论文部分内容阅读
本文介绍了一种应用在螺焊管焊缝超声波无损探伤自动线上，跟踪运动焊缝横向偏移的视觉伺服系统．该系统以单ＣＰＵ运行的软件进行实时图象识别与系统控制．文中提出了一种在线建立特征模型的匹配识别策略，使系统能够很好地适应引起焊缝视觉特征变化的工作环境、条件的变化，准确识别不同特征的焊缝，同时具有较好的识别快速性．在实际检测线上，系统表现了良好的跟踪特性．
This paper introduces a visual servo system which is used in the ultrasonic non-destructive flaw detection of spiral welded pipe welds to track the lateral displacement of the moving weld seam. The system uses a single CPU running software for real-time image recognition and system control. In this paper, a matching identification strategy of feature model is established online to make the system adapt well to the change of working environment and conditions that cause the change of the visual characteristics of the weld, to accurately identify the weld with different features, and to have better recognition speed Sex. In the actual test line, the system showed good tracking characteristics.