Closed ajsnyde closed 1 year ago
--cap-add IPC_OWNER
?--ipc host
?--privileged
?export LIBGL_ALWAYS_SOFTWARE=1
before running stella_vslam?sudo docker run -it --rm --cap-add IPC_OWNER --ipc host --privileged -e DISPLAY=$DISPLAY -v /tmp/.X11-unix/:/tmp/.X11-unix:ro stella_vslam-desktop
Finally, I ensured that export LIBGL_ALWAYS_SOFTWARE=1
was run before running the basic tutorial command:
./run_video_slam -v ./orb_vocab.fbow -m ./aist_living_lab_1/video.mp4 -c ../example/aist/equirectangular.yaml --frame-skip 3 --no-sleep --map-db-out map.msg
, of course after downloading the vocab and materials listed in the tutorial.
As a workaround, I'm using the server/socket to run and display right now.
Do not use sudo
to do docker run
. See https://docs.docker.com/engine/install/linux-postinstall/
yes, on the host machine it runs fine
What about when run inside a container?
Can I visualize without gpu? I have just Cpu
This is a problem related to OpenGL and docker. The solution depends on the environment. If you are willing to help us, please provide the following information
--cap-add IPC_OWNER
?--ipc host
?--privileged
?export LIBGL_ALWAYS_SOFTWARE=1
before running stella_vslam?Can I visualize without gpu? I have just Cpu
It worked without GPU on Ubuntu 18.04. I think it will work on Ubuntu 20.04 or later, but it doesn't seem to work in some environments.
In my environment, I solved the problem by adding --privileged
. documentation has been updated.
Not sure what could be causing this, but the map viewer window is "blank" and shows a fixed image of what was behind it when it opened. The camera/video feed, however, functions normally. Running the basic tutorial after successfully building the docker. Host machine is Ubuntu 22.04 with xfce installed on top. https://user-images.githubusercontent.com/11302823/220725247-fdf29955-0595-46f3-8a39-a3d24d41aa92.mp4