| landscan data [message #46955] |
Sun, 15 January 2006 07:45  |
matthias.demuzere
Messages: 32
Registered: January 2006
|
Member |
|
|
Dear all,
Is there anyone that has experience in reading the Landscan data in
IDL? The LandScanTM Dataset comprises a worldwide population database
compiled on a 30" X 30" latitude/longitude grid. The LandScan Dataset
files are available in an ESRI raster binary format.
Now I have all the data, but I really don´t succeed in reading in this
dataset. The data includes also a header file with the following
information:
ncols 43200
nrows 20880
xllcorner -180
yllcorner -90
cellsize 0.00833333333333
NODATA_value -9999
byteorder LSBFIRST
and a .prj file with following contents:
Projection GEOGRAPHIC
Datum WGS84
Zunits NO
Units DD
Spheroid WGS84
Xshift 0.0000000000
Yshift 0.0000000000
Parameters
Is their anybody who could help me with this problem?
Thanks,
Matthias
|
|
|
|
| Re: landscan data [message #47035 is a reply to message #46955] |
Tue, 17 January 2006 05:09  |
Paolo Grigis
Messages: 171
Registered: December 2003
|
Senior Member |
|
|
matthias.demuzere@geo.kuleuven.be wrote:
> Thanks for the reply David.
>
> I tried again to read in the whole dataset with 43200x20880 fields on
> another PC, but that didn't work either. So I probably have to look to
> see if I can expand my memory capacity....
The problem with the 32-bit versions of IDL is that no matter how much
*memory* you have in the system, you cannot allocate more than 2GB for *one*
array (the reason is likely due to the fact that internal array indexing
is done with a long signed integer, ranging up to 2^31-1).
So the maximal numbers of *elements* in one array is approximately
2^31/size, where size is the byte-size of your data type (1 for byte,
2 for int, 4 for longs etc.).
Possible solutions: break up the array in smaller chunks or switch
to 64-bit IDL versions.
Ciao,
Paolo
>
> Cheers,
> Matthias
>
|
|
|
|
| Re: landscan data [message #47036 is a reply to message #46955] |
Tue, 17 January 2006 03:18 |
Ben Panter
Messages: 102
Registered: July 2003
|
Senior Member |
|
|
matthias.demuzere@geo.kuleuven.be wrote:
> Thanks for the reply David.
>
> I tried again to read in the whole dataset with 43200x20880 fields on
> another PC, but that didn't work either. So I probably have to look to
> see if I can expand my memory capacity....
Wow!
That's one big array. I have a 2Gb Linux machine here, and can't get a
43200x20880 float array to fit in memory. I can do it with an integer
array, however it's not possible to do much with it afterwards.
You might want to go down the problem-splitting route after all.
cheers, and best of luck,
Ben
--
Ben Panter, Garching, Germany.
Email false, http://www.benpanter.co.uk
or you could try ben at ^^^^^^^^^^^^^^^
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
