Academic

UR3 Robot Vision Sorter

I worked on an academic sorting cell using UR3 arms, a conveyor, and RGB cameras. OpenCV detected the cubes by color and located each cube’s centroid to help align the gripper for pick-and-place motion.

Two UR3 robot arms beside a conveyor, with camera hardware and a laptop at the sorting bench. — The UR3 vision sorter: robot arms, conveyor, camera hardware, and control station.

Finding the center of a cube

OpenCV processed the RGB-camera images for basic cube-color detection. The detected centroid gave a position reference for aligning the gripper over the cube.

From vision to handling

The bench combined two UR3 arms, the conveyor, camera hardware, and a control station. PolyScope programming handled the robot motion within the sorting workflow.

Technical details

Vision software: OpenCV for cube-color detection and centroid position.
Vision hardware: RGB cameras.
Robot cell: Two UR3 arms, a conveyor, and PolyScope robot programming.

A closer look

UR3 vision sorting demonstration. Full 1-minute 51-second video.