Import a point cloud file and perform poisson 3D surface reconstruction algorithm, integrated with third-party libraries (e.g. open3d, pymeshlab...)
Project description
Poisson Surface Reconstruction: 3D point cloud
Import a point cloud file and perform poisson 3D surface reconstruction algorithm, integrated with third-party libraries like open3d and pymeshlab
Dependencies
- python 3 <= 3.8.x
Recommended: Use pyenv to install and manage Python versions
- numpy >= 1.20
- open3d >= 0.12
- pymeshlab >= 0.2
Development dependencies
- setuptools: For installation via
setup.py
- setuptools-scm: To generate version numbers from git tags
- wheel: Built packages
.whl
to install packages with PIP - twine: Publish packages to https://pypi.org
- tqdm: CLI progressbar when publish a package
Development guide
For local installation and develop new features for this package, follow the steps below:
Windows
Run unassigned powershell scripts on Windows requires change your execution policy with
Set-ExecutionPolicy
toAllSigned
, orBypass -Scope Process
.
See: Execution Policies
Run the script setup.ps1
with powershell
# Default virtualenv folder is "venv"
.\setup.ps1
# Optionally, pass a custom virtualenv folder
.\setup.ps1 -VirtualEnvFolder myenv
If you wish run just an specific function from setup.ps1
(e.g. Install-Python
), execute:
powershell -command "& { . .\setup.ps1 'venv' -Execute 0; Install-Python }"
Unix systems (Comming soon)
# Default virtualenv folder is "venv"
./setup.sh
# Optionally, pass a custom virtual enviroment folder
./setup.sh --virtual-env-folder myenv
For more detailed installation, see the Wiki pages in: Installation Development
Install this package
Quick way:
pip install surface_reconstruction
Or clone this repository and build the .whl
file from the project root:
cd [your-project-root]
# Build the .whl package file
python -m build --sdist --wheel .
# Install locally from the .whl file
# Where: x.x.x is the generated version from git tag
pip install dist/surface_reconstruction-x.x.x-py3-none-any.whl
For more detailed, see the Wiki pages in: Generate package
Run the unit tests
# Run all tests of the module "surface_reconstruction_test`
python -m unittest tests/surface_reconstruction_test.py
Usage
Import a .ply
file with point cloud vertices, and generate the mesh file
from surface_reconstruction import SurfaceReconstruction
import os
# Pass a method/library that contains a Poisson algorithm implementation
surface = SurfaceReconstruction(
method_type='open3d',
point_cloud_file=os.path.join('files', 'point_cloud.ply'),
output_file=os.path.join('files', 'terrain_mesh.ply')
)
# Call the method from the specific library, and export a mesh file
surface.poisson_mesh()
You can pass custom filters/parameters for the specific library. This is important because poisson algorithm requires some pre-filters before to be applied (e.g estimate normals in the point cloud)
# ...
parameters = {
'estimate_normals': {
'fast_normal_computation': False,
'normals': (1, 3)
}
}
# Unpack the dictionary "parameters" as a **kwargs
surface.poisson_mesh(**{'filters': parameters})
PS: See the unittests inside tests folder for more usage examples
Extending: Add new libraries
Is possible create and register custom strategies to allow others libraries (Python
, C++
bindings...)
from surface_reconstruction import SurfaceStrategy, SurfaceReconstruction
# Create a class that inherit from "SurfaceStrategy"
class MyCustomSurface(SurfaceStrategy):
def __init__(self, my_custom_param: dict):
"""
Custom constructor with custom parameters
"""
super().__init__()
def load_file(self, file_path: str):
"""
Custom load point cloud file implementation here
"""
pass
def poisson_mesh(self, save_file=True, **params: {}):
"""
Generate the mesh file with faces/triangles here
"""
pass
# Register your custom strategy here
SurfaceReconstruction.register_type(MyCustomSurface)
# Pass a method/library that contains a Poisson algorithm implementation
surface = SurfaceReconstruction(
method_type='mycustom', # Don't pass the "surface" suffix
my_custom_param={'extra_config': 'some_value'},
)
# Call the method from the specific library, and export a mesh file
surface.poisson_mesh()
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
Hashes for surface_reconstruction-0.1.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 01385edec3317230b86c4e8b02f794185459ac6b30561d634462fc03410fa334 |
|
MD5 | 871058922ea173b0d49b86e94eace203 |
|
BLAKE2b-256 | 9e261c2852fdea74cc31868f800edf8a9c680966bd33416a03f584a2ca716f78 |