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。 Soft capsule quality testing is a weakness in the development of soft capsule industry, the detection of soft capsule size is the main indicator of quality testing. The method of machine vision was used to detect the capsule images. First, the capsule images were obtained, the images were automatically segmented by ostu, the morphological filtering was used to eliminate the noise, the area was marked, the shape of the effective area was extracted, and the shape parameters and weights Relationship, find the weight of the capsule, to achieve the capsule sorting. The study developed a robotic vision-based automated soft capsule sorting system that detects unqualified capsules on the production line in real time and rejects them through knockout mechanisms. The qualified product accuracy rate was 95.6%, the unqualified product accuracy rate was 90.2%, the overall accuracy rate was 94.1% and the daily output reached more than 200,000.