Skip to content

reading img files with rgdal

3 messages · Roger Bivand, Matthew Landis

Original
Matthew Landis
# 
Dear R-sig-Geo,

I'm trying to get started with rgdal, and not having much luck so far.  
I'm running R 2.5.0 on an i386-pc-mingw32 with rgdal 0.5-9 and sp 0.9-14. 

For starters, I'd like to read in an .img file created in ArcGIS 9.2. As 
you can see below, the file is a global map in geographic coordinates 
with 0.5 degree resolution.

 > GDALinfo('swe20070615.img')
Closing GDAL dataset handle 0x020d8660...  destroyed ... done.
rows        360
columns     720
bands       1
ll.x        -180
ll.y        90.0321
res.x       0.5
res.y       0.5
oblique.x   0
oblique.y   0
driver      HFA
projection  +proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs
file        swe20070615.img

The result of readGDAL is initially promising, but ultimately fails, 
apparently due to a problem with the projection.

 > test <- readGDAL('swe20070615.img')
swe20070615.img has GDAL driver HFA
and has 360 rows and 720 columns
Closing GDAL dataset handle 0x056988a8...  destroyed ... done.
Closing GDAL dataset handle (nil)... done.
Closing GDAL dataset handle 0x020dade8...  destroyed ... done.
Closing GDAL dataset handle 0x05699470...  destroyed ... done.
Error in `proj4string<-`(`*tmp*`, value = <S4 object of class "CRS">) :
        Geographical CRS given to non-conformant data
 > str(test)
Error in str(test) : object "test" not found

Can anyone suggest what might be going wrong and how to fix it?

In addition, I've had some trouble finding a good point of entry for how 
to use rgdal.  I have checked the standard documentation in the R help 
files and the GDAL websites, but having a bit of trouble penetrating it 
all.  If anyone can point me to a concise overview of how to use rgdal, 
I'd be most grateful.

Many thanks in advance,

Sincerely,

Matt Landis
Roger Bivand
#
On Wed, 11 Jul 2007, Matthew Landis wrote:

            

      
      
      
    

        
This is the problem, since > 90N is not defined for geographical 
coordinates, and leads to the projection being rejected. Why Arc wants to 
believe in > 90N is probably not known. If you can find out where the fuzz 
is coming from, perhaps you can remove it?  If correcting the data is not 
an option, please ask again, and a work-round may be possible (reading the 
data without its projection information, and placing it in a correct 
grid).

Roger

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Matthew Landis
    

    

      
      #
    

  


  

      
Dear Roger,

Many thanks for your rapid reply!  I guess I was so caught up in the 
details of rgdal that I completely failed to see that obvious problem.   
I've  managed to fix the coordinates, and now rgdal reads the image just 
fine. 

I'm still searching for a good overview of rgdal or gdal though - any 
pointers would be wonderful.

Sincerely,

Matt

        
Roger Bivand wrote: