| upsampling images [message #84396] |
Mon, 03 June 2013 17:56  |
Oana Coman
Messages: 27
Registered: December 2011
|
Junior Member |
|
|
Hi all,
I've been trying to tinker with figuring out how to upsample images since I need very specific output dimensions and cannot get that accomplished using rebin factors in RESIZE_DOIT.
I basically need to upsample images to match my highest resolution image so I can do a bunch of band math on them.
I thought I had found my saving grace in ENVI_LAYER_STACKING_DOIT, but for some images it gives me results that are just 1 pixel off, so I can't always use band math.
Is there another trick that I'm not aware of to try to resize images to higher resolutions? Something where I can enter the output dimensions and have IDL resize my image to those specific dimensions?
Thanks!
|
|
|
|
| Re: upsampling images [message #84571 is a reply to message #84396] |
Tue, 04 June 2013 15:00  |
Klemen
Messages: 80
Registered: July 2009
|
Member |
|
|
If you know, the number of cols /lins and you know how the pixels are positioned, you can use just interpolate function. But if you know, that each upscaled pixel should consist of e.g. 5 by 5 original pixels, then this example (upsample the MODIS from original swath to 5 by 5 averaged pixels). The whole code is at another thread:
https://groups.google.com/forum/?hl=en&fromgroups#!topic /comp.lang.idl-pvwave/aI6Vh_op7Lc
Cheers, Klemen
; Average original data to "geolocation frame"
;first prepare indexes
out_size = size(m_lat) ;the output will have a reduced spatial
resolution (corresponding to the geolocation)
;the position 0,0 in geolocation corresponds to pixel 2,2 in original
data
;the geolocation is 5 times downsampled
out_indx_col = indgen(out_size[1]) * 5L + 2L ;corresponding coloumns
of orig. data in downsampled grid
out_indx_lin = indgen(out_size[2]) * 5L + 2L ;corresponding lines of
orig. data in downsampled grid
out_indx_col = rebin(out_indx_col, out_size[1], out_size[2])
out_indx_lin = rebin(reform(out_indx_lin,1,out_size[2]), out_size[1],
out_size[2])
out_indx = out_indx_lin * in_size[1] + out_indx_col ;one dimensional
index of original data in downsampled grid
- hide quoted text -
;compute mean value for the radiance
m_count = make_array(out_size[1],out_size[2]) ;array containing the
number of good maeasurements
m_mean = make_array(out_size[1],out_size[2]) ;array containing the
mean maeasurements
for j=-2,2 do begin
for i=-2,2 do begin
indx = out_indx + out_size[1]*j + i
tmp = m_modis[out_indx]
indx_good = where(tmp le 32767) ;do not use nodata, etc.
m_count[indx_good] = m_count[indx_good] + 1
m_mean[indx_good] = m_mean[indx_good] + tmp[indx_good]
endfor
endfor
m_mean = m_mean / m_count
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
