Pointcloud tiling software for .e57/.pts files

[Inception]

I'm searching for a software suite/plugin that provides a simple method for pointclouds to be segmented by 'tiles'.

Some co-workers of mine have used lastool (lasclip) for previous mobile scanning data, however this only supports .las/.laz files. We are looking for an equivalent program that can deal with a wider variety of point cloud formats (.e57/.pts)



Any help or points in a certain direction would be appreciated

[jedfrechette]

PDAL works great for this:

https://pdal.io/apps/tile.html

It supports las/laz, e57, pts, as well as a bunch of other formats.

[rollorigby]

FME also does this quite well across a number of formats

[smacl]

SCC has a wide range of tiling options and works with all open formats. This includes tiling and filtering based on grids, volumes (3d grids), chainage / offset ranges (e.g. distance along roads, rails, etc..), point count and arbitrary input polygons.

[landmeterbeuckx]

forgive my ignorance but what's the advantage of tiling?

[smacl]

forgive my ignorance but what's the advantage of tiling?

A few that come to mind;

• For very large models, such as road and rail, it enables multiple people to be working on the deliverable at one time. So for example, if you'd captured a 40k road network with a Pegasus system, your could give 4 operators 10k each and be able deliver the final job much faster.

• Huge point cloud, e.g. 10s or 100s of billions of points, may very slow or impossible to process on some hardware and software configurations. In this scenario, tiling allows you to work with very large datasets on modest hardware.

• Many geographical processing algorithms have a complexity greater than order n or O(n), e.g. O(n log n) or even O(n * n). If you throw a huge amount of unsegmented data at such algorithms, they will take an unacceptably long time to complete.

• Tiled datasets can be faster to load, save and archive.

• In very large areas, projects can be tiled for geodetic reasons, where each tile has an effective scale factor of 1.0 for engineering purposes and its own parameters to move between cartesian and geodetic coordinates. The London Survey Grid would be an example of this. High flown national mapping LAS data can be sparse and cover huge areas

[landmeterbeuckx]

forgive my ignorance but what's the advantage of tiling?

A few that come to mind;

• For very large models, such as road and rail, it enables multiple people to be working on the deliverable at one time. So for example, if you'd captured a 40k road network with a Pegasus system, your could give 4 operators 10k each and be able deliver the final job much faster.

• Huge point cloud, e.g. 10s or 100s of billions of points, may very slow or impossible to process on some hardware and software configurations. In this scenario, tiling allows you to work with very large datasets on modest hardware.

• Many geographical processing algorithms have a complexity greater than order n or O(n), e.g. O(n log n) or even O(n * n). If you throw a huge amount of unsegmented data at such algorithms, they will take an unacceptably long time to complete.

• Tiled datasets can be faster to load, save and archive.

• In very large areas, projects can be tiled for geodetic reasons, where each tile has an effective scale factor of 1.0 for engineering purposes and its own parameters to move between cartesian and geodetic coordinates. The London Survey Grid would be an example of this. High flown national mapping LAS data can be sparse and cover huge areas

Thanks Shane.

So you can select the tile to load in case of a road project for exemple and then the data itself will be loaded on the machine?

[smacl]

forgive my ignorance but what's the advantage of tiling?

A few that come to mind;

• For very large models, such as road and rail, it enables multiple people to be working on the deliverable at one time. So for example, if you'd captured a 40k road network with a Pegasus system, your could give 4 operators 10k each and be able deliver the final job much faster.

• Huge point cloud, e.g. 10s or 100s of billions of points, may very slow or impossible to process on some hardware and software configurations. In this scenario, tiling allows you to work with very large datasets on modest hardware.

• Many geographical processing algorithms have a complexity greater than order n or O(n), e.g. O(n log n) or even O(n * n). If you throw a huge amount of unsegmented data at such algorithms, they will take an unacceptably long time to complete.

• Tiled datasets can be faster to load, save and archive.

• In very large areas, projects can be tiled for geodetic reasons, where each tile has an effective scale factor of 1.0 for engineering purposes and its own parameters to move between cartesian and geodetic coordinates. The London Survey Grid would be an example of this. High flown national mapping LAS data can be sparse and cover huge areas

Thanks Shane.

So you can select the tile to load in case of a road project for exemple and then the data itself will be loaded on the machine?

Hi Lieven,

Yes, a tile is essentially just another point cloud which is a subset of a larger point cloud. Sometimes you use different tiles on different PCs, other times on the same PC to overcome speed and memory limitations. In SCC in addition to automated tiling on import, we also allow creation of tiles on the fly, which is a bit like having a persistent limit box and navigation tools to interactively change tiles,

Shane