基于深度学习的工业自动化包装缺陷检测方法

目的针对目前工业自动化生产中基于人工特征提取的包装缺陷检测方法复杂、专业知识要求高、通用性差、在多目标和复杂背景下难以应用等问题,研究基于深度学习的实时包装缺陷检测方法。方法在样本数据较少的情况下,提出一种基于深度学习的Inception-V3图像分类算法和YOLO-V3目标检测算法相结合的缺陷检测方法,并设计完整的基于计算机视觉的在线包装缺陷检测系统。结果实验结果显示,该方法的识别准确率为99.49%,方差为0.0000506,只使用Inception-V3算法的准确率为97.70%,方差为0.000251。结论相比一般基于人工特征提取的包装缺陷检测方法,避免了复杂的特征提取过程。相比只应用图像分类算法进行包装缺陷检测,该方法在包装缺陷区域占比较小的情况下能较明显地提高包装缺陷检测精度和稳定性,在复杂检测背景和多目标场景中体现优势。该缺陷检测系统和检测方法可以很容易地迁移到其他类似在线检测问题上。     

基于无监督深度神经网络的卷烟小包拉线缺陷视觉智能检测方法

目的 为减少ZB47包装机小包拉线缺陷投诉,基于无监督深度神经网络构建一种小包拉线缺陷视觉智能检测方法。方法 首先,在ZB47包装机CH转塔部位设计并加装小包图像采集装置,获得实时高清晰度小包图像。其次,将小包图像根据拉线位置进行固定位置的裁剪,从而减轻不同工况的环境背景影响并且加快检测速度。然后,构建自编码器–编码器结构的主干网络,同时叠加生成对抗网络中的判别器模块组成缺陷判别模型,并综合采用图像间、图像隐空间以及图像特征间的信息构建模型的损失函数。最后,使用裁剪后的正常小包拉线图像对构建的缺陷判别模型进行训练,并基于所有的正常小包图像得到异常阈值。结果 实际验证阶段,待检测图像的得分大于异常阈值即判断为异常图像,触发CH转塔部位的小包剔除装置将该缺陷小包剔除。生产现场测试表明,所提方法可以对典型小包缺陷进行快速准确检测,缺陷检测准确率为99.99%。结论 该方法能够满足生产现场卷烟小包拉线缺陷检测的准确性和实时性要求。     

基于图像处理的包装印刷缺陷检测方法

目的为了提高包装印刷品缺陷检测的准确度和适用性,基于图像处理设计一种包装印刷缺陷检测方法。方法采用几个关键步骤包括图像配准、配准区域自动选取、缺陷检测等,并根据包装印刷品图像特征选择配准区域,同时给出一种快速图像配准算法,利用改进差影匹配算法实现缺陷检测;基于DSP和FPGA设计控制系统硬件平台,主要包括控制单元、图像采集与处理单元、成像单元等,并进行实验研究。结果所述方法能够准确识别出细微的刀丝、拉条类缺陷。结论该方法具有较高的可靠性、通用性,可实现包装印刷品缺陷的快速检测。     

基于视觉成像技术的高速包装机模盒内烟支检测的研究

目的 解决机械探针式烟支检测在检测模盒内烟支时存在误检和漏检的问题。方法 设计一种基于视觉成像技术的高速包装机模盒内烟支检测装置，研究其关键技术，包括工业相机、LED光源和控制器的安装，动态图像采集和图像处理算法，PLC控制程序的设计等。结果 该烟支检测装置安装使用后，对模盒内烟支缺陷产品检测的准确率≥98%，误检率≤1%。结论 该烟支检测装置能有效解决机械探针式烟支检测误检和漏检的问题，提高了产品质量，降低了废烟消耗，提高了设备有效作业率，可推广应用于行业内所有包装设备上。     

