Welcome to the fascinating point cloud world! If you like Python, contributions are welcome at https://github.com/daavoo/pyntcloud . /end recruiting.
What are the most commonly used point cloud file formats?
- General pourpose 3D file formats
Well ... almost any general pourpose 3D file format implicitly supports point clouds because those formats usually store a bunch of polygons, and in order to define a polygon you must define it's vertices.
If you think carefully, those vertices are just point clouds.
A good example of this is the .ply file format. Wich is a common choice to export point clouds in many photogrammetry (one of the ways to generate point clouds from the real world) softwares:
Other common formats of this type are .obj , .stl , .off , and more.
This formats are designed for storing point clouds + metadata (mostly geospatial stuff).
Probably the most common is .las .
Many Lidar manufactures also have their own privative format.
If someone is into robotics, he will be into ROS.
If someone is into ROS and point clouds, he will use sensor_msgs/PointCloud2.
If someone uses sensor_msgs/PointCloud2, he will convert it to .pcd because PCL is the de facto library for working with point clouds in ROS.
In addition to all this, you can store the point cloud information in a plain text file separating the columns with comma, tab, spaces or wathever and name it .xyz, .pts or something like that. There is no standard and people just do whatever they like. ASCII formats are slow to read and memmory inefficient for big point clouds.
Which point cloud file formats are supported through PCL?
Many of the above.
http://docs.pointclouds.org/trunk/group__io.html
Is there any collections of point cloud files available? Ideally, XYZRGB to play with.
You can use one of the above mentioned photogrammetry softwares to generate point clouds yourself with just a camera. It's a lot of fun.
You can also use some 3D modeling software like Blender and then generate a point cloud from the exported mesh. I wrote a post about it:
https://medium.com/@daviddelaiglesiacastro/3d-point-cloud-generation-from-3d-triangular-mesh-bbb602ecf238
https://github.com/PointCloudLibrary/data
https://people.sc.fsu.edu/~jburkardt/data/ply/ply.html
http://opentopo.sdsc.edu/datasets?listAll=true
Spain has a national project with lidar data for all the country free to download:
http://centrodedescargas.cnig.es/CentroDescargas/index.jsp
Maybe other countries have something similar.
