Choose 2D
If parts lie in one known plane and contrast or contour fully defines the position, 2D is generally simpler to integrate and maintain.
View 2D robot guidanceSolution · Robotics
3D robot guidance for variable height and orientation
When parts are not always presented flat or at the same height, a 2D position may be insufficient. A 3D system estimates the spatial pose, transforms it into robot coordinates and guides the gripper to a part in a controlled presentation.
Application boundary
Between fixed 2D presentation and full bulk picking
3D robot guidance fits a table, conveyor, shallow tray, rack or machine position where height or orientation varies but parts remain sufficiently visible and reachable.
Choose 3D guidance
If the part remains visible but its height, tilt or free orientation changes within a bounded working volume.
Choose bin picking
If parts lie randomly in a deep bin, substantially hide one another and require new grasp candidates and collision checks for every pick.
View 3D bin pickingImplementation
From part trial to a manageable robot cell
Define the process
Record part variants, presentation, working volume, tolerance, cycle time and exceptions.
Measure and test
Test representative parts for visibility, surface behaviour, motion and extreme poses.
Calibrate
Establish camera, robot and tooling in one repeatable coordinate chain.
Integrate
Connect pose, quality status and error codes to the robot controller and, where needed, the PLC or line control.
Validate
Test normal production, changeovers, exception handling and maintenance against agreed criteria.
Production example
Machine tending with varying part orientation
An operator picks cast parts from a shallow carrier, turns them to the required orientation and loads them into a machine. The parts remain largely visible but do not always lie flat. A 3D camera estimates position and orientation. After transformation into robot coordinates, the robot can pick the part and load the machine.
The cell rejects a measurement when too little distinctive geometry is visible or the estimated pose falls outside the agreed confidence limits. This prevents an uncertain move and allows the controller to capture again or request assistance.
Investment and return
Build the business case from your process
Current situation
Record manual hours per shift, ergonomic load, required output, changeover time, rejects, downtime and operator-intervention cost.
New situation
Include camera, robot, gripper, engineering, safeguarding, commissioning, training, maintenance and expected service life. Use only measured capacity and exception handling from a representative trial.
Feasibility
Test parts and process limits before selecting the system
Share representative parts, the current presentation, robot platform and required process outcome. We determine the trial needed and whether 2D, 3D guidance or full bin picking is the appropriate level.