no process_data.py in folder "act_plus_plus"

[VictoriYang]

Hello again!
I'm trying to reproduce the policy training process. However, it occurs to me there's no process_data.py as described in README.md.
Please help me check it out when it's convenient for you. Thanks a lot!

[xshenhan]

Hello again!
Apologies for the late response. I’ve uploaded the missing process_data.py file and updated the README.md with additional explanations. Please check the repository for the latest updates.
Let me know if you have any further questions—happy to help!

[VictoriYang]

Thank you for your reply!

Another question for:

    Prepare the data:
    Place the images in the folder /path/to/input/images.
    Place the image for alignment in /path/to/input/align_image.png.

I got None results when running
corners, ids, _ = cv2.aruco.detectMarkers(gray, charuco_dict), and a possible reason is my unproper image. Could you please provide an example folder including images for quick exp. and some detailed instructions for preparing these images (like the camera view or aruco-markers)? It would be very helpful!

[xshenhan]

The issue might be due to a mismatch between the markers you’re using and the dictionary specified in the code. The relevant lines in get_marker2base_aruco.py are:

charuco_dict = cv2.aruco.getPredefinedDictionary(cv2.aruco.DICT_6X6_250)
board = cv2.aruco.CharucoBoard((5, 5), 0.04, 0.03, charuco_dict)

To resolve this, you can either create a high-precision calibration board or download one from calib.io and print it. If you already have an ArUco marker, you can also modify the code to match its parameters. The script get_marker2base_aruco.py will work as long as some of the images in the joint_positions pair from real-deployment/calibration/hand_in_eye_shooting.ipynb contain clear and complete markers. If the automatic script fails to capture the markers (e.g., you may not want to put the marker in centered in the scene), you can manually control the robotic arm to take additional images and use them as input for get_marker2base_aruco.py. Note that the images used by hand_in_eye_calib.py and get_marker2base_aruco.py do not need to be the same.

I’ll also provide example images with properly prepared images and detailed instructions for the alignment step soon.