基于逆向P-M扩散的医用输液容器组合盖缺陷检测系统  

Defect detection system of medical infusion container combination cover based on reverse P-M diffusion

作　　者：张辉[1,2,3] 师统 何世超 王海洲[2] 

机构地区：[1]长沙理工大学电气与信息工程学院,长沙410012 [2]湖南大学电气与信息工程学院,长沙410082 [3]湖南千山制药机械股份有限公司,长沙410100

出　　处：《电子测量与仪器学报》2015年第5期692-700,共9页Journal of Electronic Measurement and Instrumentation

摘　　要：针对医用输液容器组合盖在生产制造中出现的黑点、毛发、气泡等缺陷问题,设计了基于逆向P-M扩散的医用组合盖缺陷检测系统。首先设计了机械执行、电气控制、图像采集系统,然后采用逆向P-M扩散增强缺陷区域,通过差分后二值化提取缺陷区域,并进行图像滤波,接着利用SVM支持向量机对缺陷与非缺陷区域进行分类,通过交叉验证法自动选择最佳分类参数,解决了医用组合盖表面纹理对缺陷检测的干扰,实现了医用组合盖缺陷检测功能,有效提高了分类器的性能。实验结果表明,该方法要求训练样本少,适用于不同组合盖缺陷类型检测和检测环境,准确率95%以上。In order to solve the defects of medical infusion combined cover appeared in the manufacturing process, such as black spots, hair, bubbles, etc, a new detection system based on reverse diffusion P-M algorithm is designed for detecting medical combined cover. Firstly, the mechanical actuator, the electrical control and the image acquisi- tion system are designed, and then the reverse P-M diffusion algorithm is used to enhance the defect area. This area is extracted by binarization after difference, and the image is filtered. Secondly, SVM algorithm is used to classify the defective and non-defective area automatically, and the best classification parameter is selected by cross validation method. The interference in defect detection of surface texture of medical combined cover is solved, the function of defect detection is realized, and the performance of the classifier is improved. The results show that the method re- quires less training samples, and is suitable for detecting the defect types of combined cover and environmental tes- ting, and its accuracy rate is more than 95%.

关 键 词：医用组合盖 机器视觉 缺陷检测 逆向P-M扩散 支持向量机 

分 类 号：TN911.73[电子电信—通信与信息系统]