comp.lang.idl-pvwave archive: archive » non-rectangular array subset

Home » Public Forums » archive » non-rectangular array subset

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

non-rectangular array subset [message #89446]

Fri, 17 October 2014 04:25

jecca.baker
Messages: 12
Registered: October 2014

Junior Member

Hello,

I have an array of global data (MODIS LAI) and need to select data from the Amazon basin only. Is there a way to do this by selecting a number of rectangular regions and concatenating them if the rectangular regions are different sizes? Eg, is it possible to concatenate the following?

0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0
0,0,0,0,0,0,0,0

0,0,0,0,0,0
0,0,0,0,0,0

0,0,0
0,0,0

Or for a more sophisticated selection, is there a way to only select data that falls under an amazon basin shapefile?

Any tips would be much appreciated.
Many thanks,
Jess

Report message to a moderator

Re: non-rectangular array subset [message #89447 is a reply to message #89446]

Fri, 17 October 2014 04:36

Fabzi
Messages: 305
Registered: July 2010

Senior Member

Hi,

On 17.10.2014 13:25, jecca.baker@gmail.com wrote:
> is there a way to only select data that falls under an amazon basin shapefile?

IDLanROI is a tool designed for this purpose. You'll have to transform
the shape coordinates into your grid coordinates and then use IDLanROI:

http://www.exelisvis.com/docs/IDLanROI.html

Fabien

Report message to a moderator

Re: non-rectangular array subset [message #89448 is a reply to message #89446]

Fri, 17 October 2014 04:39

Fabzi
Messages: 305
Registered: July 2010

Senior Member

I forgot to answer to your first question about concatenation. Once you
have non-regular subsets you have to work with masks and where() to
access your data.

(e.g.: https://www.idlcoyote.com/ip_tips/xroi.html)

On 17.10.2014 13:25, jecca.baker@gmail.com wrote:
> Eg, is it possible to concatenate the following?

Report message to a moderator

Re: non-rectangular array subset [message #89449 is a reply to message #89448]

Fri, 17 October 2014 05:22

jecca.baker
Messages: 12
Registered: October 2014

Junior Member

Hi Fabien.
Thanks for your tips. Could you possibly expand on how to use masks and where for subsetting? My data isn't an image file, rather a large array containing a grid of data every month for several years, so not sure how I would get XROI to work in the way David Fanning uses in the example you posted.
Many thanks!

On Friday, October 17, 2014 12:39:09 PM UTC+1, Fabien wrote:
> I forgot to answer to your first question about concatenation. Once you
>
> have non-regular subsets you have to work with masks and where() to
>
> access your data.
>
>
>
> (e.g.: https://www.idlcoyote.com/ip_tips/xroi.html)
>
>
>
> On 17.10.2014 13:25,
>
>> Eg, is it possible to concatenate the following?

Report message to a moderator

Re: non-rectangular array subset [message #89450 is a reply to message #89449]

Fri, 17 October 2014 06:42

Fabzi
Messages: 305
Registered: July 2010

Senior Member

Hi,

A 2d grid of data is the same as an "image", and third or fourth
dimensions (e.g. time or atmospheric levels) are simply a "pile of images".

Depending on the size of your array you could do some dimension juggling
(https://www.idlcoyote.com/tips/rebin_magic.html, case study 1) or use a
loop over the other dimensions.

Fabien

Report message to a moderator

Re: non-rectangular array subset [message #89452 is a reply to message #89450]

Fri, 17 October 2014 14:49

Chris Anderson
Messages: 3
Registered: August 2014

Junior Member

Jess,

Subsetting your data using the 'where' function will allow you to subset your data from your image arrays into one-dimensional vectors that can then be concatenated together into a single array. I'm sure this can be done in a cleaner fashion, but one potential example, given two images and some sort of threshold could be:

index1=where(image1 gt threshold,cnt)
if (cnt gt 0) then output_data=image1[index1]

index2=where(image2 gt threshold,cnt)
if (cnt gt 0) then output_data=[output_data,image2[index2]]

Your output_data vector then will contain a one-dimensional array of the values you have subset from your original images. All of your LAI values for the regions you are using will be in one array.

On Friday, October 17, 2014 6:42:07 AM UTC-7, Fabien wrote:
> Hi,
>
>
>
> A 2d grid of data is the same as an "image", and third or fourth
>
> dimensions (e.g. time or atmospheric levels) are simply a "pile of images".
>
>
>
> Depending on the size of your array you could do some dimension juggling
>
> (https://www.idlcoyote.com/tips/rebin_magic.html, case study 1) or use a
>
> loop over the other dimensions.
>
>
>
> Fabien

Report message to a moderator

Previous Topic:	Converting days since 0-01-01 to Julian/Gregorian days - how can IDL recognise 0 AD?
Next Topic:	File time issue

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 07:23:02 PDT 2025

Total time taken to generate the page: 0.43366 seconds