An image tracking system for welded seams using fuzzy logic

Nowadays, automatic welding is an important technology in the shipbuilding industry since it can improve the quality of the welding process. In the past a large volume of manpower was necessary to carry out welding, but, due to recent improvements in welding methods and automation technology, the current trend is towards semi-automatic welding. However, some manpower is still required to guide the semi-automated welding system during welding. Due to the high temperatures involved in the welding process and the need to protect the welded area using gas, operators are not able to monitor and control the process very closely. Human operational error and natural welding environmental factors lead to errors in the weld seam coordinates, which influence the quality of the welding. Rectification of these problems consumes much time and manpower. The aim of this paper is to address this problem by using fuzzy theory coupled with image processing techniques to analyze the welded seam and to send an appropriate control signal to the numerical control (NC) machine. This will reduce the manpower required to supervise and correct the welding process, and will move the welding process towards the goal of fully automatic welding. The proposed welding monitoring and control process is as follows. The initial step is to use CCD equipment to obtain images of the actual weldment and welded seam. Fuzzy theory and edge-operator-detection methods are then used to generate the gray level feature factors and the membership function of the image. This information is input into decision-making logic, which, through linear regression, determines the correct welded seam coordinates. In this study, three different welds are considered: a straight line, a kinked straight line and a curved line. These coordinates are transferred to the control code of the NC machine to correct the welding process. It was found through experimentation that adjusting the lighting environment directly influences the process of image tracking of the welded seams.