采用机器视觉技术的软胶囊药板质检系统设计  

Design of a Soft Capsule Plates Testing System Using Computer Vision

作　　者：彭彦卿[1] 刘成[1] 陈李清 许国澎 

机构地区：[1]厦门理工学院电气工程与自动化学院,福建厦门361024 [2]国药控股星鲨制药(厦门)有限公司,福建厦门361026

出　　处：《厦门理工学院学报》2016年第3期23-27,共5页Journal of Xiamen University of Technology

基　　金：厦门市科技计划项目(3502Z20141150)

摘　　要：采用机器视觉技术设计软胶囊药板质量检测系统.通过图像采集将被检测软胶囊药板转换成图像信号,图像处理系统根据像素亮度、颜色分布等信息对目标特征进行提取,并综合使用包括聚类算法、形态学分析、图像滤波、边缘检测及模板匹配在内的数字图像处理算法,对软胶囊常见缺陷模式进行识别.系统如果判断药板存在质量问题,则控制现场设备将不合格产品分检出来.实验结果表明:该系统每分钟可检测软胶囊2 400粒,对黑点的检测精度可达0.3 mm,检测准确率达到99.7﹪以上.A soft capsule plates testing system was designed using computer vision. Soft capsule drug plates to be detected were transformed into image signals through the image acquisition and the image processing system got extraction of the target features according to brightness and color distribution of pixel information. Through various techniques employing digital image processing algorithms including clustering algorithm,morphological analysis,image filtering, edge detection and template matching, pattern recognition of soft capsules＇ common defects was then done. Drugs board detected for quality faults was sorted out by local devices. The experimental results show that the system can test 2 400 soft capsules per minute with a testing accuracy of 0. 3 mm for black spots at a 99. 7% accuracy rate and higher.

关 键 词：软胶囊药板 质量检测 系统设计 机器视觉 铝塑泡罩包装机 

分 类 号：TQ460.7[化学工程—制药化工] TP391.41[自动化与计算机技术—计算机应用技术]