摘要
软胶囊的质量检测是软胶囊行业发展的一个软肋,软胶囊大小的检测是质量检测中的主要指标。研究采用机器视觉的方法进行检测,首先获取胶囊图像,用ostu自动取阈法分割图像,采用形态滤波消除噪声,进行区域标记,对有效区域进行形状特征的提取,建立形状参数与重量之间的关系,求出胶囊的重量,实现对胶囊的分选。该研究开发了一套基于机器视觉的软胶囊自动分选系统,能实时检测出生产流水线上的不合格胶囊,并通过剔除机构将其剔除。合格品准确率为95.6%,不合格品准确率为90.2%,总体准确率为94.1%,日产量达到20万粒以上。
The quality inspect of soft capsules is a weakness in soft capsule development of the industry and the size detection of soft capsules is key indicators of the quality detection. Because of many shortcomings in the artificial separation, this study used machine vision approach to the test, first of all, access to capsule images, automatic threshold by ostu to segmentation image, using morphological filter to eliminate noise, regional markings, shape the effective region feature extraction, the establish the relationship between the shape parameters and the weight, obtain the capsule weight, achieve the separation of capsule. This study developed a set of machine vision-based automatic sorting system on soft capsule capable of real-time detection of the production pipeline failure capsule, and will be removed through its omissions. Conforming accuracy was 95.6%, unqualified goods for 90.2% accuracy rate, the overall accuracy was 94.1%, production per day is more than two hundred thousand.
出处
《食品工业》
北大核心
2008年第5期64-66,共3页
The Food Industry
基金
国家863科技项目(2002AA248051)
江苏省自然基金重点项目(BK2006707-1)
关键词
软胶囊
机器视觉
在线分选
soft capsules
machine vision
on-line grading
