摘要
设计了一套基于机器视觉技术的全自动灯检系统,以实现大输液中可见异物的在线实时自动检测。为了有效地解决在线高速实时检测这个关键性问题,将制约系统实时性的弱小目标识别过程分离出来,交由高速DSP芯片进行专门处理。通过理论分析和试验测试证明,设计的基于 TMS320C6416 DSP 的图像处理平台能够很好地满足大输液异物检测系统对处理速度和精度的要求。研制的全自动灯检系统已经在某公司的大输液生产线上试验运行。试验表明,该系统对粒径大于50微米微小异物的识别准确率和速度均高于熟练的灯检工。
To detect the possible foreign and particulate matter in pharmaceutical liquor online, we designed an automatic vision inspection system based on machine vision. In order to effectively deal with the key problem of high-speed real-time on-line detection, we assigned high-speed DSP chip to process the particle detection program independently. It was proved theoretically and experimentally that the advanced TMS320C6416 DSP image-processing platform could meet well the actual requirement of on-line inspection for pharmaceutical liquor in both speed and accuracy. Experimental results show that the system can effective detect foreign particles whose sizes are bigger than 50 micron in liquor. Its detection accuracy and speed is higher than skilled workers are.
出处
《电气自动化》
北大核心
2008年第1期51-53,共3页
Electrical Automation
基金
山东省自然科学基金资助项目(Z2006F05)
关键词
微小异物检测
机器视觉
DSP
输液灯检
particle detection machine vision DSP pharmaceutical liquor inspection