基于局部区域灰度矩图像边缘定位方法的研究

在采用机器视觉测量零件尺寸的过程中,提出了一种基于局部区域灰度矩图像边缘定位方法。该方法先利用中值滤波对采集到的图像进行处理,然后通过Prewitt算子对边缘进行初步定位,在此基础上,对图像边缘附近的9个像素的灰度值进行一维灰度矩定位,最后采用误差分析和处理的方法,获得更精确的边缘位置。实验证明,该方法能有效提高边缘检测精度,在检测分辨率为1000DPI情况下,尺寸检测误差小于2um。     

基于HALCON的双目立体视觉工件尺寸测量

为了实现对大批量生产工件尺寸的精密测量，该文基于双目立体视觉视差原理，设计了一种工件尺寸实时测量系统；系统通过两个工业数字摄像头实时采集工件不同角度的两幅图像，并基于机器视觉软件HALCON的HDevelop开发环境，对两幅图像进行相关预处理，获取感兴趣区域。提出一种基于canny的亚像素边缘检测、边缘细化、基于atukey权重函数的最小二乘法边缘拟合及边缘均匀取点算法获取精确的边缘特征点，并通过双目系统标定、极线约束与NCC相结合的立体匹配、世界坐标转换、几何计算相结合的非接触式双目视觉方法进行尺寸测量。通过大量实验证明，该方法可有效地获取工件特征点的三维坐标值，不借助外部测量仪器实现工件关键尺寸的高精度实时检测，检测精度达±0.2mm以上，简单快捷，具有较好的准确性和实时性。     

基于神经网络的零件缺陷机器视觉识别系统

零件缺陷检测是保证零件使用安全的重要手段。传统的零件缺陷检测法需要有操作人员参与其中,易受主观因素影响,检测的效率及精度得不到良好的保证。而采用机器视觉技术的检测法可实现实时在线的自动检测,无需人工参与,这就极大的提高了生产效率。本文以小轴承表面为研究对象,针对微小轴承的表面结构、尺寸、检测精度和缺陷特征,设计了基于BP神经网络的零件缺陷机器视觉在线自动检测系统,其采用机器视觉技术,构建了BP神经网络检测识别模型,采用进行图像特征提取的间接识别方法,对微小轴承缺陷进行实时检测。实验结果证明了人工神经网络模型的检测能力的可靠性。     

基于机器视觉的圆孔尺寸测量系统研究

针对常见的圆形目标检测,开发了一套基于机器视觉技术的圆孔尺寸测量系统,采用混合噪声滤波、模糊和遗传边缘检测方法从采集的零部件图像中,提取出圆孔图像的边缘,通过插值圆度检测方法,精确测量出圆孔的半径、面积等特征参数。测试表明,测量系统的平均值误差为0.010 5 mm,远远低于人工测量的平均值误差0.027 3 mm,能够满足工业应用中的精确性要求。     

基于点阵标定板的视觉测量系统的标定方法EI北大核心CSCD

在应用机器视觉技术进行测量时,测量系统的像素当量、系统误差和光源强度等因素均会对测量精度造成影响,因此必须对像素当量和系统误差进行标定,并分析光源强度对工件图像边缘位置的影响。提出一种基于点阵标定板的视觉测量系统综合标定方法,在提取标定圆圆心坐标的基础上,计算圆心距物理尺寸和像素尺寸的比值,得到像素当量;建立圆心实际坐标和理论坐标的二元三次误差模型,并利用最小二乘法拟合求解误差模型系数;通过确定光源强度引起的图像边缘位置误差补偿量,实现测量系统的综合标定。实验结果表明,该方法可以有效提高系统的测量精度。     

基于小波变换和数学形态学的孔径测量研究

针对视觉测量中硬盘圆孔孔径容易受到圆孔周围局部强反射、外部噪声的影响和测量精度不高的问题,提出一种基于小波变换、数学形态学和机器视觉相结合的圆形零件孔径测量方法。对摄像头采集圆孔图片,通过小波变换分解出高低频图像,利用小波变换对高频部分进行边缘提取,数学形态学对低频部分进行边缘提取,然后进行小波边缘融合,获取有效的圆孔边缘。利用最小二乘法对边缘进行孔径尺寸计算,经实验验证,该算法与其他算法相比,在有效地保留圆孔周围边缘信息的情况下,对周围噪声进行了有效的抑制,检测精度为0.01 mm以内,实验结果表明此方法简单易行,且具有较高的精度。     

基于点阵标定板的视觉测量系统的标定方法

在应用机器视觉技术进行测量时,测量系统的像素当量、系统误差和光源强度等因素均会对测量精度造成影响,因此必须对像素当量和系统误差进行标定,并分析光源强度对工件图像边缘位置的影响。提出一种基于点阵标定板的视觉测量系统综合标定方法,在提取标定圆圆心坐标的基础上,计算圆心距物理尺寸和像素尺寸的比值,得到像素当量;建立圆心实际坐标和理论坐标的二元三次误差模型,并利用最小二乘法拟合求解误差模型系数;通过确定光源强度引起的图像边缘位置误差补偿量,实现测量系统的综合标定。实验结果表明,该方法可以有效提高系统的测量精度。     

