search

xdspacelab / openvslam

OpenVSLAM: A Versatile Visual SLAM Framework
https://openvslam.readthedocs.io/
2.97k stars 870 forks source link

cannot extract any keypoints #453

Open roboticsai opened 3 years ago

roboticsai commented 3 years ago

When i run the openvslam run_slam process. It works for few moments than it is not able to detect any keypoints in image. Here is my command line which i'm using to run the slam process: ./run_camera_slam -v ../data/orb_vocab/orb_vocab.dbow2 -c ../data/aist_living_lab_2/config.yaml --frame-skip 3 --no-sleep --map-db map.msg -n -1 System setup: hardware: raspberry pi 4B ( 4 gb ram ) OS: ubuntu server 20.04 camera: logetech 270 dataset: aist_living_lab_1

Isn't their any vocabulary dataset which is generic like this one provided by orb_slam2? Here is the demo video.

mirellameelo commented 3 years ago

Hi. I truly believe the vocabulary used in Openvslam is generic as well. The link for the video you provided is not available. Could you please double-check it? If you're trying to run the SLAM system with your personal camera, you should give the path to its respective camera parameters. So, in order to build a map, change "../data/aist_living_lab_2/config.yaml" to the path of your yaml file.