Skip to content
Extract geometry features from 3D point clouds
Python Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Extract geometry features from 3D point clouds


The project contains the following folders:

  • geo3dfeatures contains source code
  • docs contains some mardown files for documentation purpose and images
  • examples contains some Jupyter notebooks for describing data
  • tests; pytest is used to launch several tests from this folder

Additionally, running the code may generate extra subdirectories in a chosen data repository (./data, by default).

How to install

This projects runs with Python3, every dependencies are managed through poetry.

Installation from source

$ git clone ssh://
$ cd geo3dfeatures
$ virtualenv -p /usr/bin/python3 venv
$ source venv/bin/activate
(venv)$ poetry install



Run commands

In order to get the available program commands, consider the program help (geo3d -h):

usage: geo3d [-h] {info,sample,index,featurize,cluster,train,predict} ...

Geo3dfeatures framework for 3D semantic analysis

positional arguments:
    info                Describe an input .las file
    sample              Extract a sample of a .las file
    index               Index a point cloud file and serialize it
    featurize           Extract the geometric feature associated to 3D points
    cluster             Cluster a set of 3D points with a k-means algorithm
    train               Train a semantic segmentation model
    predict             Predict 3D point semantic class starting from a
                        trained model

optional arguments:
  -h, --help            show this help message and exit

Any further CLI documentation may be printed with geo3d <command> -h.


Some documentation is available, that describes the set of considered geometric features, the fixtures (i.e. dummy datasets) used for test purpose and a practical pipeline use case:


The following example has been generated starting from a CANUPO dataset (file, with 500k points, 50 neighbors and all the features):


The example that is run in the test (b9.ply) comes from the CGAL repository. Thanks to their maintainers (for more details, please refer to CGAL, Computational Geometry Algorithms Library,


Oslandia – 2019-2020

You can’t perform that action at this time.