印刷电路板无铅焊点假焊的检测

通过分析由三色LED环形结构光源和3-CCD彩色相机获取的印刷电路板(PCB)焊点图像特征,设计了一种PCB无铅焊点假焊的检测方法,以提高自动光学检测系统检测焊点质量的准确率,降低误判.该方法通过对焊点及其元器件的定位,采用重心法检测假焊;根据焊点及元器件的边缘确定焊点区域及元器件区域,采用面积法二次检测假焊;最后,采...     

基于图像特征统计分析的PCB焊点检测方法

提出了一种基于图像特征统计分析的炉后焊点检测方法,以提高在线自动光学检测系统的检测性能和可操作性.提出双阈值的AdaBoost算法用于设计分类器,在训练的同时进行最优特征选择和分类器的增强,实现了焊点图像特征的自动提取和检测参数的自动设定.采用分类和回归树方法将焊点缺陷决策方法优化为一棵二叉决策树,提高了检测速度.实验结果表明,该方法训练速度较快,可以满足实际生产需要.与目前已经实用化的图像对比算法和图像分析算法相比,在保持现有检测速度基本不变的情况下,该方法的检测精度更高.     

基于多分类支持向量机算法的PCB焊点检测研究

针对PCB焊点检测分类的应用需求,采用基于机器视觉的"图像特征+分类器"的技术方案,研究总结了焊点图像特征提取方法,并采用多分类支持向量机算法实现焊点分类;在检测算法具体实现上,综合考虑多类别样本空间均衡、类间离散程度以及分类器数量,在算法性能、分类器训练与分类执行效率3方面取得平衡。实验结果表明,该检测方法对焊点类型的正确识别率达到了97.9%,证明了方法的有效性。     

一种基于混沌优化算法的PCB板元件检测方法

先进电子制造生产中经常要对PCB板元件进行检测与识别,介绍了一种基于图像模板匹配算法的PCB板元件自动快速检测方法.从检测速度和准确度出发,首先提出了一种图像相似性度量参数指标,并提出一种利用并行混沌算法融合单纯形的算法,来优化搜索图像相似性,给出了算法实现的全过程.用实际拍摄的PCB板元件进行性能测试,验证了该优化算法能提高检测速度.     

基于改进SOLOv2的金丝球焊焊点检测方法

为实现对金丝球焊焊点的精确检测,提出了一种基于改进SOLOv2网络的金丝球焊焊点检测方法。该方法以SOLOv2网络为主体框架,设计了一种孪生结构编码器,可以同时输入同轴光图像和低环光图像并准确提取焊点特征。在SOLOv2网络结构中将编码器同一层级的特征图通过跳跃连接建立同轴光图像和低环光图像之间的特征联系,实现了图像特征信息互补以及焊点检测准确率的提高。通过孪生结构隔离开同轴光图像和低环光图像特征提取模块的权重参数,避免编码器权重参数在训练过程中偏向于某一种光照场景,提高了网络的泛化能力。改进后的SOLOv2网络在测试集上的交并比IoU和F1-score值分别提升了0.021 9和0.013 7,并且使用在COCO数据集上的预训练权重来初始化编码器,以加速网络的收敛。相对于语义分割网络,SOLOv2网络也能够有效解决黏连焊点特征提取问题,满足实际检测需求。     

基于机器视觉的PCB裸板缺陷自动检测方法

提出了通过比较标准图像与待测图像差异并分析差异区域边界进行印刷线路板(PCB)缺陷检测与识别的算法。在同一位置采集多幅标准PCB图像并计算其灰度平均值从而得到标准图,将待测PCB图与其进行对比。首先使用限定区域Hough变换快速检测出图像中相互垂直相交的细短标志线,将线段的交点作为特征点并计算其坐标,进而对标准图与待测图进行仿射变换配准,差影计算后,再通过二值化、形态学处理等去除伪缺陷,即可获取缺陷区域位置。在此基础之上,对处理过的差影图进行膨胀处理,通过边界检测获取各个缺陷区域闭合轮廓各点坐标。分析各个轮廓坐标对应阈值分割后的配准待测图中点的像素值,并结合缺陷是缺料缺陷还是多料缺陷识别出缺陷类型。对合格的和有缺陷的PCB图各200幅进行算法测试,检测准确率为98.3%,基本能够稳定检测出常规缺陷。     

基于超声参数化和熵模型的汽车焊点质量识别

汽车焊点的超声波检测过程中,超声回波为非稳态信号,特征不易提取,同时焊点缺陷类型众多,导致汽车焊点质量的自动判定与识别比较困难。因此,提出一种基于超声信号参数化和判别熵模型的汽车焊点质量智能识别方法。通过建立汽车焊点超声回波信号的参数化模型,再基于EM算法思想,提出多回波超声信号的特征参数估计算法。根据提取的超声信号时频特征值,结合判别熵对特征值的有效性进行监督,提取最优特征值子集,最终实现汽车焊点质量类型的智能识别。焊点的实际检测结果验证了方法的有效性和识别的准确性。     