基于机器视觉的包装品质检测系统设计

目的 为提高包装缺陷检测精度,基于机器视觉设计一种包装品质检测系统。方法 介绍机器视觉系统硬件结构,包括载物平台、工业相机、光室、计算机等。在此基础上,以条烟外包装检测为主要研究对象,详细阐述图像处理算法。首先利用中值滤波消除原始图像噪声;然后基于Canny算子实现图像边缘锐化;最终通过图像配准判断条烟外包装是否存在缺陷。结果 通过实验验证,结果表明该系统对合格样本的识别率可以达到100%;不合格样本的识别率也可以达到98.67%;整体识别率可以达到99.33%。结论 机器视觉系统具有识别精度高、性能稳定等特点;图像处理算法可准确区分条烟是否存在缺陷,具有实际应用价值。     

C800包装机烟包外观检测装置的设计

目的为了解决GDX6S包装机出口至烟包输送通道、烟包存储装置、C800包装机入口处烟包输送、存储、包装等过程中容易导致烟包外观质量缺陷,而且无任何检测装置对其进行有效检测与剔除的问题。方法设计一种新型烟包外观视觉检测装置,通过对其关键技术的研究,综合工业相机及触发脉冲信号的安装,全方位可旋转支架的设计制作与安装,动态图像的采集与处理,PLC控制程序的设计等,实现烟包外观缺陷产品的动态检测及准确剔除。结果该外观视觉检测装置安装后,对烟包外观缺陷产品检测剔除率≥99%,误检率≤0.9%。结论该烟包外观检测装置能提升产品质量,降低物耗,提高设备作业率,降低有缺陷烟包进入下道工序的风险,可推广应用于行业内所有包装机组设备上。     

水松纸在线检测系统设计

目的针对烟支卷制完成后出现的水松纸质量缺陷等问题,结合生产实际要求,设计基于PLC和机器视觉的水松纸在线检测系统。方法采用线阵相机对图像进行采集,工业计算机利用Vision-Pro图像处理软件进行图像处理,采用SIFT改进算法简化图像处理过程,实现对水松纸质量的在线检测,同时完成参数设定、状态监控和数据记录等功能。结果实际运行结果表明,成品烟合格率由92%提高到100%。结论系统运行可靠,可与现有系统配合实现在线监测,减少了原材料的浪费,提高了成品烟支合格率。     

一种药品包装自动检测系统的研究与实现

结合药品包装检测的实际环境,通过应用线性光中心位置的提取、数字图像的中值滤波和自适应模板匹配算法,设计了一套药品包装智能检测系统.该系统能准确对药品包装中出现的装错、漏装、缺损等质量问题进行实时在线检测,以实现药品包装自动化.     

BV包装机条烟外观质量检测装置的设计

目的 解决BV包装机在生产过程中,容易出现条烟外包透明纸丢失、泡皱、粘贴不牢、反条包装、拉线丢失、拉线错位等外观包装缺陷不能被有效检测及剔除等问题。方法 设计一种新型的BV包装机条烟外观视觉检测系统,通过动态图像采集、图像识别与处理、安装支架设计、工业相机安装、触发脉冲采集、剔除装置设计等,实现对外观存在缺陷条烟的自动准确检测剔除。结果 该外观视觉检测系统实施后,外观缺陷条烟检测和剔除率≥99.8%,误剔率≤0.01%。结论 该外观视觉检测系统能提高产品外观质量,降低物耗,为精益生产、精益加工起到了一定的保障作用,该系统和技术可推广应用于行业内的单机包装设备上。     

基于机器视觉的工业炸药包装缺陷在线检测系统

针对炸药药卷包装缺陷,开发了药卷包装缺陷在线视觉检测系统。对系统结构、视觉成像系统进行了描述,并详细介绍了检测方法。采用该方法对获取的药卷图像预处理,并计算药卷轮廓特征,将得到的药卷特征参数与预先建立的药卷标准模板进行对比。试验和应用表明,该方法对3类常见包装缺陷检测准确率不低于96%,所开发的系统可以快速、准确地检测药卷缺陷,能满足实际炸药包装生产过程在线检测的需要。     

军用飞机复合材料构件损伤的检测技术

介绍国内外各种复合材料损伤的无损检测技术,对其应用在军用飞机复合材料构件损伤检测的优势和局限性进行分析和探讨.电阻抗检测方法具有设备小型化、抗干扰能力强、检测速度快、可靠性高等特点,适用于军用飞机复合材料构件损伤的外场原位检测,具有良好应用前景,是复合材料构件外场原位检测研究的重点.     

