Abstract: Abstract: A seam tracking algorithm based on multi-information sensor of vision and arc is proposed. The algorithm is applied to the automatic control process of MIG pulse welding quality. This paper captures different description information for the effective feature extraction and transmission with the vision sensor and the arc sensor, and can be used in multi-sensor information fusion algorithm for seam tracking. The vision sensor obtains image information by industrial CCD for the lateral deviation control of the weld torch, the arc sensor obtains current information by data acquisition card for the height deviation control of the weld torch, The two kinds of sensor information in a complementary way to integrate for the lateral and height corrective control of the welding process. Experimental results show that the proposed algorithm can improve the welding quality, and thus verify the algorithm efficiency and rationality.

Key words: MIG pulse welding, vision sensor, arc sensor, multi-sensor information fusion algorithm

摘要： 随着焊接传感技术以及信号处理技术的快速发展，人们越来越注重对焊接过程质量控制的研究，而多传感器信息融合技术是焊接质量控制的一种重要的方法,采用该技术实现熔化极气体保护焊(MIG)脉冲焊接质量的自动控制。该技术采用了视觉传感器和电弧传感器将采集的不同的描述信息进行了有效的特征提取和传输，并运用多传感器信息融合算法进行焊缝的跟踪。视觉传感器利用工业电荷耦合元件(CCD)获取图像信息控制焊炬的横向偏差信息，电弧传感器利用数据采集卡获取电流信息控制焊炬的高度互补偏差信息和横向的冗余信息，将两种传感器得到的冗余信息和互补信息在特征层下进行融合实现焊接过程横向和高度的纠偏控制。冗余信息的融合可以实现视觉传感的图像去噪，而互补信息的融合可以进一步提高焊缝跟踪的精度。实验结果表明，所提算法能够较好地提高焊接的质量，从而也验证了算法的有效性和合理性。

关键词: MIG脉冲焊, 视觉传感器, 电弧传感器, 多传感器信息融合