Skip to content

KopiSoftware/VINS_Lite_GPU

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

VINS_Lite_GPU

VINS FUSION GPU lite version.

vins_lite_gpu-2


Update:

  • demostration video: https://youtu.be/q81QQVIy3yA?feature=shared

  • Some old opencv api has been updated.

  • Disable show_track in configuration yaml if it start up with a remote ssh.

  • Rebundent dictionary file has removed.

Our works

  • Memory leak problem is fixed.
  • We trained a DBOW vocabulary with LIP6Indoor dataset. This model achieves 95% of the performance of the original model, and much smaller.
  • We have test this model on Nvidia Jetson Nano 2GB version with MYNTEYE s1030-ir. The RAM consumption is about 60MB when program starts up. The usage of RAM will increase with the number of keyframe or running time. The GPU consumption is about 20%.


Performance

It can be very easy to reach 30fps on Jetson Nano 2GB with half feature number cut down. Please check the pose_graph in RVIZ rather than tracking image view. The pose_graph is realtime, but the tracking image view has a serious delay. Here is my configuration:

multiple_thread: 4
max_cnt: 75
freq: 30                # frequence (Hz) of publish tracking result. At least 10Hz for good estimation.
max_solver_time: 0.02  # max solver itration time (ms), to guarantee real time
max_num_iterations: 4   # max solver itrations, to guarantee real time

Installation

Please follow original VINS FUSION GPU. The installation process is identical. You may need vitrual memory to compile the program.

Acknowledgment

The Jetson devices in our experiment is supported by NVIDIA Jetson Nano 2GB Developer Kit Grant Program.

About

VINS FUSION GPU lite version.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published