Conveyor and Camera Calibration from an iPad

Setting up a vision-guided pick-and-place cell has traditionally taken hours. With the conveyor and camera position calibration commands in the Adeeb Robot Controller, that setup now takes minutes — and you can run it from an iPad standing next to the robot.

All you need is a calibration pattern and a calibration tool. The entire setup is a two-step process.

Why calibration decides pick accuracy

A conveyor-tracking cell only works if the controller knows exactly where everything sits relative to the robot. In the Robot Controller configuration, both the camera and the conveyor are defined by a six-value position relative to the world frame — X, Y, Z in millimeters and three rotations in degrees.

Those numbers are what turn a part detected in the camera image into a target the robot can intercept on a moving belt. If the camera frame is off, every pick is off by the same error. If the conveyor frame is off, the robot tracks the belt along the wrong direction or at the wrong height.

Entering those values by hand means measuring offsets and angles on a physical installation, then iterating until the picks land. That iteration is where setup hours go.

Two steps instead of an afternoon

The controller provides dedicated calibration commands for both frames:

• Conveyor position calibration — establishes where the conveyor sits relative to the robot.
• Camera position calibration — establishes where the camera sits, so detections in the image map to real positions on the belt.

Using a calibration pattern and tool, the whole setup completes in a two-step process, with the controller determining the positions instead of an engineer measuring and computing them by hand.

This is why “auto-calibration in under 5 minutes” is listed as a core feature of the Adeeb Robot Controller.

Why an iPad works

The controller serves a web-based UI on port 55000. From there you can modify the configuration, jog the robot, view the 3D simulation, and run Python motion and I/O scripts.

Because it is a browser interface, you are not tied to the PC running the controller. As the Motion Controller API docs describe, any phone or tablet on the same LAN can connect by browsing to the server’s LAN IP address on port 55000. In practice that means the person doing calibration can stand at the conveyor with a tablet, watch the robot and the pattern directly, and step through the setup without walking back to a workstation.

When you still need to recalibrate

For Adeeb turnkey Delta solutions, the camera ships fixed-mounted on the robot, so its position rarely changes. As covered in Running the Adeeb Robot Controller on Linux, you may still need to correct the rotational angle RZ slightly if the mount was not perfectly aligned at installation.

Other common triggers for recalibration:

• Moving the robot or re-leveling the frame
• Replacing or re-routing the conveyor
• Re-mounting or swapping the camera
• Commissioning a new cell from scratch

With a two-step process that runs in minutes, recalibrating after a mechanical change stops being a project and becomes a routine check.

Where this fits

Calibration is the bridge between the Adeeb Delta Robot, its vision system, and the conveyor it serves. The Robot Controller handles native conveyor tracking, and the calibration commands make sure the frames it tracks against are correct from the first pick.

See it on your line

If commissioning time is a bottleneck in your automation projects, take a look at the Adeeb Robot Controller or book a demo to see conveyor and camera calibration on a live cell. Have a specific setup in mind? Contact our team to talk it through.