Photoneo bin picking

Photoneo builds structured-light 3D scanners (the PhoXi range for static parts, MotionCam for parts moving on a belt) and pairs them with software that turns a raw point cloud into ranked, collision-checked grasp poses. The pipeline is aimed squarely at random bin picking.

In a cell the scanner sits above a bin or tote, captures the scene, and hands the robot a specific grasp target plus an approach path that clears the bin walls and neighboring parts. That path-planning step is what separates a bin-picking product from a general-purpose 3D camera like Zivid 3D.

Software integrates with FANUC, ABB, Yaskawa, KUKA, and Universal Robots controllers, so the picked pose is delivered in a form the robot's own motion planner can execute directly. That removes most of the coordinate-frame plumbing a from-scratch integration would require.

Choose Photoneo when the part is presented randomly and the shop has already tried and failed to fixture around it, when the pick has to work across many SKUs on the same cell, or when a project has been quoted around a vendor package rather than a custom vision stack.

Compatibility caveats are physical, not software. Shiny, black, or transparent parts scatter or absorb the projected pattern and produce sparse point clouds. Deep bins with tall walls limit the grasp angles the scanner can see and the arm can reach, so bin depth and part orientation need review during cell design.

Commissioning trouble tends to be: bin fill level assumptions that break as the bin empties, gripper tooling that can't reach the corners, ambient light bleeding into the scanner enclosure, and grasp-planning parameters (approach angle, retract clearance) that need part-specific tuning.

Typical end effectors: soft grippers for delicate or variable parts, Piab vacuum for flat sheet metal, and heavier parallel or two-finger tools like Schunk grippers or the Robotiq 2F gripper for machined parts, often on the arm via an ATI tool changer.