3D bin picking

3D bin picking systems for reliable robot automation.

Manual loading can limit machine capacity or require repeated handling of heavy or awkward parts. Sedeco engineers 3D bin-picking systems around the actual parts, bin, gripper and destination. A feasibility trial establishes the usable presentation range before camera, software and robot integration are finalised.

Complete system integration

Validate detection, gripping and placement together

A detected part is not yet an accepted placement. The cell must find an accessible grip, plan the approach and withdrawal, confirm pickup and deliver the part within the receiving process limits. Define the outcome for an empty bin, no reachable grip and a failed pickup before acceptance.

01

See the bin in 3D

Test usable depth and feature visibility at full, partly depleted and nearly empty bin levels. A surface hidden by another part cannot be assumed to be measured; evaluate which visible features support a valid part estimate.

02

Choose a feasible pick

Evaluate the gripper against available contact surfaces, required holding force and clearance at the walls. Include interlocking parts and double-pick detection where relevant. Validate the approach and withdrawal path with the actual tooling.

03

Pick and look again

Confirm pickup and the required orientation or position at delivery. Measure from the start of the attempted cycle until an accepted placement, including retries and recovery. A fast successful demonstration does not describe the full production cycle.

Why is it difficult?

Every bin creates different requirements

A working solution follows from the combination of part, surface, bin, gripper and required speed.

Size

Small parts require sufficient detail. Large parts require a wider measurement area and enough visibility of the bin.

Speed

Image capture, calculation and robot movement must fit the available cycle time together.

Surface

Matte parts are generally easier to observe than glossy, dark or translucent surfaces.

Shape and overlap

Parts can interlock, lie on top of each other or remain only partly visible.

Bin and reach

Deep bins, high walls and corners restrict what the camera, robot and gripper can reach.

Gripper

Vacuum, clamping and magnetic gripping create different requirements for clearance and gripping points.

Project stages

From sample trial to on-site acceptance

Feasibility and cell design

During feasibility, test representative fills and surface variants with candidate optics and grips. Deliver the observed limits, failed conditions and a preliminary timing budget. During design, document camera positions, calibration frames, gripper interfaces and the required robot clearance.

Integration and handover

During integration, test part/result association, robot commands, pickup confirmation and recovery. At site acceptance, run full-to-empty sequences with the actual replenishment and receiving process. Handover includes configuration and calibration backups, operator recovery instructions and a tested restore procedure.

Agree the scope in stages: sample feasibility, cell design, integration and on-site acceptance. Each stage should produce evidence for the next decision. Record unresolved conditions and who will test them before committing to the final configuration.

Available 3D hardware

An RGB-D camera as part of the solution

The camera supplies image and depth information. Its suitability depends on part size, measurement area, surface, working distance and speed.

Hikrobot MV-DB500S-S RGB-D camera
RGB-D camera

MV-DB500S-S

View the product page for the available specifications. We can then assess whether its measurement area, the surface and application fit this camera.

View MV-DB500S-S →

Software connects the steps

Measure accepted placements across bin depletion

Agree acceptance by bin level and part variant. Record attempted picks, accepted placements, failed or double picks, remaining unreachable parts and operator interventions. Divide total elapsed trial time by accepted placements to report time per accepted part, stating whether replenishment and interventions are included. Specify placement tolerance and test it independently. Include restart, changed bin position and loss of robot or camera communication.

Explore our software →

Start with the actual bin

Ask Jeroen to assess feasibility

Provide parts, surface variants, bin and destination drawings, grip restrictions and the required accepted output. For the business case, record manual loading hours, ergonomic demands, refill and recovery work. Budget hardware, tooling, integration and validation hours × rate, maintenance and expected changeovers. Avoid counting recovered capacity and the same downtime reduction as separate benefits.

Jeroen van den BergJeroen van den Berg

First decide what is needed

Not every pick from a bin requires full 3D bin picking

The decision guide compares 2D guidance, constrained 3D guidance and random bin picking based on part presentation, overlap and room to move.

Read when 3D is needed →Read what 3D bin picking is →