[Shankar G]

Hello RoboDK Team,
I’m currently working on a robotic vision project for my final-year engineering.
Setup Details:

• Robot: UR3 (connected to RoboDK)
  
• Camera: Intel RealSense D435 (USB connection)
  
• AI Model: YOLOv8 (trained with custom metal surface dataset)
  
• Controller: Python API with RoboDK integration
  
• Goal: To detect and pick up metallic objects in real time using RoboDK + YOLOv8 detection results.
  

The YOLOv8 model runs correctly and detects the material in real-time using Python, but I need help integrating the detections with RoboDK to control the UR3’s movement automatically.
Questions:

1. How can I send YOLOv8’s detected (x, y, z) coordinates from the RealSense camera to RoboDK in real time?
   
2. Should I use RoboDK’s Python API or the URSim interface for coordinate synchronization?
   
3. Is there an example or documentation for camera-based robot movement in RoboDK?
   

Any sample scripts, tips, or project examples would be extremely helpful.
Thank you,
Shankar G.
(Final Year Robotics Engineering Student)

[Albert]

I recommend you to take a look at these examples which show how to simulate and use 3D depth cameras in RoboDK:
https://robodk.com/doc/en/PythonAPI/exam...-camera-3d