中文摘要：

研究一种基于图像质心的焊缝识别新方法。根据焊接熔池视觉图像质心算法及视觉传感器工作原理，推导并建立焊接图像质心与焊接区温度及焊缝中心关系模型。通过焊接温度场的数值化，得到图像质心与焊缝中心关系模型在一定焊接参数条件下的数值解。采用数值仿真分析模型的影响因素，结果表明焊接电流、焊接速度和焊缝宽度对模型曲线影响最为显著。在分析质心与焊缝中心数值解变化趋势的基础上，得到质心偏差与焊缝偏差的经验表达式。在此基础上，对不同焊接参数条件下的模型曲线进行最小二乘拟合，建立质心偏差模型特征参数的样本空间。然后进行多元线性回归分析，确定出经验表达模型的特征参数，从而建立质心偏差关于焊接参数为变量的实用理论模型的函数表达式。焊接工艺试验表明，依据图像质心算法可以较好地识别焊缝中心位置。

英文摘要：

A novel method of seam center detection based on the welding visual image centroid is presented. Based on the weld pool image centroid arithmetic and CCD work principle, a centroid model whose characteristic corresponds with the centroid value, workpieces temperature field and the seam center position is established. The model numerical solution under a certain welding condition can be acquired after the welding temperature field is numerically simulated. The analysis result of model influence factors indicates that the factors such as welding current, traveling velocity and seam width influence the model apparently. Experiment formula is obtained according to the tendency analysis of model solution. Also, the least-square method is employed to build the specimen spatial data according to various welding parameters. Then the particular parameters of model are obtained according to multi-linear regressive analysis. Therefore, the centroid-offset formula with actual welding parameters is acquired. It is shown by actual welding experiments that the welding position can be detected by using the image centroid arithmetic.