基于机器视觉的汽车仪表盘指针检测

基于机器视觉构建汽车仪表自动检测系统已成为实现仪表生产测试的主要途径。基于图像空间到参数空间的收敛映射提出了改进CM-Hough变换检测算法对汽车仪表盘的指针位置实施检测。首先,介绍了基于机器视觉对汽车仪表的符号片及指针进行自动检测的系统组成。接着,针对经典Hough变换发散映射运算量大的缺点,提出了基于收敛映射CM-Hough变换进行指针检测的算法,给出了利用Hough变换的检测结果直接解算指针位置的方法。最后,将算法由软件编程实现,通过测试实验验证了算法的实时性和有效性,测试数据分析显示了指针位置误差小于1%。     

基于机器视觉的高精度测量与装配系统设计

在对某些精密产品实现自动化生产过程中,存在难以对装配该产品所需的多种装配小零件进行高精度自动测量与装配的问题。针对该问题,搭建了基于机器视觉技术的自动化测量与装配系统。基于Halcon图像处理软件平台,对零件图像进行了中值滤波、图像增强等预处理;采用了Canny算法对零件求取像素精度的边缘,并运用椭圆曲线拟合法获取了亚像素精度边缘;建立了两种相机镜头畸变模型,采用径向排列约束(RAC)标定法与张正友标定法对相机进行了标定,并对标定精度进行了对比;实验结果表明：本系统的装配同轴度精度能达到0.05mm,零件尺寸测量标准差低于3.8μm,满足工业需求,可以解决工业实际问题。     

显微视觉检测系统精度分析及校准技术

为提高显微视觉检测系统的检测精度,建立了检测系统的误差模型,分析了影响检测系统精度的因素,给出了运动系统及视觉系统的误差补偿方法;提出了一种开环控制加位置检测的运动控制方式和一种基于多项式拟合的图像边缘检测算法;实验结果表明,提出的运动控制方式和边缘检测算法有效降低了显微视觉检测系统的系统误差,提高了微小型零件的检测精度。     

Real-time application of critical dimension measurement of TFT-LCD pattern using a newly proposed 2D image-processing algorithm

A critical dimension measurement system for TFT-LCD patterns has been implemented in this study. To improve the measurement accuracy, an imaging auto-focus algorithm, fast pattern-matching algorithm, and precise edge detection algorithm with subpixel accuracy have been developed and implemented in the system.The optimum focusing position can be calculated using the image focus estimator. The two-step auto-focusing technique has been newly proposed for various LCD patterns, and various focus estimators have been compared to select a stable and accurate one.Fast pattern matching and subpixel edge detection have been developed for measurement. The new approach, called NEMC, is based on edge detection for the selection of influential points; in this approach, points having a strong edge magnitude are only used in the matching procedure. To accelerate pattern matching, point correlation and an image pyramid structure are combined.Edge detection is the most important technique in a vision inspection system. A two-stage edge detection algorithm has been introduced. In the first stage, a first order derivative operator such as the Sobel operator is used to place the edge points and to find the edge directions using a least-square estimation method with pixel accuracy. In the second stage, an eight-connected neighborhood of the estimated edge points is convolved with the LoG (Laplacian of Gaussian) operator, and the LoG-filtered image can be modeled as a continuous function using the facet model. The measurement results of the various patterns are finally presented.The developed system has been successfully used in the TFT-LCD manufacturing industry, and repeatability of less than 30 nm (3σ) can be obtained with a very fast inspection time.     

基于图像处理的零件二维尺寸检测方法

为了提高零件二维尺寸测量精度和效率,提出一种基于图像处理的非接触式检测方法。对零件图像进行灰度化、采用自适应中值滤波算法来减小图像数据量以及除去噪声;通过最大类间方差法自动确定阈值获得二值图像;采用形态学梯度算子提取图像边缘,并用最小二乘法拟合像素尺寸;通过标准件法对系统进行标定获得零件尺寸。以6204-Z型深沟球轴承作为实验对象,获得轴承内、外径的尺寸。与游标卡尺测量结果比较,所提出的图像处理方法获得的内、外径平均误差达到9.6μm和0.7μm。通过实验对比与分析,方法快速有效,具有较高的准确性和良好的鲁棒性。     

Accurate measurement method of the thickness of a quartz pendulous reed by combining polarized reflectance and vision image

The detection of thickness provides important information on the production process of the quartz pendulous reed (QRP). Through real-time detection of QPR thickness, high-quality products can be produced, and defective products can be identified and removed. In order to improve detection accuracy and efficiency, a novel noncontact method for measuring the QPR thickness by combining the polarized reflectance and vision image is presented, which can be automated. Based on Snell’s law of fundamental optics, the image of the laser spot is obtained by a laser vision system. The selection criterion of the laser incident angle is analyzed. An improved edge detection algorithm is proposed to locate the subpixel edge of the laser spot, and the calculation method of the laser spot center is presented based on a set of subpixel edge-point data obtained by defining an error function which is to optimize the distribution of data points and minimize it. Experimental results show that the method proposed has good stability and high measurement accuracy, which can achieve noncontact accurate measurement of QPR thickness.     

