Vision-guided component packing at high line takt

Placing small furniture components into packages at high line takt is fast and unforgiving: components arrive on conveyors in varying positions and materials, every package must be filled correctly each cycle, and a starved or overfilled conveyor stops the whole line.

QiPICK guides four robot loading stations: neural networks trained per part type find position and rotation of 3D objects on the moving conveyor — no 3D scanner needed — and direct up to 90 picks per minute into the packages, communicating directly with the sidebot robots. Lighting is controlled by the vision system itself, dimmed per program, so detection stays stable without burdening the line control. In parallel, four Q2L smart modules with external cameras monitor conveyor fill levels with programs per material and signal the line before feeding becomes a problem.

Parts are loaded at full line speed with machine-verified placement, and conveyor fill is watched continuously instead of by someone walking the line. Delivered with integrator Stimba and extended in follow-up projects directly with the customer.