| 基于机器视觉的日用瓷表面缺陷检测 |
| |
| 引用本文： | 谢森林,曾辉,董晓庆.基于机器视觉的日用瓷表面缺陷检测[J].韩山师范学院学报,2014(6):43-48. |
| |
| 作者姓名： | 谢森林 曾辉 董晓庆 |
| |
| 作者单位： | 韩山师范学院物理与电子工程系,广东潮州,521041 |
| |
| 摘 要： | 为了检测釉面陶瓷表面的斑点、划痕、崩角等常见缺陷,提出了一种采用同轴光照方式和CCD为图像获取工具的数字图像实时在线检测系统.该系统根据地方企业生产的白釉面陶瓷的特点,通过边界提取和图像形态学闭运算获得图片的ROI（感兴趣区域）,再与经过分离中值滤波算法处理过的ROI进行对比,从而确定陶瓷的缺陷信息.实验表明：该方法不像传统的神经网络和SVM算法需要建模,所以检测速度较快,且准确度高,检测效率超过3个熟练工人,可以满足当地企业生产陶瓷种类多且外形不定的检测要求.
|
| 关 键 词： | 机器视觉 边界提取 闭运算 表面缺陷检测 |
Daily-use Porcelain Surface Defect Detection Based on Machine Vision |
| |
| XIE Sen-lin,ZENG Hui,DONG Xiao-qing.Daily-use Porcelain Surface Defect Detection Based on Machine Vision[J].Journal of Hanshan Teachers College,2014(6):43-48. |
| |
| Authors: | XIE Sen-lin ZENG Hui DONG Xiao-qing |
| |
| Institution: | (Department of Physics and Electronic Engineering, Hanshan Normal University, Chaozhou, Guangdong, 521041 ) |
| |
| Abstract: | In order to detect common defects on the surface of glazed ceramic, such as spots, scratches,cracks, this paper puts forward a way of digital image real-time on-line detection system, which uses coaxialillumination and CCD as the image acquisition tools. According to the characteristics of white glaze ceramic —the product of a local enterprise, the system firstly acquires the ROI（Region of Interest） from the pretreatmentof image, and then uses separation of median filtering algorithm to process the ROI. By comparing the pro-cessed ROI with the original ROI, we can determine ceramic defect information. Experiments show that thismethod provides faster detecting speed and higher accuracy, compared with the traditional neural network andSVM algorithm, which a model needs to be built. Therefore, the method put forward in this paper can better sat-isfy the requirements of local ceramic enterprises, whose products and their appearances are various. |
| |
| Keywords: | machine vision boundary extraction closed operation surface defect detection |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
