A Leap in Automation: Bin of Metal Bars Powered by 3D Vision for Metal Loading & Unloading

Metal bar loading is a critical link between metal processing and heat treatment. Nine specifications of metal round bars are neatly arranged in standard bins. Manual handling one by one is not only inefficient, but reflective metal surfaces also make stable imaging difficult for traditional 2D vision.

TransferTech deployed the Epic Eye Laser L V2S 3D industrial camera together with Epic Pro vision software to deliver automated upgrades for this process on a metal processing production line. The 3D camera is fixed-mounted above the bin. A single scan captures the position and pose of all bars inside the bin and guides the robotic arm to pick and load workpieces.

Loading Multiple Bar Specifications: Testing Vision Performance

The incoming workpieces on this line are metal round bars in nine specifications, with diameters ranging from 34 mm to 60 mm and lengths from 345 mm to 670 mm. The workpieces are placed inside standard bins sized 1070 mm × 840 mm × 670 mm. Parts are arranged in order with consistent orientation and densely packed within each bin.

This process previously relied on manual piece-by-piece handling: operators retrieved bars from the bin and placed them into the downstream guide chute, where parts travelled along the conveyor line into the heat treatment station. Continuous repetitive handling consumed labor resources. Meanwhile, reflective and overlapping surfaces of metal bars made stable recognition challenging for conventional vision solutions, limiting production takt time and product changeover efficiency.

Solution Concept: Enabling the Robotic Arm to "See and Precisely Pick"

TransferTech’s solution is straightforward. One Epic Eye Laser L V2S 3D smart camera is fixed-mounted roughly 1400 mm directly above the bin. Its single field of view fully covers the entire bin.

Once the bin is in position, the control system triggers a scan command. The camera initiates blue laser galvanometer scanning to capture a complete 3D point cloud on the surfaces of metal bars. Epic Pro vision software automatically calculates the position and pose of the top-layer bars, computes the optimal grasp coordinates, and transmits the data to the robotic arm via the TCP/IP protocol.

The robotic arm end effector uses a V-shaped electromagnet to pick bars and place them into the guide chute. Parts are then conveyed to the heat treatment process. The vision system repeats recognition and pick guidance until the bin is empty, then sends an empty-bin signal to the control system to prompt bin replacement. The whole workflow requires no manual intervention. Multi-specification changeover is triggered by model signals issued from the control system, with no on-site recalibration required.

Core Hardware: Large Field-of-View Laser Camera + Graphical Vision Software

The hardware adopted is TransferTech Epic Eye Laser L V2S laser galvanometer 3D industrial camera.

This model is engineered for mid-to-long range, large field-of-view industrial applications, with a working distance ranging from 1200 mm to 3000 mm. It features a binocular structure with a 400 mm baseline and a 450 nm blue laser galvanometer scanning path. It delivers strong immunity to ambient light interference in workshop environments and generates continuous, complete point clouds even for highly reflective metal components.

The software is Epic Pro. Its graphical drag-and-drop interface supports camera connection, hand-eye calibration and grasp parameter configuration, and can interface with mainstream industrial robots and PLCs.

On-site Results: Precision, Takt Time and Usability

After project deployment, core metrics of the vision system remain stable: recognition accuracy ±0.5 mm, recognition rate ≥99.9%, and the takt time per workpiece from scan to calculation completion ≤4 seconds.

During on-site validation, camera scanning takes approximately 2.1 seconds and algorithm processing takes around 1.8 seconds, leaving ample time for robotic arm movements.

Nine bar specifications share one unified vision configuration. Model switching is activated via signals from the control system. IP65 protection and wide-temperature-range design allow the equipment to operate reliably long-term in workshop environments with dust, oil mist and stray light.

Takeaways for Similar Projects

For manufacturers planning metal bar and shaft-type workpiece loading/unloading automation retrofits, this project offers three key takeaways:

Large bins with orderly arranged workpieces suit fixed-mounted large field-of-view cameras. A single scan covers the whole bin without complex motion mechanisms.

For highly reflective metal workpieces, verify point cloud quality at the product selection stage. Blue laser galvanometer scanning delivers good adaptability for such working conditions.

The vision system handles position and pose guidance, while posture correction is delegated to downstream fixtures. This shortens commissioning cycles and enables flexible multi-specification changeover.

Empowering flexible robotic loading with 3D vision, TransferTech continuously refines its self-developed cameras and algorithms to advance automation upgrades across the metal processing industry.