摘要
针对传统印刷品缺陷检测流程中出现的轮廓误检问题做出了改进,首先利用Sobel算子对模板图像进行边缘提取,然后将边缘图像经过数学形态法处理后掩盖在差影图像上,最后再次利用基于数学形态法的开运算对图像进行全局处理以消除轮廓伪影和噪声。实验表明,此方法可以有效消除因纸张双重反射效应和机械振动等原因造成的轮廓伪影,提高检测过程的准确率。
In view of the problem of edge effect, this paper puts forward some new print defect detection methods which are advantageous over the traditional ones. Sobel operator is first used to extract edge and processing edge by morphological method, with the subtraction image by the edge covered. The morphological method is used again to resist the shadow of contour and image noise. Experimental results show that this method could effectively eliminate the false detections due to the double reflections of paper and the mechanical vibration, with the accuracy of the de- tection process improved.
作者
陈恺煊
刘昕
CHEN Kaixuan;LIU Xin(School of Printing, Packaging Engineering and Digital Media Technology, Xi＇an University of Technology, Xi＇an 710048, China)
出处
《西安理工大学学报》
北大核心
2017年第4期486-491,共6页
Journal of Xi'an University of Technology
关键词
缺陷检测
机器视觉
边缘检测
defect detection
machine vision
edge detection
