The joining process of large diameter pipes through submerged arc welding is a multiple pass process and is very time consuming. If the process could be automated, a large amount of man hours could be removed from the process while increasing the level of accuracy and quality of the welding seam. The use of mechanical and automated systems has been previously implemented to varying degrees of success throughout the years to achieve this. However, as the weld becomes more complex, (for example where the weld area is covered in flux or when welding a cap) these basic mechanical and visual based systems tend to losequality and accuracy. To combat this an advanced visual based system needs to be implemented that can actively scan the area to be welded and control the torch position, and welding parameters.