作　　者： (黄静文）; (金晅宏）;

机构地区： 上海理工大学光电信息与计算机工程学院,上海200093

出　　处： 《软件导刊》 2017年第9期185-187,共3页

摘　　要： 设计了一种图像缺陷检测算法,该算法将标准显示图像作为模板,将待测面板图像与模板图像进行比对,确定缺陷大小、位置。首先采用最小二乘法获取屏幕边界所在直线,用所得的4条直线作为角点,根据透视投影原理利用得到的4个角点进行图像几何校正,将校正后的模板和待测图像配准,最后将校正后的图像进行形态学处理,突出缺陷后将其与模板进行差模运算获取缺陷。其中,模板图像为标准显示图像经过与待测面板相同的校正、形态学处理后得到的图像。经试验检测,这种方法适用于多种显示模式,可以有效地检测出面板的显示缺陷。 An image defects detection algorithm is designed. then, compare the panel images with the template images to The algorithm collected standarddisplay images as a template and determine the defect size, location. At first, the least square meth od was used to detect straight lines in order to get the image boundary, then we used the four corners of the image to correct the image, ensure the basic alignment template and the image under test. At last ,we adjusted image morphology processing, high- lighted the defect after comparing with the template for defects. The template image as a standard display images was processed as the panel under test after correcting and morphology processing. Experiments prove that this method is suitable for a variety of display mode and it can effectively detect the display detection.

关 键 词： 缺陷检测 图像校正 点阵面板 最小二乘法 面板