基于数学形态学的手机屏缺陷检测算法设计  

Design of Defect Inspection Algorithm for Mobile Screen Based on Mathematical Morphology

作　　者：许祖鑫[1] 毕明德[1] 孙志刚[1] 

机构地区：[1]华中科技大学控制科学与工程系信息所,湖北武汉430074

出　　处：《电气自动化》2013年第3期99-101,105,共4页Electrical Automation

摘　　要：针对手机屏颜色单一、缺陷形状特征明显等特点,提出了一种基于数学形态学的表面缺陷自动检测方法。采用CCD工业相机获取单帧图像进行滤波预处理,然后运用灰度形态学方法进行缺陷分割;并通过提取缺陷的形态特征进行缺陷分类。实验结果表明:算法可靠,能有效地识别缺陷,同时准确地判别缺陷类型。Aiming at the characteristics of mobile screens,such as single color and obvious defect shape feature,an automatic detection method for surface defects based on mathematical morphology is introduced in this paper.In this method,single frame images are taken with a CCD industrial camera to go through the filtering pretreatment,next the defect segmentation by means of gray scale morphological approach,and finally the defect classification according to the extracted morphological features.The results show that the algorithm is reliable,can not only identify detects effectively,but also judge the type of defects accurately.

关 键 词：灰度形态学 手机屏 中值滤波 特征提取 缺陷分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]