First, model and gripper CAD files and actual RealSense video captures are imported and prepared for use in PoseStream Designer. Next, the user matches the 3D CAD to the video captures and programs the desired grips. 

Model settings such as symmetry are configured to optimize robotic performance and precision during operations.

Our cloud service processes files within 24 hours. Customers prepare their files, email them to us, 

and we will promptly send the results back via email.

The PoseStream Vision System inference model 

leverages advanced algorithms to accurately predict 

object poses and collision free grips in real-time. This model enables precise 

object detection, positioning, and handling, enhancing 

the efficiency and accuracy of robotic operations.

In the final step, robot communication ensures that 

the robotic system receives and interprets data from 

the PoseStream Vision System correctly. This enables 

the robot to execute precise actions based on the real

-time information provided, completing the automated 

task efficiently.

Speed: 16 FPS 

Accuracy: 1mm / 0.3mm (Cluttered / Singulated)

Use cases: Pick parts from bins, trays, etc. Precision assembly or placement of parts to PCB, tray, etc. Inspection of part placement

CONTACT US