基于YOLOv4及其变体的白车身焊点外观检测

白车身焊点的质量直接影响汽车的牢固度和安全性,而其检测目前以人工目测为主,其检测效率和检测质量都不足以满足汽车产业发展的要求.基于YOLOv4网络模型对白车身焊点的外观分类做深度学习并成功实现了焊点的识别与分类.并针对焊点图像的特点对YOLOv4模型进行了变体开发,提出一种针对小目标的、快速运算的YOLOv4变体.首先收集了大量白车身焊点图像,根据车身焊点外观质量要求对图像进行了分类,输入YOLOv4网络模型作为训练样本,使目标检测准确性指标mAP值达到97.29％.其次,根据焊点小目标、图像特征单一等特点去除了SPP模块和最后一个下采样层,结果发现,改进后的模型不仅可以保证较高的目标检测精度,且运算效率提高了2.9％,同时参数数量减少17.6％.研究证明,基于变体YOLOv4深度学习框架的焊点外观检测方法,不仅可以实现焊点目标的检测及缺陷分类,且参数量更少,计算速度更快,本研究有助于白车身制造智能化水平与焊接质量的提高.     

基于最小二乘拟合法的焊点形状检测

自动光学检查是利用图像处理算法进行电路板故障检测与定位的技术。电路板焊点形状不规则故障，涉及到焊点几何形状的测量。本文提出一种新颖的基于最小二乘拟合法的焊点形状检测算法，对焊点形状轮廓图像进行曲线拟合，确定焊点形状近似的理想圆参数，即圆心和半径，以及焊点形状与理想圆之间的偏差，实验证实了该算法的性能。     

基于改进多尺度分形特征的红外图像弱小目标检测

为提高红外图像弱小目标检测的准确率和实时性,在分析用于红外图像增强的分形参数K相关的多尺度分形特征(MFFK)基础上,提出了一种基于改进多尺度分形特征(IMFFK)的红外图像弱小目标检测算法。首先,将基于地毯覆盖法的分形维数计算公式代入MFFK计算公式,提出了一种改进多尺度分形特征(IMFFK)用于图像增强。其次,对IMFFK特征计算进行简化,采用自适应阈值分割得到感兴趣目标区域,提出了一种具有较高计算效率的红外图像弱小目标检测算法。最后,通过仿真图像分析了主要参数对图像增强和算法耗时的影响,采用红外真实图像进行了算法检测性能测试,并与当前基于局部对比度测度的目标检测算法进行了对比。实验结果表明,提出的算法虽然在一些检测场景具有较多虚警,但能同时适用于弱小目标和较大目标检测,且无论目标为亮目标或暗目标。提出算法对于低分辨率红外图像(320×240)检测接近30 frame/s。提出算法具有较强的适用性,能够检测出红外图像中具有较高局部对比度的目标。     

稀疏重构SAM芯片焊点检测方法研究

提出了基于稀疏表示的声扫描显微镜(Scanningacousticmicroscope,SAM)图像超分辨率重构方法,以解决其空间检测分辨率受超声波频率和穿透深度的限制,原始SAM图像分辨率较低,不利于封装缺陷辨识等问题。通过字典设计训练和稀疏系数α求解获得了重构的高分辨率SAM图像,利用Levenberg-Marquardt算法改进BP神经网络(LM-BP),并用于倒装芯片焊点缺陷识别。与原始图像及双三次插值图像相比,稀疏重构图像的峰值信噪比明显增大,提高了SAM图像质量,减小了芯片焊点的错误识别数目,错误率降至2.76%。试验结果表明稀疏表示的SAM重构算法和LM-BP神经网络训练速度快、识别精度高,可用于高密度半导体封装缺陷的检测及可靠性评估。     

Automated inspection of PCB components using a genetic algorithm template-matching approach

Automated inspection of surface mount PCB boards is a requirement to assure quality and to reduce manufacturing scrap costs and rework. This paper investigates methodologies for locating and identifying multiple objects in images used for surface mount device inspection. One of the main challenges for surface mount device inspection is component placement inspection. Component placement errors such as missing, misaligned or incorrectly rotated components are a major cause of defects and need to be detected before and after the solder reflow process. This paper focuses on automated object-recognition techniques for locating multiple objects using grey-model fitting for producing a generalised template for a set of components. The work uses the normalised cross correlation (NCC) template-matching approach and examines a method for constraining the search space to reduce computational calculations. The search for template positions has been performed exhaustively and by using a genetic algorithm. Experimental results using a typical PCB image are reported.     