数字式电脑多频涡流检测仪在电厂中的应用

目的 消除电站热交换管板信号干扰，提高检测信噪比。方法 采用4个独立频率的检测通道，混频处理后干扰信号，结果 在理论分析与实验室大量模拟试验基础上开发出国内第一台数字式电脑多频（四频）涡流检测仪，并在电厂中得到应用。结论 现场使用证明该仪器具有操作方便，检测速度快，抗干扰灵敏度高等特点。     

Automatic recognition of welding defects in real-time radiography

This paper describes a real-time radiography configuration for the automatic inspection of welds. The optimal geometrical imaging configuration is evaluated and discussed in relation to conventional film radiography. For the automatic inspection of X-ray images, a two-step analysis was adopted: a fast search for defective regions, followed by fine identification and location of defects. Two different algorithms, based on the relative irregular behaviour of a defect, were developed for the fast search procedure. The second step, fine identification, can be achieved by a sequential similarity detection algorithm or by a thresholding algorithm. The different methods were applied to various X-ray images of welds and the automatic inspection was evaluated and compared with visual inspection.     

一种基于结构特征的票据印刷号码识别方法

目的为保证含有号码印刷错误的票据不流入社会,研究一种新的用于票据印刷在线检测系统的号码识别方案。方法提出一种基于结构特征的票据号码识别方法。先对票据图像进行采集、灰度化、二值化、去噪、倾斜校正、字符定位、单字符分割及归一化等一系列预处理。建立一种基于结构特征的号码识别分类器,再根据票据中每个号码的结构特征值,对号码进行分类识别。结果实验结果表明,利用文中提出的结构特征方法,票据号码识别率达到99%以上。结论经过对大量实际发票号码的识别测试实验,该方法有较强的抗干扰性,识别算法速度快、精度高。     

Nondestructive identification of multiple flaws using XFEM and a topologically adapting artificial bee colony algorithm

We present a novel algorithm based on the extended finite element method (XFEM) and an enhanced artificial bee colony (EABC) algorithm to detect and quantify multiple flaws in structures. The concept is based on recent work that have shown the excellent synergy between XFEM, used to model the forward problem, and a genetic‐type algorithm to solve an inverse identification problem and converge to the ‘best’ flaw parameters. In this paper, an adaptive algorithm that can detect multiple flaws without any knowledge on the number of flaws beforehand is proposed. The algorithm is based on the introduction of topological variables into the search space, used to adaptively activate/deactivate flaws during run time until convergence is reached. The identification is based on a limited number of strain sensors assumed to be attached to the structure surface boundaries. Each flaw is approximated by a circular void with the following three variables: center coordinates (x_c, y_c) and radius (r_c), within the XFEM framework. In addition, the proposed EABC scheme is improved by a guided‐to‐best solution updating strategy and a local search (LS) operator of the Nelder–Mead simplex type that show fast convergence and superior global/LS abilities compared with the standard ABC or classic genetic algorithms. Several numerical examples, with increasing level of difficulty, are studied in order to evaluate the proposed algorithm. In particular, we consider identification of multiple flaws with unknown a priori information on the number of flaws (which makes the inverse problem harder), the proximity of flaws, flaws having irregular shapes (similar to artificial noise), and the effect of structured/unstructured meshes. The results show that the proposed XFEM–EABC algorithm is able to converge on all test problems and accurately identify flaws. Hence, this methodology is found to be robust and efficient for nondestructive detection and quantification of multiple flaws in structures. Copyright © 2013 John Wiley & Sons, Ltd.     

