采用机器视觉的铝压铸件表面缺陷检测  

Surface Defect Detection of Aluminum Die Casting Using Machine Vision

作　　者：郑晓玲[1,2] 刘斌 

机构地区：[1]数字化视觉测量厦门市重点实验室,福建厦门361021 [2]黎明职业大学机电及自动化学院,福建泉州362000

出　　处：《华侨大学学报（自然科学版）》2016年第2期139-144,共6页Journal of Huaqiao University(Natural Science)

基　　金：国家自然科学基金面上资助项目(51175191)

摘　　要：对铝压铸件常见的缺陷形态进行分析,提出适用于铝压铸件表面缺陷的检测算法.首先,采用阈值分割与形态学相结合的方法分割出可疑区域;然后,根据基于面积、亮度均值、亮度均值差、灰度曲线分析的4个剔除原则,剔除伪缺陷的干扰.试验结果表明:文中的检测方法具有低成本、高精度、可操作性强等优点,能有效提高铝压铸件生产过程中的检测效率.The common defects of aluminum die casting were analyzed in this paper,then a detection algorithm suitable for aluminum die casting surface defect was put forward.First,by using the combined method of threshold segmentation and suspicious area.Then,based on four removing principles：the area,the mean brightness,the mean brightness difference,gray curve analysis,false defects interference was eliminated.The experimental results showed that the detection method had the advantages of low cost,high precision,strong operability,and could effectively improve the detection efficiency in the process of aluminum die castings production.

关 键 词：铝压铸件 缺陷检测 伪缺陷 机器视觉 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]