手机屏幕图像缺陷检测方法的研究
合集下载
相关主题
- 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。
关键词:缺陷检测;ROI 识别;特征提取;手机液晶屏
I
安徽大学硕士学位论文
手机屏幕图像缺陷检测方法的研究
Abstract
With the rapid progress of technology and industry, LCD screen electronic products have become an integral part of our lives and production. Reality shows a higher and higher request on the quality of LCD screen of these products, while the current level of technology can not avoid sorts of defects. So, the first problem ,which LCD manufactures need to face, is how to identify and solve the defects of the LCD screen rapidly and accurately. Current image defects of LCD products mainly rely on manual inspection that neither meet the accuracy detection of defects in LCD screen, nor guarantee the stability of test results. This thesis is precisely solves this problem as a starting point by using the mobile LCD screen as object of study and combining digital image processing, pattern recognition and computer technology, and in view of the common image defects, referring to the requirements of industrial production detection algorithm on efficiency and accuracy. This thesis have researched and designed an efficient detection algorithm. Simulation experiment results demonstrated my algorithm is efficiency and accurate on the image defects detection. First to standardize design of the mobile transmission system, and equip with acquisition equipments of appropriate models of high-speed image acquisition card, monitoring camera and computer equipment, complete the preparations of the hardware conditions. In the detection process, use surveillance camera to get the mobile image at first. And the images data are captured from the buffer of image acquisition card quickly by directshow technology. Then the weighted-averaged frame of bad frames could reduce the bad effect of the bad frames which are brought about by the harsh environment of image acquisition. In the image preprocessing stage, the noise will be removed by Gaussian pyramid sampling. Dynamic threshold value, getting from each RGB 3-channels by recursive iteration, will be helpful to detecting the screen rectangle by identifying and extracting shape feature of the image, and then using image two-dimensional geometric transformation to auto-correct the mobile
II
安徽大学硕士学位论文
Abstract
position, extracting ROI. There is the end of image preprocessing. Finally, it detects the contours of defect using the Canny algorithm, and combine with Douglas-Peucker algorithm and Freeman chain code. It extracts the information of defect, and then testing the image defects of mobile screen: the number of bad pixels, geometric distortion, chromatic aberration. This algorithm is stable and efficient. It can be widely used in LCD products for image defect detection relying on the relevant national standards, and it is worth promoting.
安徽大学 硕士学位论文 手机屏幕图像缺陷检测方法的研究 姓名:刘波 申请学位级别:硕士 专业:信号与信息处理 指导教师:李新华 2010-05
安徽大学硕士学位论文
摘
要
摘
要
随着科技与工业水平的突飞猛进, 液晶屏类电子产品逐渐成为我们生活和生 产中不可或缺的要素。现实对电子产品的液晶屏幕的品质要求日益提高,而当前 的科技水平还无法避免各种各样的缺陷。那么,如何快速、准确地检测、识别液 晶屏幕的缺陷将是这类产品生产企业首先必须面对和解决的问题。 当前液晶屏类 产品图像缺陷主要依靠人工检测, 这种检测手段不能满足工业生产对液晶屏幕缺 陷检测准确性的要求,也不能保证检测结果的稳定性。 本课题正是以解决这一问题作为出发点,以手机液晶屏幕为研究对象,结合 利用数字图像处理、模式识别以及计算机技术,针对液晶屏幕的常见图像缺陷, 参照工业生产中对检测算法高效性、准确性的要求,研究并设计了一种高效的检 测方案。仿真实验的结果证实了本检测方案能够实现对图像缺陷的自动、高效、 准确的检测。 先对手机传送系统进行标准化设计,并配备合适型号的高速图像采集卡、监 控摄像头以及计算机等采集设备,完成硬件条件的准备工作。检测过程中,首先 用监控摄像头采集手机图像。再利用 directshow 技术从图像采集卡缓存区快速 获取图像数据;其次,将多帧图像加权平均处理,用以剔除恶劣的图片采集环境 造成的坏帧影响。在图像预处理阶段,通过高斯金字塔采样去噪及分别从 RGB 三 通道递归迭代获得自适应分割阈值;识别并提取图片中的形状特征,获得屏幕矩 形外框信息;运用图像二维几何变换自动校正手机姿势,继而提取 ROI,完成图 像的预处理;最后,利用 Canny 算法检测缺陷轮廓,结合 Douglas-Peucker 算法 与弗里曼链码提取缺陷信息,最终检测手机屏幕图像缺陷:坏点数目,几何失真 度,色差。算法稳定、高效,依托相关国家标准,可广泛应用于液晶屏类产品的 图像缺陷检测,具有一定的推广价值。
图 3-3 RGB 颜色模型示意图 ...........................................................................24 图 3-4 图 3-5 图 3-6 图 3-7 图 3-8 图 3-9 图 3-10 原点在结构元素的外围情况下的腐蚀运算 ........................................26 原点在结构元素的内部情况下的腐蚀运算 ........................................27 原点在结构元素的外围情况下的膨胀运算 ........................................27 原点在结构元素的内部情况下的膨胀运算 ........................................28 图像金字塔示意图 ................................................................................31 高斯金字塔建立流程图 ........................................................................32 高斯金字塔下采样图像 ......................................................................33
Keywords:Defect detection;ROI identification;Feature extraction;Mobile LCD screen
III
安徽1 图 1-2 图 1-3 图 2-1 图 2-2 图 2-3 图 2-4 图 2-5 图 2-6 图 3-1 图 3-2 电脑屏幕检测标准 GB/T9813-2000 对坏点的定义 ..............................5 色差对比示意图 ......................................................................................6 电磁波谱中可见光谱的范围 ..................................................................8 视频采集所需的相关硬件设备 ............................................................12 手机生产线示意图 ................................................................................12 DirectShow 视频采集系统结构图 ........................................................14 手机屏幕中心经过二维几何变换移到标准位置 ................................18 产品线上手机放置情况 ........................................................................18 手机屏幕的图像缺陷检测方案设计流程图 ........................................21 坏帧图像 ................................................................................................23 多帧图像加权处理后的图像 ................................................................23