis of Self-Structured Surface Textures

Currently available directional fractal signature (DFS) methods are not suited for self-structured surface textures since they base on the assumption of Brownian fractal or they do not use the entire image data in calculation. To address these difficulties, two new DFS methods were developed in this study, i.e., an augmented blanket with rotating grid (ABRG) method and a blanket with shearing image (BSI) method. The performance of these methods in measuring surface roughness and directionality, the capacity for quantifying multi-patterned textures, and the ability to detect differences between textures of self-structured surfaces were evaluated. The methods were compared against a blanket with rotating grid (BRG) method. Computer-generated images of self-structured surface textures with different roughness, directions and patterns, and atomic force microscope images of real self-structured surfaces were used. The computer texture images were generated using a specially developed motif-based texture generator. Results obtained showed that the ABRG method is more accurate and reliable than the BRG and BSI methods.     

Fast Defect Detection in Textured Surfaces Using 1D Gabor Filters

In this paper, we present a fast machine vision method for the automatic inspection of defects in textured surfaces. Traditional 2D Gabor filtering schemes have been shown to be very effective for detecting local anomalies in textured surfaces of industrial materials. However, they are computationally expensive and sensitive to image rotation. In order to alleviate the limitations of 2D Gabor filtering, we first use a 1D ring-projection transformation to compress a 2D grey-level image into a 1D pattern, and then employ a 1D Gabor filter to detect defects embedded in a homogeneous texture. Given a problem with image size N × N and filter window W × W, the computational complexity can be reduced significantly from O(W ² N ² ) in the 2D Gabor space to O(WN ² ) in the 1D Gabor space, and the detection results are invariant to rotation changes of a texture. The experiments on structural textures such as a wooden surface, an LCD display, and a machined surface, and statistical textures such as granite, leather, and sandpaper have shown the efficiency and effectiveness of the proposed method. ID="A1" Correspondence and offprint requests to: Dr Du-Ming Tsai, Department of Industrial Engineering and Management, Yuan-Ze University, 135 Yuan-Tung Road, Nei-Li, Tao-Yuan, Taiwan. E-mail: iedmtsai@saturn.yzu.edu.tw