Re: [R-sig-Geo] LiDAR LAS file import into R via RSAGA

[Chris Reudenbach]

Hi Clay,

RSAGA provides a comfortable platform independent system wrapper for 
SAGA but as stated  no two-way bindings.


If you want to make use of SAGA algorithms () it is IMHO the most 
straightforward way to import and export the raster files using the e.g. 
gdalUtils. If you use raster() directly please note that it may produce 
problems writing SAGA files directly. To avoid format complications  I 
use  most of the time GTiff. You can easily integrate the power of 
rLidar in this workaround concept.

A typical work flow may look like:


 # gdalwarp is used to convert the data format to SAGA
 # you may assign the projection information to the data (like example 
or change it with t_srs=)
  gdalwarp(file.in, file.out, overwrite=TRUE, 
s_srs=paste0('EPSG:',epsg.code), of='SAGA')


 # do something in SAGA using the files on the disk storage as "interface"
 # note SAGA works much more stable under differnet platforms when you 
provide the env settings


  # e.g. calculate wood's terrain indices   wood= 
1=planar,2=pit,3=channel,4=pass,5=ridge,6=peak

rsaga.geoprocessor('ta_morphometry',"Morphometric Features",env=myenv,
   list(DEM='mp_dem.sgrd',
FEATURES='mp_wood.sgrd',
SLOPE='mp_slope.sgrd',
LONGC='mp_longcurv.sgrd',
CROSC='mp_crosscurv.sgrd',
MINIC='mp_mincurv.sgrd',
MAXIC='mp_maxcurv.sgrd',
TOL_SLOPE=14.00
TOL_CURVE=0.1
EXPONENT=0.00
ZSCALE=1.00))

# read whatever you wand directly using raster()
peak.area<-raster('mp_wood.sdat')

For using SAGA and/or GRASS in a "mixed" environment with R you may find 
some help at our course pages:

 
http://moc.environmentalinformatics-marburg.de/doku.php?id=courses:msc:advanced-gis:lecture-notes:ag-ln-06

cheers Chris


Am 12.02.2016 um 07:51 schrieb Clay S:

Thanks Michael Sumner

rLidar works well.  But how do I pass the lidar data to another SAGA module
within R without saving the output of each step as a file in a whole series
of SAGA module calls from R?

I would like something like the dplyr %>% pipe command.

Thanks
Clay

On Fri, Feb 12, 2016 at 7:44 PM, Michael Sumner  wrote:


Try rLidar::readLAS as a straightforward alternative, it's not bullet
proof but works fine for a lot of examples. Happy to help extend it if your
format is not supported.

Cheers, Mike

On Fri, 12 Feb 2016 7:41 am Clay S  wrote:


Hi

Is it possible to import a Lidar LAS file into R with the
rsaga.geoprocessor function of the RSAGA package.  The following code runs
without an error, but all I get in R is the SAGA console output.

The R code below does save a SAGA point cloud file called "test" in my
working directory.  If I delete the `POINTS = "test"` parameter, then
nothing happens as expected.  However, I want to work with the data in R,
and pass the point cloud to other SAGA modules. This can be done in SAGA
without having to save the point cloud.  Can the same be done from R?

How can I get the point cloud imported into the R environment?  Do I need
to coerce it into a matrix, data frame, or Spatial Points object?


R code:

###

library(RSAGA)


# Set RSAGA environment.

env <- rsaga.env()


# SET WORKING DIRECTORY FIRST!!

dir <- getwd()


# Get module parameters

lasfilelist <- list.files(dir, pattern =".las$", full.names=TRUE,
recursive=FALSE)


# send command to RSAGA

test <- rsaga.geoprocessor(lib = "io_shapes_las", module = "Import LAS
Files",
param = list(FILES = lasfilelist[1], POINTS = "test",
 T=TRUE, i=TRUE, a=FALSE, r=TRUE, c=FALSE,
u=FALSE,
 n=TRUE, R=FALSE, G=FALSE, B=FALSE,
e=FALSE,
d=FALSE,
 p=FALSE, C=FALSE, VALID=FALSE,
RGB_RANGE=FALSE),
env = env)

###


Here is the console output (the test variable stores the same thing):

###

library path: C:\OSGEO4~1\apps\saga\modules\io_shapes_las.dll
library name: Import/Export - LAS
tool name   : Import LAS Files
author  : O. Conrad, V. Wichmann (c) 2009

Parameters

Input Files: "C:/Users/whcms0/'.CPT/Tarewa/Lidar/AX300711.las"
Point Clouds: No objects
gps-time: yes
intensity: yes
scan angle: no
number of the return: yes
classification: no
user data: no
number of returns of given pulse: yes
red channel color: no
green channel color: no
blue channel color: no
edge of flight line flag: no
direction of scan flag: no
point source ID: no
rgb color: no
Check Point Validity: no
R,G,B value range: 16 bit

Parsing AX300711.las ...
Save point cloud: test...




Here is my env:




env

$workspace

[1] "."


