基于机器视觉的玻璃瓶在线检测系统研究与开发

Research and Development of Glass Bottle Online Detection System Based on Machine Vision

玻璃瓶作为酒饮料的主要包装容器,其产品(特别是瓶口)质量直接影响酒的品质与安全。根据透明玻璃瓶在线检测中缺陷信息弱、生产线抖动等特点,研究与开发了一种基于机器视觉的在线检测系统,应用最大熵法实现瓶口图像的分割以及改进的Hough变换算法实现瓶口图像的定位。实际运行表明,系统动态图像采集效果良好,且图像处理算法快速准确,满足当今高速酒饮料灌装在线检测的要求。

Glass bottles play a very important role in maintaining the quality and safety of liquor. As the main part of the liquor bottle, bottle mouth is even more critical. The goal of this work was to develop a vision - based online detection system that can solve the problems of weak defect infor- mation and vibration of product line. We used the maximum entropy algorithm to fulfill the bottlemouth image segmentation and a new improved Hough transformation way to achieve the location of the bottle mouth. These techniques were worked out successfully, which are suitable and fast enough to be included in this system and satisfied the ne- west demands of high - speed online liquor bottle detection system.