High Dynamic Range Machine Vision for Automated Seam Tracking

Xiris Automation is releasing WeldStudio Tracker, an artificial intelligence-driven machine vision software that consolidates seam tracking and weld monitoring into a single optical device. This technology addresses the requirements of robotic and automated welding applications by providing simultaneous real-time process correction and visual feedback for operators.

AI-Driven Optical Seam Tracking
Automated and robotic welding systems typically rely on tactile, through-the-arc, or laser optical trackers to compensate for metal warping, heat distortion, and poor edge preparation. However, these conventional tracking mechanisms operate independently of visual process monitoring systems, preventing operators from visually confirming whether the process remains in control. The new software integrates machine vision directly into near-infrared, high dynamic range (HDR) cameras operating at over 140 decibels. This extended dynamic range manages the extreme brightness of the welding arc, allowing the system to capture distinct, simultaneous images of the welding torch, arc, melt pool, and joint seam without oversaturation.

Zero-Offset Measurement and Payload Reduction
By utilizing artificial intelligence models trained on specific industrial automation processes, the software identifies and tracks key topographical features of the root, weld bead, and joint across multiple passes. Because the measurements are extracted directly from the camera's live field of view at the arc, the system eliminates the physical spatial offset typically required between a standalone seam sensor and the electrode. The camera image captures both the torch and the groove, allowing the system to calculate their absolute positions and relative movements continuously.

Consolidating process monitoring and seam tracking into a single hardware unit reduces the overall payload weight on the end of the robotic arm and simplifies hardware integration. This unified approach allows the automated welding cell to execute autonomous trajectory corrections while giving the human operator a continuous live feed to evaluate process inputs and weld quality.

Additional Context
This section details technical specifications and competitive benchmarking not included in the original news release.

In the automated welding market, seam tracking is traditionally dominated by laser triangulation systems from manufacturers such as SERVO-ROBOT and Meta Vision Systems. Laser triangulation provides high-precision topographical profiles of the weld joint but requires a secondary camera system if operators require visual process monitoring, which increases the end-of-arm tooling weight and complexity. Through-the-arc seam tracking (TAST), commonly integrated into advanced power sources by companies like Fronius and Lincoln Electric, adds zero weight to the torch but completely lacks visual feedback and is generally limited to weaving weld processes on thicker materials.

The Xiris integration of 140+ dB HDR near-infrared imaging with AI-based tracking competes directly by reducing the robotic payload and eliminating the spatial measurement offset between the sensor and the weld pool. While laser scanners provide geometric absolute measurements based on structured light, vision-based AI tracking relies on specific algorithm training to identify joint features in 2D space, offsetting the need for fragile laser emitters near the spatter zone.

Edited by Aishwarya Mambet, Induportals Editor, with AI assistance.

www.xiris.com