$cmd

[1] "saga_cmd.exe"


$path

[1] 

Re: [R-sig-Geo] LiDAR LAS file import into R via RSAGA

[Michael Treglia]

In case it's of any use to others, I've been scripting some LiDAR
processing from R using LASTools, focused on dealing with many individual
las files.  At this point we've just been extracting a bunch of metadata
fields to get a sense of what we have, fixing some metadata errors that
we're aware of, and reprojecting data. We've got lots of data for a region
that are in various projections, and sometimes the LAS metadata are just
incorrect, so the functions as written are designed to help us deal with
those issues, but can easily be adapted for other purposes.

These functions use foreach and doparallel to use multi-core processors to
speed things along, running LasTools via system commands.  This work is
still in fairly early stages and admittedly needs some better organization
and documentation.  At this point, we've exclusively been calling on
LasTools functions that do not require a paid license, and plan to keep it
as such for now.

Also, we were exploring use on both linux and windows systems on an
external server, and the linux setups have been a lot faster, so functions
ending in unix.R are the most up to date [the changes to go back to windows
are pretty simple].  (as I said, we need to organize, but it is functional
for our purposes now...).

The code is available in a github repository:
https://github.com/mltConsEcol/OK_LiDAR_Processing

Feel free to contact me with questions on this code, suggestions, or want
to contribute any additions (and I'd be curious to hear if anybody does end
up using it). I think some future functions will call on the rLidar
package, and we'll see where our next steps take us...

Best,
mike






On Fri, Feb 12, 2016 at 2:32 AM, Chris Reudenbach  wrote:

> Hi Clay,
>
> RSAGA provides a comfortable platform independent system wrapper for SAGA
> but as stated  no two-way bindings.
>
> If you want to make use of SAGA algorithms () it is IMHO the most
> straightforward way to import and export the raster files using the e.g.
> gdalUtils. If you use raster() directly please note that it may produce
> problems writing SAGA files directly. To avoid format complications  I use
> most of the time GTiff. You can easily integrate the power of rLidar in
> this workaround concept.
> A typical work flow may look like:
>
>
>  # gdalwarp is used to convert the data format to SAGA
>  # you may assign the projection information to the data (like example or
> change it with t_srs=)
>   gdalwarp(file.in, file.out, overwrite=TRUE,
> s_srs=paste0('EPSG:',epsg.code), of='SAGA')
>
>  # do something in SAGA using the files on the disk storage as "interface"
>  # note SAGA works much more stable under differnet platforms when you
> provide the env settings
>
>   # e.g. calculate wood's terrain indices   wood=
> 1=planar,2=pit,3=channel,4=pass,5=ridge,6=peak
> rsaga.geoprocessor('ta_morphometry',"Morphometric Features",env=myenv,
>list(DEM='mp_dem.sgrd',
> FEATURES='mp_wood.sgrd',
> SLOPE='mp_slope.sgrd',
> LONGC='mp_longcurv.sgrd',
> CROSC='mp_crosscurv.sgrd',
> MINIC='mp_mincurv.sgrd',
> MAXIC='mp_maxcurv.sgrd',
> TOL_SLOPE=14.00
> TOL_CURVE=0.1
> EXPONENT=0.00
> ZSCALE=1.00))
>
> # read whatever you wand directly using raster()
> peak.area<-raster('mp_wood.sdat')
>
> For using SAGA and/or GRASS in a "mixed" environment with R you may find
> some help at our course pages:
>
> http://moc.environmentalinformatics-marburg.de/doku.php?id=courses:msc:advanced-gis:lecture-notes:ag-ln-06
>
> cheers Chris
>
>
>
> Am 12.02.2016 um 07:51 schrieb Clay S:
>
>> Thanks Michael Sumner
>>
>> rLidar works well.  But how do I pass the lidar data to another SAGA
>> module
>> within R without saving the output of each step as a file in a whole
>> series
>> of SAGA module calls from R?
>>
>> I would like something like the dplyr %>% pipe command.
>>
>> Thanks
>> Clay
>>
>> On Fri, Feb 12, 2016 at 7:44 PM, Michael Sumner 
>> wrote:
>>
>> Try rLidar::readLAS as a straightforward alternative, it's not bullet
>>> proof but works fine for a lot of examples. Happy to help extend it if
>>> your
>>> format is not supported.
>>>
>>> Cheers, Mike
>>>
>>> On Fri, 12 Feb 2016 7:41 am Clay S  wrote:
>>>
>>> Hi

 Is it possible to import a Lidar LAS file into R with the
 rsaga.geoprocessor function of the RSAGA package.  The following code
 runs
 without an error, but all I get in R is the SAGA console output.

 The R code below does save a SAGA point cloud file called "test" in my
 working directory.  If I delete the `POINTS = "test"` parameter, then
 nothing happens as expected.  However, I want to 

Re: [R-sig-Geo] LiDAR LAS file import into R via RSAGA

[Clay S]

Another option would be to create a SAGA tool chain XML file following the
tutorials by Michael Bock, Olaf Conrad, Volker Wichmann, and drag that XML
file into the SAGA module library.  Does anyone know if you can access your
own SAGA tool chains with the RSAGA::rsaga.geoprocessor command?  I suppose
the parameters for each module in the chain would be set according to the
XML file.  If that is the case, then I also suppose you could edit the XML
file with R before you run the tool chain.  But I am getting ahead of
myself.

