胶囊药丸包装的机器视觉与射频识别技术的应用研究

Research on the Application of Machine Vision and RFID Technology in Capsule Pill Packaging

针对胶囊药丸包装品质参数的人工目视检测或接触性检测,存在着作业效率低的问题,研究了一种光学检测药丸包装品质的方法。首先使用二台彩色及灰阶CCD摄相机对待测胶囊药丸分别取二张相片;其次利用相片中色调或是外形特征实现药丸颜色辨识,同时把灰阶影像采用Ostu算法进行二值化,并利用灰阶投影统计图进行字符分割,再采用动量法和互相关比对法,实现药丸使用期限OCR字符辨识;最后利用无线射频识别(RFID)读取射频识别标签的品管内容,与光学检测内容比对,作为品质管制及产品内容的验证,并将品管结果输出为Excel档案。试验数据表明,该系統的辨识成功率达到95%以上,符合产业应用的要求。此影像检测系统可实现自动化生产检测,降低生产成本和提高产品的优良率。

In view of the problem of low efficiency in manual visual or contact inspection of quality parameters of capsule pill packaging,a method of optical detection for the quality of pill packaging was studied.First,two color and gray-scale CCD cameras were used to take two photos of the capsule pill under test.Second,the pill color was distinguished by hue or shape feature of picture.Binarization of grayscale image was realized with Ostu algorithm.And character segmentation was obtained using grayscale projection statistics.Pill life identification was realized with OCR character by the momentum method and cross correlation comparison method.Finally,the quality control content of RFID tag was read using RFID,which was used as the verification of quality control and product content by comparing with visual inspection content,and the quality control results were outputted as excel file.The experimental result shows that the identification success rate of the system was over 95%,which meets industrial application requirements.The studied image detection system can realize automatic production detection,reduce production cost and improve the excellent rate of products.