基于K邻域链码拐点的胶囊端面缺陷检测算法  

Capsule end face defect detection algorithm based on K-neighborhood chain code inflection point

作　　者：刘孝星 吴哲[2] 郑力新[2] 周凯汀[1] 

机构地区：[1]华侨大学信息科学与工程学院,福建厦门361021 [2]华侨大学工学院,福建泉州362021

出　　处：《微型机与应用》2015年第24期53-55,59,共4页Microcomputer & Its Applications

基　　金：福建省科技厅项目(2013H2002)

摘　　要：针对目前空心胶囊端面缺陷图像采集困难等问题,提出了一种特殊照明的方式对胶囊端部进行成像。将图像锐化、改进后的局部自适应阈值、轮廓提取用于胶囊端面图像的预处理,并针对胶囊端面缺陷漏检率高、检测效果差等问题,提出了一种基于K邻域链码的拐点检测算法,统计出胶囊端面缺陷图像中的拐点个数并做出缺陷判别。实验结果 证明,该检测算法实时性好,对常见的5种颜色的胶囊漏检率和误检率均控制在2%~9%。In view of difficulty in images acquisition of vacant capsules end face defects, a special way lighting mode of capsules end face defects imaging was proposed. Firstly image sharpening, improved local adaptive threshold, and contour extraction were used for pre-processing. And according to problems of high missing rate and unsatisfactory detection of capsules end face defects, K-neighborhood chain code for inflection point detection algorithm was proposed to calculate out the number of inflection point in capsule end face defects in images and realize defect detection. The experimental results show that the detection algorithm is of good real-time performance, and the missing rate and error rate of common five color capsules is 2%-9%.

关 键 词：K邻域链码 空心胶囊 端面缺陷检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]