LasTile: Tile Size

[Brian]

I am working with a large las file (757+ mil points) and when using lastile a tile size needs to be specified.  My question, is there an easy way to decide what the tile size should be without having to run the tool till you get all your tiles down in size that is usable in other tools such as lasclassify?  I have included a screenshot of lasinfo results the las file I am tiling.

Thanks,

Brian A. Ritter RF,CA,GISP,PGISS

Forestry PhD Candidate

Graduate Teaching Assistant

Clemson University

Department of Forestry and Environmental Conservation

[Terje Mathisen]

Brian wrote:
> I am working with a large las file (757+ mil points) and when using
> lastile a tile size needs to be specified. My question, is there an
> easy way to decide what the tile size should be without having to run
> the tool till you get all your tiles down in size that is usable in
> other tools such as lasclassify? I have included a screenshot of
> lasinfo results the las file I am tiling.

From that lasinfo dump it looks like your points fit into a rectangle
of about 4.3 x 2.3 km, or about 10 sq km, right?

This means that you have an average of 76 returns per square meter! Is
this a terrestrial scan or just a _very_ dense heli capture?

If you want to be able to use the free tools (assuming this isn't a
commercial venture) you should aim at about 1 M points/tile, or a tile
size of maximum 114 m.

I would try to use 100x100 m tiles with a 5 m buffer zone, for a total
of 110x110 = 12100 sq m per tile!

Terje

>
> Thanks,
>
> Brian A. Ritter RF,CA,GISP,PGISS
>
> Forestry PhD Candidate
>
> Graduate Teaching Assistant
>
> Clemson University
>
> Department of Forestry and Environmental Conservation
>
>
>

> --
> Download LAStools at
> http://lastools.org
> http://rapidlasso.com
> Be social with LAStools at
> http://facebook.com/LAStools
> http://twitter.com/LAStools
> http://linkedin.com/groups/LAStools-4408378
> Manage your settings at
> http://groups.google.com/group/lastools/subscribe


--
- <Terje.M...@tmsw.no>
"almost all programming can be viewed as an exercise in caching"

[Brian]

Thanks for the reply, I have a licensed version of lastools and wound up using a 250 tile size with a 10 m buffer, however that was after I ran it at 750, 500, and 300 to try and get <20 mil pts per tile.  My study boundary is slightly over 7 sq km so it would fall into your estimated area..  The actual point density (156/sqm) I am getting is pretty high due to the cell resolution (2.5 cm).  I understand your logic here and it makes sense to look at the point density across the boundary to determine tile size.  How did you determine using a 5 m buffer?

[Terje Mathisen]

Brian wrote:
> Thanks for the reply, I have a licensed version of lastools and wound
> up using a 250 tile size with a 10 m buffer, however that was after I
> ran it at 750, 500, and 300 to try and get <20 mil pts per tile. My
> study boundary is slightly over 7 sq km so it would fall into your
> estimated area.. The actual point density (156/sqm) I am getting is
> pretty high due to the cell resolution (2.5 cm). I understand your
> logic here and it makes sense to look at the point density across the
> boundary to determine tile size. How did you determine using a 5 m
> buffer?

Just a reasonable percentage of the suggested tile size to avoid most
boundary artifacts. With your densities you don't need a big buffer!

I use 250+35 for my own much more sparse projects.

Terje

>
> On Wednesday, April 22, 2015 at 2:59:38 PM UTC-4, Brian wrote:
>
> I am working with a large las file (757+ mil points) and when
> using lastile a tile size needs to be specified. My question, is
> there an easy way to decide what the tile size should be without
> having to run the tool till you get all your tiles down in size
> that is usable in other tools such as lasclassify? I have
> included a screenshot of lasinfo results the las file I am tiling.
>
> Thanks,
>
> Brian A. Ritter RF,CA,GISP,PGISS
>
> Forestry PhD Candidate
>
> Graduate Teaching Assistant
>
> Clemson University
>
> Department of Forestry and Environmental Conservation
>
>
>

[Brian]

Thanks, one more question, do you feel that a buffer is needed in all cases?

[Terje Mathisen]

Brian wrote:
> Thanks, one more question, do you feel that a buffer is needed in all
> cases?

Depends on what you are doing, but if you want anything like a smooth,
continuous cover over the entire area after final processing, then a
buffer is obviously needed.

The main thing now is that since about 1997, Moore's law have stopped
for single-threaded operations: You really need to run as many parallel
processes as there are available cores, and buffered tiles is the
best/most obvious way to do this. :-)

Terje

> - <Terje.M...@tmsw.no <javascript:>>