针对烟标(烟品包装)印刷质量缺陷检测自动化的需要,设计了一种烟标在线检测系统,该系统可以高效地检测烟标的印刷缺陷,并对缺陷数据进行分析处理.系统的关键技术包括利用几何学方法实现图像的快速匹配,通过Soble边缘检测算法消除图像求差后的边缘效应,用区域膨胀、腐蚀方法消除纹理噪声等.实验表明,该系统在检测效率和精度方面能够满足实际应用需要.
According to the requirement of automation of cigarette label (cigarette packaging) printing defect detection, this paper design a online detection system, which can detect printing defects of cigarette label effectively, analyses and process the defect data. The key technologies include using geometric methods to achieve fast image matching, applying soble edge detection algorithm to eliminate the edge effect after image subtractive, utilizing dilation and erosion of region to eliminate texture noise. The experiments show that the system in the detection efficiency and accuracy to meet the needs of practical application.