VINS-Fusion-Lite

VINS FUSION lite version.

Our works

We trained a DBOW vocabulary with LIP6Indoor dataset. This model achieves 95% of the performance of the original model, and much smaller.
We adapted program with a higher opencv version. It can be run on Ubuntu20.04 with opencv 4.2.0 now.
Memory leak problem is fixed.
This program is related to our program: VINS_Lite_GPU. It can be run on a embedded device with CUDA such as Nvidia Jetson Devices.

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

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
camera_models		camera_models
config		config
docker		docker
global_fusion		global_fusion
loop_fusion		loop_fusion
support_files		support_files
vins_estimator		vins_estimator
LICENSE		LICENSE
README.md		README.md