摘要
为了解决换热器、冷凝器、高压加热器等工业器件中管板焊接自动化程度低、工作效率低下的问题,文中建立了一套基于被动视觉的全自动管板焊接系统.在此基础上,对视觉传感下管板焊接过程中焊接轨迹的检测与提取进行了研究,并针对管板焊接工件非均匀打磨情况下的二值化图像特征,提出了通过提取圆管内圆特征结合管壁厚度代替直接提取焊接轨迹的方法.并提出了一种多次最小二乘法迭代去除背景噪点的方法,实现了圆管内圆边缘的准确提取.通过对比人工测算圆管内圆圆心位置以及半径大小对该方法的提取精度进行了误差分析.结果表明,x轴方向和y轴方向的圆心位置偏差均为-3 pixel,半径相对误差为1.49%.
In order to solve the problem of low automation and inefficiency in the tube-sheet welding of industrial components such as heat exchanger,condenser,high-pressure heater and so on,a set of passive vision based automatic tube-sheet welding system was established. And on this basis,the detection and extraction of welding trajectory were studied in tube-sheet welding by the visual sensing. The direct extraction method of welding trajectory was replaced by the method of combining with the tube thickness and the dimensions of the inside tube circle,through the study of the characteristics of the binary image to non-uniform polished tube-sheet weldment. A multiple iterationleast squares method was also proposed in this paper to remove the background noise. And the edge of the inside tube circle was accurately extracted by using the method. The error analysis of extraction accuracy to the method of this paper was done by comparing with manual measuring result. The result showed that the deviation of the inside circle center position was-3 pixel and-3 pixel in x and y directions,and relative error of the radius was about 1. 49%.
出处
《焊接学报》
EI
CAS
CSCD
北大核心
2017年第10期117-120,共4页
Transactions of The China Welding Institution
关键词
视觉传感
管板焊接
图像处理
visual sensing
tube-sheet welding
image processing
