基于最小生成树和图像矩的陶瓷过滤器表面缺陷检测  

Surface Defect Detection of Ceramic Filters Based on Minimum Spanning Tree and Image Moments

作　　者：周静 刘旭[1,2,3] 董子昊 李金屏 ZHOU Jing;LIU Xu;DONG Zihao;LI Jinping(School of Information Science and Engineering,University of Jinan,Jinan 250022,Shandong,China;Shandong Provincial Key Laboratory of Network Based Intelligent Computing,University of Jinan,Jinan 250022,Shandong,China;Shandong College and University Key Laboratory of Information Processing and Cognitive Computing in 13th Five-year,University of Jinan,Jinan 250022,Shandong,China)

机构地区：[1]济南大学信息科学与工程学院,山东济南250022 [2]济南大学山东省网络环境智能计算技术重点实验室,山东济南250022 [3]济南大学山东省“十三五”高校信息处理与认知计算重点实验室,山东济南250022

出　　处：《济南大学学报（自然科学版）》2022年第4期433-439,451,共8页Journal of University of Jinan(Science and Technology)

基　　金：山东省重点研发计划项目(2017CXGC0810);山东省高等学校科学技术计划项目(J18KA346,J18KA371)。

摘　　要：针对人工检测陶瓷过滤器堵孔、裂缝缺陷效率低、误检率高的问题,提出一种基于最小生成树和图像矩的缺陷检测算法;对输入缺陷图像进行灰度化和去除噪声处理,利用阈值分割方法对图像进行二值化,根据陶瓷过滤器表面孔洞空间分布及面积变化,利用滑动窗口遍历图像,结合最小生成树与图像占空比检测堵孔缺陷;根据过滤器裂缝的灰度和形状特征,采用基于图像矩的等价椭圆的方法检测裂缝缺陷。结果表明,所提出的算法能够有效地检测出陶瓷过滤器堵孔、裂缝缺陷,自建数据集的检测准确率达到95%以上。Aiming at the problem of low efficiency and high error detection rate in manual detection of plugging hole and crack defects in ceramic filters,a defect detection algorithm based on minimum spanning tree and image moments was proposed.The input defect images were processed to be grayed and denoised,and threshold segmentation method was used for binarization of the images.According to spatial distribution and area change of holes on ceramic filter surface,the plugging hole defects was detected by using sliding window to traverse the images as well as combining minimum spanning tree and image duty cycle.Equivalent ellipse method based on image moments was used to detect the crack defects according to gray scale and shape features of filter cracks.The results show that the proposed algorithm can effectively detect plugging hole and crack defects,and the detection accuracy of self-built data sets is greater than 95%.

关 键 词：陶瓷过滤器 缺陷检测 最小生成树 占空比 图像矩 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]