基于机器视觉的酒液智能检测系统研究与开发

Research and Development on Intelligent Inspection System for Wine Based on Machine Vision

设计了一种基于机器视觉的酒液智能检测系统,主要用于酒液中杂质的在线检测。设计了系统的直线式检测机械结构和翻转机械手以及基于工控机和可编程控制器的控制系统,通过特定的给光方式和高速摄像机,获取酒液序列图像,采用亚像素边缘定位方法确定检测区域,提出了在检测区域里采用累积差分和基于空间最小交叉熵的阈值分割相结合的方法获取运动目标,根据运动轨迹判断目标是否为杂质,并对提取的杂质进行测量,计算是否在允许范围内。大量实验验证了检测算法的有效性和装置的可行性。

A machine--vision--based intelligent inspection system was presented. The linear con- vection mechanical structure and electric control system based on industrial computer and programma- ble logical controller were illustrated in detail. A method using high--speed cameras was applied for obtaining the image sequences. The sub--pixel edge location method was used for confirming the in- spection region. The methods of cumulative difference and the minimum space of the cross--entropy threshold segmentation were used to acquire the moving targets. The foreign substances were judged by the moving trajectory. Finally,according to the metrical data of foreign substances,a robot gave an judgment that the product was qualified or unqualified. A great deal of experiments demonstrate that the detection algorithm is effective and the device is feasible.