Geo Point Cloud Viewer

This project is an implementation of a novel technique for massive point cloud rendering on the Unity game engine. The project as a whole transforms a set of LAS files in any SRID recognized by the EPSG and generates a Unity scene in which the point cloud differentiates different point classes by color. It consists in two parts:

Model Generator

This Python program contains a preprocessing stage that structures the point cloud on a hierarchical structure designed for efficient out-of-core rendering.

The usage is defined as:

pc_model PC_MODEL_NAME [-h] [-d D] [-f F [F ...]] [-o O] [-e E] [-n N] [-s S]

optional arguments:

 -h, --help    show this help message and exit 
 -d D          Folder with LAS files inside
 -f F [F ...]  Paths to LAS files
 -o O          Path to output folder (default wd)
 -e E          EPSG num (default 32631)
 -n N          Max points per node (default 65000)
 -s S          Subsample PC in parent nodes (default True)

For example:

pc_model PC_MODEL_NAME -f las1.las las2.las -d path/to/las/folder -o path/to/out -p 32631 -n MAX_POINTS_NODE -s True

Unity Renderer

Unity project that loads a pre-set point cloud model and renders it following the double material strategy described in our work published at Eurographics 2019.

@article{santana2019visualization, title={Visualization of Large Point Cloud in Unity}, author={Santana N{'u}{~n}ez, Jos{'e} Miguel and Trujillo Pino, Agust{'\i}n Rafael and Ortega Trujillo, Sebasti{'a}n Eleazar}, journal={Eurographics technical report series}, year={2019} }

The Unity project contains a MonoBehaviour called PointCloudViewer which must reference to the folder created by the Model Generator in order to load the model. The PointCloudViewer also references a listener who handles point selection events. The technique shaders are prepared for Desktop platforms and has been tested on PC and Mac both for mono and stereoscopic rendering (VR).

See it in action on Youtube.