An image acquisition system for real-time seam tracking

Abstract

Gas metal arc welding (GMAW) process is one of the most widespread welding processes used in industries for their excellent quality, reliability, productivity, and cost-effectiveness. To develop an automatic GMAW system, vision capability in the system is a necessary component supplying real-time information about weld pool and seam tracking. In this research work, an automatic seam tracking system is presented, where the automatic tracking of welding path and torch positioning are performed by a newly developed image acquisition system. The system aims to add a vision capability to the GMAW system. A CCD camera is configured with a welding torch to acquire real-time images. The acquired images are processed through newly developed software for real-time detection of welding seam location and characteristics. The software encapsulates the acquired image input facility, image filtering technique, strategy to measure the seam gap, strategy to position torch at the starting point of welding, user interface for automatic guide, and the strategy to correct the torch movements. The seam recognition accuracy was verified during several welding experiments on linear weld seam. Real-time measurements of the seam gap and the seam tracking have achieved a high accuracy

3D Digital Reconfiguration of an Automated Welding System for a Railway Manufacturing Application

“…the images captured by the vision system, are processed in order to give as output the desired features. Based on these features, process parameters such as welding torch position, torch speed, or electrical parameters can be determined based on a suitable decision making procedure, such as a neurofuzzy approach [6][7][8].…”

An on-line shape-matching weld seam tracking system

“…Hence, the seam center that is detected cannot be immediately used to rectify the position of the heat source. Nele et al [17] reported a tracking error of 0.66 mm due to the lag distance of 50 mm between the weld torch and the vision sensor (CCD camera). In practical scenarios where the welding direction changes frequently rather than walking a straight line [17,18], the tracking error is significant.…”

“…Nele et al [17] reported a tracking error of 0.66 mm due to the lag distance of 50 mm between the weld torch and the vision sensor (CCD camera). In practical scenarios where the welding direction changes frequently rather than walking a straight line [17,18], the tracking error is significant. An alternative solution to this problem is to perform seam detection and actual welding in two different passes [19].…”

Weld seam profile detection and feature point extraction for multi-pass route planning based on visual attention model

“…The real intelligent welding encompasses at least three parts. First, guidance of initial welding positions [3,4] should be implemented. Next, accurately weld seam tracking for different weld joints [5,6] has to be included and third, real-time weld defect detection [7][8][9] and welding quality control must be carried out.…”