> - We have breaklines that are roads. Consequently, there are
> intersections between breaklines. If it is a problem (as you seem to
> mention), how can we use roads as breaklines?
As long as polygonal line segments go not "cross" you are fine. They
can meet at any vertex but cannot cross. If you have crossing line
segments you need to insert a vertex at the point where they are
crossing.
> - The area containing breaklines is greater than the area of our .las
> file. What happens with breaklines located outside the area of
> our .las file?
Not sure. (-: Can you send me an example? If you have LAS tiles
generated with lastile you can use option '-use_tile_bb' to only
rasters the points and breaklines inside the tile. For the case you
describe I may have to add a '-use_bb' option.
> - When we make las2dem -i file.las -creeks breaklines.shp -step 2 -
> v, we receive warning messages saying that some breaklines form a
> closed loop. Are these warnings important? Are these breaklines
> ignored and breaklines that don't form a closed loop are used?
All breaklines are used. These are just warnings for control purposes.
The breakline functionality in las2dem is brand new and I want to have
all the information possible if there are issues. These warning will
disappear in a future version.
If you could send me some example data to look at the "breaklines
located outside the LAZ file" scenario that would be great.
Regards,
Martin @lastools
PS: Has anyone else tried using breaklines in las2dem yet? I'd be
curious to hear how the new functionality works for you.
-----
Hi Martin,
First of all, thank you for adding breakline capability into lastools.
I’m sure many people will find that useful. Since you asked in another
post that you were curious how the breaklines work for others, here’s
my take on it.
Also, feel free to post this message to the lastools user group – I
just didn’t want to put it there because of the attached zip file,
which contains the files I used for my little testing of the
breaklines functionality. I’ll try to explain below what I did with
the files.
I tried las2dem with:
1 no breakline
2 a polygon (lake) breakline
3 polygon (lake) and line (creek) breakline
4 the same as (3) but using a denser (more vertices) creek breakline
The commands I used are as following:
las2dem -i uw.las -o 01_nobreakline.tif -v -step 0.5 -kill 50
las2dem -i uw.las -o 02_lakebreakline.tif -v -lakes water_poly_z.shp
-step 0.5 -kill 50
las2dem -i uw.las -o 03_lake_creek_breakline.tif -v -lakes
water_poly_z.shp -creeks creek_line_z.shp -step 0.5 -kill 50
las2dem -i uw.las -o 04_lake_creekdense_breakline.tif -v -lakes
water_poly_z.shp -creeks creek_line_dense.shp -step 0.5 -kill 50
I found it interesting about the difference in the output tiffs for
the ‘densified’ creek line compared to the original creek line. I
would have assumed that it wouldn’t really matter how many vertices
are used, since the actual line doesn’t change and the z-values for
the additional vertices are linearly interpolated between the original
points, which I would have assumed is handled that way anyways when
using the original breakline to generate a raster that has a pixel
size shorter than the length between two of the original line
vertices. Maybe I’m missing something here or don’t use the breaklines
quite ‘right’. Just curious if you could shed any light on this.
[[ ANSWER: You compute a TIN of all LiDAR returns. For the creek line
no LiDAR points are removed. Hence you merely "cut" a tiny corridor
into the dense vegetation TIN. The more points you put down the more
chances there are during the TIN rasterization to hit one of these
corridor points. You should really re-run the above with the
'-keep_class 2' option and work on the terrain only. ]]
Cheers,
Mike
_______________________________
Mike Lackner, MSc.
Remote Sensing Specialist
Mapping, Analysis & Design
Faculty of Environment
University of Waterloo
mlac...@uwaterloo.ca
Tel: 519-888-4567, ext. 36563
Fax: 519-888-4325