基于标准相关匹配的铸件外观缺陷检测  

Appearance Defect Detection of Castings Based on Standard Correlation Matching

作　　者：肖庆阳 方建儒 张效民 李金钟 丛明[2] XIAO Qing-yang;FANG Jian-ru;ZHANG Xiao-min;LI Jin-zhong;CONG Ming(Dalian Yaming Automotive Parts Co.,LTD.,Dalian Liaoning 116041,China;School of Mechanical Engineering,Dalian University of Technology,Dalian Liaoning 116024,China)

机构地区：[1]大连亚明汽车部件股份有限公司,辽宁大连116041 [2]大连理工大学机械工程学院,辽宁大连116024

出　　处：《组合机床与自动化加工技术》2021年第11期72-74,共3页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：大连市科技重大专项项目(2020ZD13GX003)。

摘　　要：针对汽车发动机铸造铝合金缸体油槽表面缺陷检测的问题,提出基于标准相关匹配的铸件外观缺陷检测算法。该方法使用标准相关匹配算法从CCD获取的图像中提取出ROI(region of interest)区域,然后利用中值滤波去除图像表面噪声,通过形态学开闭运算对所得图像进行处理,进一步去除噪声的同时放大缺陷部位在图中的占比,最后使用轮廓提取算法完成对缺陷的识别与定位。实验结果表明,该算法能够有效地提高铸件外观缺陷检测速度,具有良好的检测精度,对外界光照变化以及铸件表面质量的波动具有良好的鲁棒性。Aiming at the problem of detecting the surface defect of the cast aluminum alloy cylinder block oil groove of the automobile engine,a detection algorithm of the appearance defect of the casting based on the standard correlation matching is proposed.This method uses the standard correlation matching algorithm to extract the ROI(region of interest)area from the image acquired by the CCD,and then uses the median filter to remove the image surface noise,and the resulting image is processed through the morphological opening and closing operation to further remove the noise at the same time enlarge the proportion of defects in the map,and finally use the contour extraction algorithm to complete the identification and positioning of the defects.The experimental results show that the algorithm can effectively improve the detection speed of casting appearance defects,has good detection accuracy,and has good robustness to changes in external light and fluctuations in the surface quality of castings.

关 键 词：铸件外观检测 标准相关匹配 图像轮廓提取 

分 类 号：TH162[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]