基于虚拟仪器技术与机器视觉的玻璃瓶质量检测

Quality Inspection of Glass Bottle Based on Virtual Technology and Machine Vision

玻璃瓶因具有透明度高、气密性好、外形美观、价格低廉和可回收的特点被广泛应用,但生产过程中存在崩口、划痕、含气泡、杂物和污垢等缺陷。为确保玻璃瓶质量,作者设计了一套基于虚拟技术和机器视觉的玻璃瓶缺陷检测系统。利用几何圆的特征确定圆心位置,采用Sobel算子对瓶身缺陷进行边缘检测,基于虚拟技术vision的圆度拟合函数及图像量化函数分别进行瓶口和瓶身的缺陷分析。实验结果表明该系统能够定量检测缺陷的大小,定性分析缺陷的种类,检测速度为36ms,测试精度为98%。该系统操作简单,能实现快速、准确的检测,有效降低了生产成本,满足了实际生产的需要。

Glass bottles are widely used because of their high transparency, good air tightness, beautiful appearance, low price and recyclable characteristics. However, there are some defects in the production process, such as cracks, scratches, bubbles, inclusions and dirt.In order to ensure its quality, the author designed a set of glass bottle defect detection system based on virtual technology and machine vision. The center position of the circle was determined by the feature of geometric circle. Sobel operator was used to detect the edge of bottle body defects. The roundness fitting function and image quantization function of virtual technology Vision were used to analyze the bottle mouth and bottle body defects respectively. Experimental results show that the system can quantitatively determine the size of defects, qualitative analysis of the type of defects, detection speed up to 36ms, testing accuracy up to 98%. The system is simple to operate, to achieve accurate and rapid detection, to effectively reduce the production cost and meet the actual production needs.