Referenceless segmentation of flaws in woven fabrics

The automatic segmentation of flaws in woven fabrics is achieved by applying Fourier analysis to the image of the sample under inspection, without considering any reference image. No prior information about the fabric structure or the defect is required. The algorithm is based on the structural feature extraction of the weave repeat from the Fourier transform of the sample image. These features are used to define a set of multiresolution bandpass filters, adapted to the fabric structure, that operate in the Fourier domain. Inverse Fourier transformation, binarization, and merging of the information obtained at different scales lead to the output image that contains flaws segmented from the fabric background. The whole process is fully automatic and can be implemented either optically or electronically. Experimental results are presented and discussed for a variety of fabrics and defects.     

基于现场总线的汽车检测线控制系统研究

介绍了现场总线技术,研究了基于CAN总线的汽车检测线控制系统结构和检测流程,设计了基于PCI的CAN适配卡和基于CAN总线的智能化、自动化的检测设备,分析并设计了控制系统软件。该系统将控制功能延伸到检测设备,实现了系统的全分散控制,提高了系统的灵活性、自治性。检测设备通过总线进行双向、多节点的全数字通信,提高了数据传输的实时性和可靠性。     

用HOUGH变换改进的曲率法识别平面轮廓图元

针对检测精度与检测速度两大指标,提出了用HOUGH变换改进的曲率法平面轮廓图元识别方法。开发了基于邻域值的轮廓点分类算法,采用曲率阈值法筛选轮廓点、投影高度法判别图元属性及分类轮廓点,构建了基于HOUGH变换的直线图元、圆弧图元分割与融合算法。对提出的方法分别进行了特征点检测精度与检测速度对比实验和特征点检测能力测试实验。实验结果表明,提出的方法图元识别准确、检测速度快、通用性好。     

深水钻井隔水导管挠曲方程和固有频率的计算及其研究

从深水钻井应用的隔水导管的使用实际情况出发,应用弹塑性力学中的位移复分原理的Ritz法,考虑了隔水导管所受到的轴向张力T和隔水导管受力产生的倾角α以及由于自重q所产生的自重横向分力qsinα和轴向分力qcosα,建立了较为合理的挠曲方程,并且由此得到隔水导管的固有振动频率的简易计算求解方法。由于该挠曲方程结合了钻具的实际情况,其挠曲方程的约束条件采用的是一端铰支、一端自由的形式,因此其研究更接近实际情况,研究旨在为深水钻井隔水导管的合理使用设计和选择,提供一种更快捷实用的方法,对深水钻井作业有一定的指导意义。     

A multiscale flaw detection algorithm based on XFEM

We present a novel multiscale algorithm for nondestructive detection of multiple flaws in structures, within an inverse problem type setting. The key idea is to apply a two‐step optimization scheme, where first rough flaw locations are quickly determined, and then, fine tuning is applied in these localized subdomains to obtain global convergence to the true flaws. The two‐step framework combines the strengths of heuristic and gradient‐based optimization methods. The first phase employs a discrete‐type optimization in which the optimizer is limited to specific flaw locations and shapes, thus converting a continuous optimization problem in the entire domain into a coarse discrete optimization problem with limited number of choices. To this end, we develop a special algorithm called discrete artificial bee colony. The second phase employs a gradient‐based optimization of the Broyden–Fletcher–Goldfarb–Shanno type on local well‐defined and bounded subdomains determined in the previous phase. A semi‐analytical approach is developed to compute the stiffness derivative associated with the evaluation of objective function gradients. The eXtended FEM (XFEM), with both circular and elliptical void enrichment functions, is used to solve the forward problem and alleviate the costly remeshing of every candidate flaw, in both optimization steps. The multiscale algorithm is tested on several benchmark examples to identify various numbers and types of flaws with arbitrary shapes and sizes (e.g., cracks, voids, and their combination), without knowing the number of flaws beforehand. We study the size effect of the pseudo grids in the first optimization step and consider the effect of modeling error and measurement noise. The results are compared with the previous work that employed a single continuous optimization scheme (XFEM–genetic algorithm and XFEM–artificial bee colony methods). We illustrate that the proposed methodology is robust, yields accurate flaw detection results, and in particular leads to significant improvements in convergence rates compared with the previous work. Copyright © 2014 John Wiley & Sons, Ltd.