印刷电路板焊点的智能检测

为了提高在线自动光学检测系统(AOI)的自动化程度,提出了一种基于增量聚类的智能焊点检测方法.首先,设计了在线智能AOI的系统框架.然后,根据焊点外观进行归纳分类,将关键子区域的面积特征应用于焊盘特征的量化与提取,将每类样本聚类为若干子类从而实现对多批次焊点的检测.最后,提出一种增量聚类算法,在线检测过程中系统可根据人...     

挠性印制电路板焊盘表面缺陷的检测

针对挠性印制电路板(FPC)上的焊盘表面缺陷,提出一种基于图像处理技术的智能检测方法。首先,根据缺陷的表现形式对焊盘缺陷进行归纳与分类,采用最大熵值法量化提取焊盘的颜色特征和面积特征;然后,通过评估灰度共生矩阵(GLCM)对纹理颜色变化特征与纹理结构特征量化的有效性,将其应用于焊盘纹理特征的量化与提取。实验分析显示,缺陷焊盘与非缺陷焊盘在某个或多个特征上存在着明显的差异。基于该特点,建立了BP神经网络,以焊盘的颜色、面积、纹理结构、纹理颜色变化特征值作为神经网络的输入,通过学习大量样本,获取最佳权值参数,最终实现对FPC焊盘表面缺陷的检测,检测准确率高达94.6%,50个焊盘的检测时间为300ms。实验结果表明:本文提出的检测方法不仅能够有效地对焊盘表面缺陷进行识别,而且能够满足在线检测对速度的要求。     

Algorithm of locating PCB components based on colour distribution of solder joints

This paper investigates the methodologies for locating and identifying components on a printed circuit board used for surface mount device inspection. It is the foundation of other inspections, such as solder joint inspections. The proposed approach analyzes the colour distribution patterns of solder joints under three layers of ring-shaped LEDs. All solder joints are recognized and located according to colour distribution patterns and their highlight properties. At the same time, other highlight components, such as markings and via-holes are also recognized and removed. The colour distribution direction of each solder joint is also evaluated. The determined directions offer important clues for the clustering of solder joints. All the solder joints of a component are clustered based on their colour distribution directions and some additional constraints. The protective coating of each component is located based on the position of all clustered solder joints. The experimental results show that the proposed method can recognize most of components effectively.     

内置无铅锡膏印刷机的检测算法研究

针对PCB锡膏印刷中常见的偏移、桥接、少锡、多锡等缺陷,提出了一种可集成于锡膏印刷机的机器视觉锡膏检测算法。该算法通过对锡膏印刷特征的建模、锡膏的定位、区域特征的提取,实现了锡膏印刷缺陷的检测。实验结果表明,该算法能够有效地识别常见的印刷缺陷,并且速度上可满足在线检测要求。     

Fatigue of SMT solder joint including torsional curvature and chip location optimization

The objective is to find whether fatigue of the solder joint or the location of the chip is related to the curvature and torsional curvature when SMT solder joints are under vibration or shock condition. However, the torsional curvature is always omitted. In this paper the relationship of curvature and torsional curvature is argued, and it is found that the quantity of the curvature is closer to that of the torsional curvature and shows that omitting the torsional curvature is not accurate. Also, the fatigue of the solder joint including torsional curvature is obtained. One kind of SMT printed circuit board (PCB) used in aviation and spaceflight has been obtained to test its modes. According to the modal results, the location of the SMT assembly is not very suitable and the reliability of SMT solder joints will decrease, which will make an unreasonable catastrophe appear. Meanwhile, the location optimization including curvature and torsional curvature is considered and four areas are obtained in the PCB to locate the chip for increasing the fatigue of solder joints. In this paper the results obtained by assuming the PCB is simply supported in its four boundary. The other fixed condition will be calculated using a similar method.     

Effectiveness study of an automated 3D laminography x-ray inspection system in a high-volume–low-mix SMT line

The advent of miniature surface-mount components coupled with increasing lead counts poses tremendous challenges during the assembly of printed circuit boards (PCBs). This is especially true in an electronics manufacturing service (EMS) industry, where the focus has been to increase throughput and first-pass yields at minimum expenditure. Hence, robust inspection systems are required to achieve the objectives of inspecting all the solder joints on a PCB assembly without affecting throughput and of accurately and repeatedly detecting solder-joint-related defects. Conventional inspection technologies are limited in their capability to inspect and detect solder-joint-related defects. However, these limitations have been addressed to a considerable extent by automated 3D laminography x-ray inspection systems. This research effort ascertained the effectiveness of an automated 3D laminography x-ray inspection system from a statistical and assembly perspective. The results of this investigation were analyzed and continuous improvement measures were initiated and employed to improve the performance of the automated 3D laminography x-ray inspection system and, subsequently, of the surface-mount technology (SMT) line.