Buyers often ask whether a 3D scanning welding robot can cope with real shop-floor steel: mill scale, shiny cut edges, spatter and seams that are already partly welded. The short answer is yes, with the right camera setup and a few habits. This article explains how the two-camera system works and what to do when a surface fights back.
Can a 3D scanning welding robot scan reflective or spatter-covered steel?
Yes, in most cases. A wide-view camera locates the part and a close-range camera picks precise points near each seam, which limits the effect of glare. When a point cannot be picked, check the surface for glare, spatter or existing beads, bring the close-range camera nearer to the point, or rescan with the wide-view camera.
1How does the two-camera system work?
The system uses two 3D vision cameras for different jobs. The wide-view camera captures the whole part to find where it is. The close-range camera then moves to each key point and picks it precisely. Unlike simple laser seam tracking, the system builds a model of the part first and then generates the path, so point-picking quality decides how accurately the seam is located.
In the 9-axis cantilever reference configuration the global camera works at 1.2–3 m and the local camera scans at 0.4–1 m, with a 600 × 450 mm field of view at 700 mm. Camera options are listed on the configuration page, and camera setup is covered in RVC 3D camera in robotic welding.
2Which surfaces cause scanning trouble?
| Surface | Effect on the scan | What to do |
|---|---|---|
| Glare | Bright patches or shadows in the image | Change the close-range shooting position; check overhead lighting |
| Spatter | False points near the seam | Clean the seam area before scanning |
| Existing beads | Seam edge is hidden or shifted | Scan before the earlier pass, or adjust the template’s point picking |
| Wrong distance | Close-range image gets blurrier | Shoot near the selected point, within the camera’s range |
3What should you do when a point cannot be picked?
- Identify which camera took the image
Wide-view and close-range problems are solved differently.
- Look at the part surface
Glare, spatter and existing beads are the usual causes.
- Bring the close-range camera nearer to the point
Adjust the point-picking settings so it shoots within its working distance.
- Rescan with the wide-view camera
If the close-range scan still fails, a fresh wide-view scan often recovers the part position.
- Rescan and check the path preview
Weld only when the preview shows the path on the seam.
One structural steel user who had worked with another 3D scanning system said the point cloud from this setup was much better. Before suspecting the camera, rule out the surface and the shooting distance.
4How should you test your own parts before buying?
Bring your worst-case parts to the acceptance test: the shiniest cut edges, the most spatter and seams next to existing welds. A test that only uses clean samples tells you little. What to measure is covered in the adaptive welding acceptance test.
Frequently asked questions
Why does the close-range camera get blurrier as it moves closer?
It is outside its working distance or shooting at the wrong position. Adjust the point-picking settings so it shoots near the selected point.
Can spatter on the part stop a welding robot from finding the seam?
It can create false points near the seam. Clean the seam area before scanning, then rescan and check the path preview.
How do I confirm a scanning problem is fixed?
Rescan the part, check that the path preview follows the seam, then run a test weld.
Test-scan your hardest parts
Send photos of your most reflective or spatter-covered parts. We will tell you how the two-camera system handles them.
- JTCLASER field commissioning and after-sales records, 2026
- JTCLASER 9-axis cantilever welding robot workstation, published reference configuration