[R-sig-Geo] as.raster() does not work for RImageJ objects

[Agustin Lobo]

Robert,

Had tried that, sorry I forgot reporting it:
logo <- system.file( "images", "R.jpg", package = "RImageJ" )
image = IJ$openImage( logo )
r2 <- raster(image)

Error in (function (classes, fdef, mtable)  :
  unable to find an inherited method for function ‘raster’ for
signature ‘"jobjRef"’

The workaround I posted is not bad in my current case, in particular because
ImageJ does not deal with image files of >3 bands (AFAIK), and because
I have to process
each band independently anyway (in this particular application).

A much better solution would be connecting the raster package to the
OTB, ITk or openCV
libraries for image processing. I think I'll be able to do some
testing with OTB from within R,
as there is a python API and R and python talk to each other fluently.
There is also some work on integrating ITk within R
http://www.itk.org/Wiki/Proposals:Integration_with_R_Language
which should be useful for doing the same for OTB (which is sort of
ITk derivate for Remote Sensing imagery)
but have not tried myself.

Agus

On Mon, Jan 20, 2014 at 10:15 PM, Robert J. Hijmans <r.hijmans at gmail.com> wrote:
> Agus,
>
> as.raster returns a 'raster' object as defined in the graphics package
> (or thereabouts). I assume that what you want is to coerce to a
> RasterLayer (from the raster package). That is typically done with
>
> raster(image)
> # or
> as(image, 'RasterLayer')
>
> Robert
>
>
>
> On Fri, Jan 17, 2014 at 5:45 AM, Agustin Lobo <alobolistas at gmail.com> wrote:
>> ok, thanks. I did not know that there was another class raster in
>> another package.
>>
>> This means that we cannot convert from RImageJ objects to
>> raster::raster objects.
>> But we can use RImageJ to process, save to disk and read back as raster:raster:
>>
>> logo <- system.file( "images", "R.jpg", package = "RImageJ" )
>> image = IJ$openImage( logo )
>> image$show()
>> IJ$run( "8-bit" )
>> IJ$run( "Invert" )
>> IJ$save( "bw-google.gif" )
>> image$close()
>>
>> r4 <- raster("bw-google.gif")
>> plot(r4)
>> summary(r4)
>> extent(r4)
>>
>> Hopefully I'll be able to use RImageJ for applying smoothing filters
>> and other image processing
>> methods that are too slow within R, without having to leave the R
>> scripting environment.
>>
>> Agus
>>
>> On Fri, Jan 17, 2014 at 9:26 AM, Michael Sumner <mdsumner at gmail.com> wrote:
>>> You are see a clash of the grDevices::as.raster with the raster
>>> package. grDevices is loaded by default, and has a (S3) method for
>>> whatever your image is (a matrix?).
>>>
>>> It seems like the S4 raster::as.raster should allow these to "pass
>>> through" to the underlying S3 method.
>>>
>>> I can't look any closer right now
>>>
>>>
>>>
>>>
>>>
>>> On Fri, Jan 17, 2014 at 7:09 PM, Agustin Lobo <alobolistas at gmail.com> wrote:
>>>> Actually, in OSX, there is no error message at using as.raster(), but
>>>> it does not work either:
>>>>
>>>> r <- as.raster(image)
>>>> class(r)
>>>>
>>>> [1] "raster"
>>>>> str(r)
>>>> 'raster' chr [1:76, 1:100] "#ffffff" "#ffffff" "#ffffff" "#ffffff"
>>>> "#ffffff" "#ffffff" "#ffffff" ...
>>>>
>>>> which is not the structure a raster object should have.
>>>>
>>>> Then:
>>>>
>>>> r <- as.raster.jobjRef(image)
>>>> Error: could not find function "as.raster.jobjRef"
>>>>
>>>> while ?as.raster.jobjRef shows the help page
>>>> the function as.raster.jobjRef" is not in the workspace (and this is
>>>> also the case in linux):
>>>>> search()
>>>>  [1] ".GlobalEnv"        "package:RImageJ"   "package:rJava"
>>>> "package:raster"    "package:sp"
>>>>  [6] "tools:rstudio"     "package:stats"     "package:graphics"
>>>> "package:grDevices" "package:utils"
>>>> [11] "package:datasets"  "package:methods"   "Autoloads"
>>>> "package:base"
>>>>> ls(2)
>>>> [1] "IJ"              "IJWindowManager" "ImageJ"
>>>>
>>>>
>>>> R version 3.0.2 (2013-09-25)
>>>> Platform: x86_64-apple-darwin10.8.0 (64-bit)
>>>>
>>>> locale:
>>>>   [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
>>>>
>>>> attached base packages:
>>>>   [1] stats     graphics  grDevices utils     datasets  methods   base
>>>>
>>>> other attached packages:
>>>>   [1] RImageJ_0.2-146 rJava_0.9-6     raster_2.1-66   sp_1.0-13
>>>>
>>>> loaded via a namespace (and not attached):
>>>>   [1] grid_3.0.2      lattice_0.20-23 tools_3.0.2
>>>>
>>>>
>>>> Agus
>>>>
>>>> On Thu, Jan 16, 2014 at 4:09 PM, Agustin Lobo <alobolistas at gmail.com> wrote:
>>>>> Hi!
>>>>>
>>>>> While the following works fine:
>>>>> install.packages("/home/alobo/Downloads/RImageJ_0.2-146.tar.gz", repos
>>>>> = NULL, type="source")
>>>>> require(RImageJ)
>>>>> require(raster)
>>>>> logo <- system.file( "images", "R.jpg", package = "RImageJ" )
>>>>> image = IJ$openImage( logo )
>>>>> class(image)
>>>>> plot(c(100, 250), c(300, 450), type = "n", xlab="", ylab="")
>>>>> rasterImage(image, 100, 300, 150, 350, interpolate=FALSE)
>>>>>
>>>>> (note that the example in
>>>>> http://romainfrancois.blog.free.fr/index.php?category/R-package/RImageJ
>>>>> is currently wrong, it must be rasterImage() and not raster()
>>>>>
>>>>> But as.raster() (the most important for me) does not work:
>>>>>
>>>>>> class(image)
>>>>> [1] "jobjRef"
>>>>> attr(,"package")
>>>>> [1] "rJava"
>>>>>> r <- as.raster(image)
>>>>> Error in as.raster.jobjRef(image) : attempt to apply non-function
>>>>>
>>>>> see ?as.raster.jobjRef
>>>>>
>>>>> Any fix?
>>>>>
>>>>> Thanks
>>>>> Agus
>>>>
>>>> _______________________________________________
>>>> R-sig-Geo mailing list
>>>> R-sig-Geo at r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>>
>>>
>>>
>>> --
>>> Michael Sumner
>>> Hobart, Australia
>>> e-mail: mdsumner at gmail.com
>>>
>>> _______________________________________________
>>> R-sig-Geo mailing list
>>> R-sig-Geo at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo