针对18000瓶/小时1-20mL剂量的安瓿水针剂的在线自动化异物检测需求,运用机器视觉技术,设计并开发了安瓿高速自动灯检平台;对系统的机械部分、电气部分及光学成像系统设计进行了分析,在多次试验中我们确定了转速及急停至拍照的时间间隔重要参数,且提出了一种快速划定安瓿检测区域的算法,并针对运动的弱小异物特征提出了一种有效的识别算法,识别算法采用Visual C++编程并在检测平台上实现,实际运行效果验证了算法的有效性和实用性,其精度和速度都达到了生产线上的要求,该研究验证了机器视觉能有效地运用于安瓿的在线检测.%Aiming at the on-line automatic inspection of different doses of ampoule with 18000 bottles/h, an automatic and high-speed inspector platform was developed in the light of the machine-vision technology. The mechanical and electrical units as well as its optical imaging systems were designed. The spindle speed and the interval from instant stop to take pictures are determined in many experiments. A fast algorithm for the delineation of the region detection was proposed, and for the movement characteristics of small objects, an efficient algorithm to identify small objects was proposed. Under the Visual C+ + programming environment, experiments on the platform have proved the effectiveness and value of proposed algorithms, its precision and speed have met the requirements of the production line. Our research substantiates that it is feasible and efficient to machine-vision technology into the automatic on-line inspection of ampoule without the intervention of human beings.
展开▼