Unfortunately, the most recent version of SAGA that the OSGeo4W is 2.1.2-1,
which does not have the tool chains feature. That was added in SAGA 2.1.3.
Does anyone know when OSGeo4W will get updated?





On Fri, Feb 12, 2016 at 7:51 PM, Clay S  wrote:

> Thanks Michael Sumner
>
> rLidar works well.  But how do I pass the lidar data to another SAGA
> module within R without saving the output of each step as a file in a whole
> series of SAGA module calls from R?
>
> I would like something like the dplyr %>% pipe command.
>
> Thanks
> Clay
>
> On Fri, Feb 12, 2016 at 7:44 PM, Michael Sumner 
> wrote:
>
>> Try rLidar::readLAS as a straightforward alternative, it's not bullet
>> proof but works fine for a lot of examples. Happy to help extend it if your
>> format is not supported.
>>
>> Cheers, Mike
>>
>> On Fri, 12 Feb 2016 7:41 am Clay S  wrote:
>>
>>> Hi
>>>
>>> Is it possible to import a Lidar LAS file into R with the
>>> rsaga.geoprocessor function of the RSAGA package.  The following code
>>> runs
>>> without an error, but all I get in R is the SAGA console output.
>>>
>>> The R code below does save a SAGA point cloud file called "test" in my
>>> working directory.  If I delete the `POINTS = "test"` parameter, then
>>> nothing happens as expected.  However, I want to work with the data in R,
>>> and pass the point cloud to other SAGA modules. This can be done in SAGA
>>> without having to save the point cloud.  Can the same be done from R?
>>>
>>> How can I get the point cloud imported into the R environment?  Do I need
>>> to coerce it into a matrix, data frame, or Spatial Points object?
>>>
>>>
>>> R code:
>>> ###
>>> library(RSAGA)
>>> # Set RSAGA environment.
>>> env <- rsaga.env()
>>>
>>> # SET WORKING DIRECTORY FIRST!!
>>> dir <- getwd()
>>>
>>> # Get module parameters
>>> lasfilelist <- list.files(dir, pattern =".las$", full.names=TRUE,
>>> recursive=FALSE)
>>>
>>> # send command to RSAGA
>>> test <- rsaga.geoprocessor(lib = "io_shapes_las", module = "Import LAS
>>> Files",
>>>param = list(FILES = lasfilelist[1], POINTS = "test",
>>> T=TRUE, i=TRUE, a=FALSE, r=TRUE, c=FALSE,
>>> u=FALSE,
>>> n=TRUE, R=FALSE, G=FALSE, B=FALSE,
>>> e=FALSE,
>>> d=FALSE,
>>> p=FALSE, C=FALSE, VALID=FALSE,
>>> RGB_RANGE=FALSE),
>>>env = env)
>>> ###
>>> Here is the console output (the test variable stores the same thing):
>>> ###
>>> library path: C:\OSGEO4~1\apps\saga\modules\io_shapes_las.dll
>>> library name: Import/Export - LAS
>>> tool name   : Import LAS Files
>>> author  : O. Conrad, V. Wichmann (c) 2009
>>>
>>> Parameters
>>>
>>> Input Files: "C:/Users/whcms0/'.CPT/Tarewa/Lidar/AX300711.las"
>>> Point Clouds: No objects
>>> gps-time: yes
>>> intensity: yes
>>> scan angle: no
>>> number of the return: yes
>>> classification: no
>>> user data: no
>>> number of returns of given pulse: yes
>>> red channel color: no
>>> green channel color: no
>>> blue channel color: no
>>> edge of flight line flag: no
>>> direction of scan flag: no
>>> point source ID: no
>>> rgb color: no
>>> Check Point Validity: no
>>> R,G,B value range: 16 bit
>>>
>>> Parsing AX300711.las ...
>>> Save point cloud: test...
>>>
>>> 
>>> Here is my env:
>>> 
>>> > env
>>> $workspace
>>> [1] "."
>>> $cmd
>>> [1] "saga_cmd.exe"
>>> $path
>>> [1] "C:\\OSGeo4W64\\apps\\saga"
>>> $modules
>>> [1] "C:\\OSGeo4W64\\apps\\saga/modules"
>>> $version
>>> [1] "2.1.2"
>>> $cores
>>> [1] NA
>>> $parallel
>>> [1] FALSE
>>> $lib.prefix
>>> [1] ""
>>> #
>>>
>>> [[alternative HTML version deleted]]
>>> ___
>>> R-sig-Geo mailing list
>>> R-sig-Geo@r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>>
>> --
>> Dr. Michael Sumner
>> Software and Database Engineer
>> Australian Antarctic Division
>> 203 Channel Highway
>> Kingston Tasmania 7050 Australia
>>
>>
>

[[alternative HTML version deleted]]

___
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo