This video demonstrates how to train the FANUC CRX collaborative robot to locate and pick unfixtured parts using the iRVision 2D camera system. Building on the previous module’s camera setup and calibration, this training focuses on teaching the vision system to recognize parts and programming the robot to pick them accurately.

The process begins by placing the part in the center of the camera’s field of view and initiating the detection setup. Users adjust the magenta box to frame the part and remove any unwanted edges caused by glare or noise. To improve recognition accuracy, the system is trained using multiple examples of the part in different positions and orientations. At least three examples are required to complete model learning.

Once the part is trained, the video walks through setting up the offset data calculation tool to determine the part’s height relative to the calibration grid. This step ensures precise vertical positioning during pick operations.

Next, users set a reference position and record the robot’s snap and pick positions. Using the Pick/Place (Vision) icon in the program editor, the robot is programmed to approach the part, close the gripper, and retreat safely. Speed and height parameters are configured to optimize motion and ensure safe operation.

Finally, the video demonstrates how to execute the program. The robot uses vision to locate the part, moves to the pick position, activates the gripper, and completes the cycle — all without the need for fixturing.

This module equips users with the skills to implement flexible, vision-guided automation using the CRX robot and iRVision system.

6.2 - iRVision 2D Part 2 - Finding & Picking the Part