本文研究设计了一种基于机器视觉的异型瓶药液中可见异物的智能自动化检测装置,在研究了传统的差分异物提取方法的基础上,创造性的提出了基于区域分割和区块剔除的异物检测逆向思维算法。最后,以125mL中国劲酒样品为对象进行在线检测测试,实验表明,本文所设计的检测系统能很好的满足生产线的实时性和准确性要求,同时,基于区域生长和区块剔除的视觉检测算法克服了传统差分法在跟踪拍照方式下难以克服的跟踪误差和抖动等一系列问题。
This paper proposes the design of an intelligent visual inspector for visible foreign substance in special shaped bottled medical liquor based on synchronous photographing. Through studying traditional differential foreign substance inspection method, an inspection algorithm based on region-dependent segmentation and region-dependent elimination is suggested creatively. At last, 125 mL Chinese King Wine is used as the object to carry out on-line test. Experiment results show that the proposed inspection system meets the requirements of real-time characteristic and accuracy in production line. At the same time, the algorithm overcomes the difficulties of tracking error, mechanical shaking and so on in tracking photography manner of traditional differential foreign substance inspection method.