=============
Introduction
=============

.. contents:: :local:

PCL
====
| The Point Cloud Library (PCL) is a standalone, large scale, open project for 2D/3D image and point cloud processing. PCL is released under the terms of the BSD license, and thus free for commercial and research use.
| It offers multiple advanced algorithms, written in C++ for efficiency, while offering a C++ and Python (while not giving access to all functions) interfaces.

Documentation and tutorials
============================
| Tutorials can be accessed under `PCL - Tutorials <https://pcl.readthedocs.io/projects/tutorials/en/latest/>`_.
| Documentation can be found under `PCL - Documentation <https://pointclouds.org/documentation/>`_.

Toolkit
========
| The toolkit is available under following :download:`zip file <doc/pcl_toolkit.zip>`.
| For SPL members, repo is `under here <https://gitlab.hevs.ch/SPL/vlspm/pcl_toolkit>`_.
| It contains multiple templated functions implementing PCL algorithms with simpler calls, offering a faster way for further development.
| To run it, you need to have PCL installed. Please refer to their documentation for this (install from precompiled binaries or build it from source).

Pipeline principle
===================
| The basic objective of this toolkit was to implement multiple techniques allowing to align clouds, knowing or not approximate position thanks to odometry, with or without a fixed reference that could be used for positioning the scans ...
| The idea is to go through a "pipeline", a recipe that allow both clouds to be quickly aligned with enough precision.
| It is as follow :

.. figure:: img/align_pipeline.*
      :align: center
      :alt: Registration pipeline
      :width: 450px

      Pipeline for registration, https://pcl.readthedocs.io/en/latest/registration_api.html#registration-api

| The chosen pipeline is the one presented in `"An efficient development of 3D surface registration by Point Cloud Library (PCL)" by Cheng-Tiao Hsieh <https://www.researchgate.net/publication/261128259_An_efficient_development_of_3D_surface_registration_by_Point_Cloud_Library_PCL>`_.

Other interesting papers can be found :

* `Registration with the Point Cloud Library <https://www.researchgate.net/publication/283198426_Registration_with_the_Point_Cloud_Library_-_A_Modular_Framework_for_Aligning_in_3-D>`_
* `Efficient Variants of the ICP Algorithm <https://www.cs.princeton.edu/~smr/papers/fasticp/fasticp_paper.pdf>`_
* `Robot bin picking : 3D pose retrieval based on Point Cloud Library <http://tesi.cab.unipd.it/42217/1/tesi_squizzato_11-12-2012.pdf>`_

:tag:`Algorithm`
:tag:`3D Algorithm`
:tag:`PCL`
:tag:`Point Cloud`