3d2d_ann

This repository contains the code to project 3D point cloud to 2D image, doing a correspondence.

match.py

To run the match, simply use

python match.py 1326035199606919

It means it will read the 1326035199606919.bin in ./velodyne_sync, and automatically find the closest images of 5 cams.

It can project 3D point cloud to 2d.

To run the project, simply use

python project_vel_to_cam.py vel img cam_num

vel: The velodyne binary file (timestamp.bin)

img: The undistorted image (timestamp.tiff)

cam_num: The index (0 through 5) of the camera

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
beobotv3_camera_intrinsics		beobotv3_camera_intrinsics
gpt4		gpt4
proj		proj
proj_colmap		proj_colmap
proj_lego		proj_lego
proj_pcd		proj_pcd
.gitignore		.gitignore
Accurate labelling of pointclouds with semantic labels using images.ipynb		Accurate labelling of pointclouds with semantic labels using images.ipynb
CVPR_compare_SAM.py		CVPR_compare_SAM.py
README.md		README.md
all_sync.sh		all_sync.sh
cam1velo.bag		cam1velo.bag
combine.py		combine.py
ground_plot.py		ground_plot.py
labels_proj		labels_proj
manual.py		manual.py
match.py		match.py
point_cloud.ply		point_cloud.ply
point_cloud_.ply		point_cloud_.ply
project_vel_to_cam.py		project_vel_to_cam.py
sam.rec		sam.rec
sam.sh		sam.sh
semantic.yaml		semantic.yaml
sync_cam.py		sync_cam.py
sync_cam.sh		sync_cam.sh
sync_cam2.py		sync_cam2.py
sync_check.py		sync_check.py
sync_data.json		sync_data.json