Skip to content

3D viewer to display KITTI pointcloud and 3D boxes

License

Notifications You must be signed in to change notification settings

ThomAub/viewercloud

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

67 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Viewercloud

Rust general CI

Python package Release Package Release Flow

Viewercloud is a library and also a cli to read and display Pointcloud. It was initially used to display KITTI pointcloud. However it was also tested on the Lyft Level 5 Dataset pointcloud.

It will also be able to display the 3D annotations and the 3D BoundingBox computed by your favorite algorithm.

Viewercloud will open a openGL window to display the pointcloud. Press qto close.

kitti-centroid

It can also take a screen shot of the current view to save as png. Press s take screenshot.

lyft-color

Python Usage

You can install pyviewercloud the python bindings to viewercloud thanks to Pyo3 and Pyo3-numpy

poetry add pyviewercloud
pip install pyviewercloud
import numpy as np
import pyviewercloud as pyviewer

# Create a new viewer with a window size 1200x1800.
viewer = pyviewer.PointcloudViewer(1200, 1800, 15000)

# Load some pointcloud from the lyft perception dataset
# Currently only support pointcloud as numpy.ndarray Nx3 in np.float32
lyft_point_cloud_1 = np.fromfile("tests/data/lyft/host-a101_lidar0_1241893239502712366.bin", dtype=np.float32).reshape((-1, 5))[:,:3]
lyft_point_cloud_2= np.fromfile("tests/data/lyft/host-a101_lidar1_1241893239502712366.bin", dtype=np.float32).reshape((-1, 5))[:,:3]
lyft_point_cloud_3= np.fromfile("tests/data/lyft/host-a101_lidar2_1241893239502712366.bin", dtype=np.float32).reshape((-1, 5))[:,:3]

# Add them one by one to the viewer to have different color
viewer.add_pointcloud(lyft_point_cloud_1, [255, 0, 0])
viewer.add_pointcloud(lyft_point_cloud_2, [0, 0, 255])
viewer.add_pointcloud(lyft_point_cloud_3, [0, 255, 0])

# You can now display the window
viewer.show()
import numpy as np
import pyviewercloud as pyviewer

# Create a new viewer with a window size 1200x1800.
viewer = pyviewer.PointcloudViewer(1200, 1800, 15000)

# Load some pointcloud from the kitti dataset
kitti_point_cloud = np.fromfile("tests/data/kitti/velodyne/000001.bin", dtype=np.float32).reshape((-1, 4))[:,:3]
viewer.add_pointcloud(kitti_point_cloud, [255, 255, 255])

# Add some centroids to have the same color.
# Currently only support centroids as numpy.ndarray Nx3 in np.float32
centroids = np.array([[-11.5,0,-0.8]]).astype(np.float32)
viewer.add_centroid(centroids, [255, 0, 0])
viewer.show()

CLI Usage

> viewercloud --help
viewercloud 0.2.1
Thomaub <[email protected]>
Display KITTI 3D Pointcloud with annotations and your model inferences

USAGE:
    viewercloud <pointcloud-file> [ARGS]

ARGS:
    <pointcloud-file>     Path to the kitti Pointcloud .bin or .txt file
    <annotations-file>    Path to the kitti 3D annotation .txt file
    <inferences-file>     Path to your model inferences .json file

FLAGS:
    -h, --help       Prints help information
    -V, --version    Prints version information

Build

cargo build --release

It will be possible to download from github release and brew.