🧪 New tutorials are being published — build from robot arms to sensors step by step
Skip to content

Windows ​

Connect the camera and the computer ​

Shell
lerobot-find-cameras opencv

imagepng

Teleoperate and display the camera feed ​

Shell
lerobot-teleoperate --robot.type=so101_follower --robot.port=COM6 --robot.id=my_follower_arm --robot.cameras="{ front: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 60, fourcc: "MJPG"}}" --teleop.type=so101_leader --teleop.port=COM7 --teleop.id=my_leader_arm --display_data=true

A rerun.io window opens, displaying the trajectory of each servo joint and the live camera feed in real time

The images are also saved to the C:\Users\<username>\outputs\captured_images directory

imagepng

If you encounter the following error ​

The camera cannot be connected, but switching cameras in Tencent Meeting can still open it normally

imagepng

Modify the lerobot\src\lerobot\cameras\utils.py file and change the OpenCV backend to cv2.CAP_SHOW

imagepng

This is a bug that even Doubao cannot solve. It is because the lerobot library is wrapped too deeply, making it very hard for beginners to debug.

wx_camera_1768139334330.mp4

Connect multiple cameras, teleoperate and display the camera feed ​

Shell
lerobot-teleoperate --robot.type=so101_follower --robot.port=COM6 --robot.id=my_follower_arm --robot.cameras="{ front: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 60, fourcc: "MJPG"}, side: {type: opencv, index_or_path: 1, width: 640, height: 480, fps: 60, fourcc: "MJPG"}}" --teleop.type=so101_leader --teleop.port=COM7 --teleop.id=my_leader_arm --display_data=true