基于机器视觉的医用塑料瓶瓶口检测系统研究

Design and research of bottle-mouth detection system for medical plastic bottles based on machine vision

医用塑料瓶在生产过程中瓶口处会产生残次品,如豁口、瓶口过薄等。此类瑕疵药瓶,无法封口,影响药品密封性,危害药品使用者健康。目前众多企业仍然依靠人工检测,因瑕疵点并不明显,人工误检率高。为此设计了一套基于机器视觉的医用塑料瓶瓶口检测系统。一台高速工业相机,在流水线指定工位通过硬触发,由上至下,对塑料瓶顶部取相,根据图像特点设计算法,完成定位,分割,识别瑕疵点。经过实际生产测试,验证了此系统在精度和速度方面优于人工检测方式,满足企业的检测要求。

Pharmaceutical plastic bottles will produce flaws at the mouth of the bottle during the production process, such as nicks and thinner bottle-mouth. This type of medicine bottle cannot be sealed, affecting the tightness of medicines and endangering the health of drug users. At present, many companies still rely on manual inspections. Because the defects are not obvious, the rate of human error detection is high. For this reason, a set of medicine-based plastic bottle and mouth detection system based on machine vision is designed. A high-speed industrial camera, through the hard trigger in the designated station of the assembly line, takes the phase on the top of the plastic bottle, designs the algorithm according to the characteristics of the image, and the vision system completes positioning, segmentation and detection, Through actual production tests, it is verified that this system is superior to manual detection in terms of accuracy and speed, and meets the inspection requirements of enterprises.