一种FPC缺陷视觉检测方法  

A FPC Visual Defect Detection Method

作　　者：孙国栋[1] 杨林杰[1] 张杨[1] 

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430068

出　　处：《自动化仪表》2015年第12期16-18,22,共4页Process Automation Instrumentation

基　　金：国家自然科学基金资助项目(编号:51205115);湖北省教育厅青年教师深入企业计划项目(编号:XD2014112)

摘　　要：针对柔性印制电路(FPC)补强片缺陷人工检测不稳定、效率低的缺点,提出了一种基于颜色空间变换的视觉检测方法。首先通过预处理消除彩色图像噪声,然后根据片体颜色特征,选择特定颜色空间对其进行轮廓提取,并以轮廓为基底设计掩膜,获取相应的感兴趣区域图像;再基于RGB颜色空间的正交基投影理论与HSV颜色空间色调处理算法,分别提取压点、划伤和溢胶缺陷特征;最后根据缺陷几何特征与空间信息对缺陷进行识别和分类。实验表明该方法具有良好的检测效果,适用于FPC高精度检测。Due to the manual detection for reinforcing sheet defects has the disadvantage of low efficiency and instability,a visual detection method based on color space transformation is presented. Firstly,through the preprocessing to eliminate noise of color images,then according to the color feature of sheet body,a specific color space is selected to extract the contour and image mask is designed,so as to get the corresponding image region of interest. Then orthogonal projection theory based on RGB color space and tone processing algorithms based on HSV color space are proposed to extract the features of pressured points,scratches and excess glue defects. Finally,the identification and classification of defects are accomplished according to its geometrical features and spatial information. The experiments have shown the proposed method has good detection effect,is suitable for FPC high precision detection.

关 键 词：FPC 补强片 颜色空间 掩膜 视觉检测 缺陷识别 图像预处理 

分 类 号：TP399[自动化与计算机技术—计算机应用技术] TH6[自动化与计算机技术—计算机科学与技术]