基于Zernike矩的保持架直径测量方法

为了解决轴承保持架人工抽检费时费力等问题,同时提高工业生产自动化水平,简述了一种基于Zernike矩的保持架直径测量方法;以型号32007E的圆锥滚动轴承筐形保持架为例,提出了基于视觉的直径测量方法,分析CCD相机采集到的轴承保持架大小端面图像,进行图像预处理后,对Sobel算子边界点阈值进行重新设定,快速检测出保持架两端圆面可能存在的边缘点集,增加了有效圆检测算法,剔除部分偏离有效圆的点,再利用 Zernike 矩算子对有效的边缘点进行重新定位,检测出保持架两端圆面的亚像素边缘并计算其精确位置,最后对所得到的亚像素边缘点集进行最小二乘法拟合,获取保持架两端直径具体尺寸;实验表明,该方法测量结果与人工测量精度接近,甚至更高,具有良好的效果和实用价值。     

基于蚁群算法的辊式涂布涂层厚度图像检测

针对目前辊式涂布涂层厚度检测效率低、准确率低等问题，提出一种基于启发式蚁群算法的辊式涂布涂层厚度机器视觉检测方法。采用Canny算子的原理提取出边缘信息，得到了边缘点的先验知识；然后建立了改进的蚁群算法的边缘追踪模型，实现了信息素和启发信息对蚂蚁的导向作用，同时较好地避免了蚂蚁在非边缘区域的分布和行走，解决了传统蚁群算法中随机性与正反馈两种机制的协调问题，使用改进蚁群算法的机器视觉法进行测量实验，与机理建模法对比最大误差为5.74%，平均误差为4.04%，满足实际生产需要。     

基于数字图像处理方法的焦距测量技术研究

介绍了基于数字图像处理方法测量透镜焦距的玻罗板分划线快速自动甄别法。该图像处理方法鲁棒性好,采用Canny法进行分划线边缘检测。用重心法计算所有线条的重心,从重心往两边搜索线条,判断选取间隔大的刻划线对,可以提高测量精度。实验数据表明该方法的相对测量不确定度能达到8.5×10-4。为实现以数字图象处理为基础的焦距客观测量提供了一种可行的技术途径。     

Improved multi-scale wavelet in pantograph slide edge detection

The mainstream methods of pantograph slide edge detection are based on canny operator and multi-scale wavelet. The former has good single edge response but the edge is fractured, the latter performs good edge continuity but contains excessive edge points. This paper combines the advantages of both methods and proposes as an improved multi-scale wavelet edge detection method based on canny criteria. Firstly we filtered the pantograph image with edge-preserving symmetric near neighbor filter. Secondly calculated the Gaussian wavelet modulus and arguments at all levels of scale, then suppressed the non-maxima value of modulus along the corresponding arguments. At last, we integrated the modulus drawings at all levels of scale, and connected edge with applicable dual-threshold. Experiments results show that the improved algorithm has both satisfactory performances in single edge response and edge continuity, it markedly improves the efficiency of edge detection algorithm. Peak signal to noise ratio (PSNR) analysis finds that the improved algorithm exceeds canny operator and traditional multi-scale wavelet edge detection. Moreover, it has higher positioning accuracy, clearer details and better noise performance.     

Numerical study of centroid detection accuracy for Shack-Hartmann wavefront sensor

It is crucial for the wavefront sensor to reduce the influence of noise and enhance the detection accuracy of the centroid, which is also an important step for improving the performances of adaptive optics. In this paper, based on the theoretical analysis and by utilizing the Monte-Carlo simulation (MCS) technology, the results obtained by using the centroid algorithms have been compared in detail for the case, in which centroid of the signal is not at the center of the detection area and it also has a relatively large area. It is shown that the size of the signal spot and its centroid position have considerable influences on the centroid detection accuracy. The numerical results agree well with the theoretical predictions and the optimized parameters for each algorithm have been found. The study will be helpful for further optimizing the optical system and improving the centroid measurement accuracy of the wavefront sensor.     

基于机器视觉的工件尺寸和角度的测量

针对一款手机U盘芯片的二维尺寸的测量问题,提出运用图像处理的方法,实现手机U盘芯片长度和偏角的非接触式测量。通过COMS工业相机采集到U盘芯片的背光图像,在边缘检测的基础上,用Hough变换检测和定位U盘芯片边缘直线。针对在一条边上Hough变换对应检测到多条直线的情况,本文提出采用直线参数平均法拟合边缘直线,从而获得较精准的边缘位置。通过对已知尺寸的标准块进行相机标定,由此计算出精确的U盘芯片的尺寸和倾斜角度。实验表明：该算法能够满足工程上较高精度的检测要求。