| Working with 2 partially overlapping images of different array sizes [message #28870] |
Fri, 11 January 2002 12:33 |
aqueous0123
Messages: 11
Registered: September 2001
|
Junior Member |
|
|
Sorry, I'm new to this image processing stuff.
I have 2 images of different sizes and different, but overlapping,
locations on the globe. E.g., 2 satellite images from different
sensors.
|||||||||||||||| <- IMAGE_1 = bytarr(532,532)
| |
| ''''''''''''''''''
| ' | '
| ' | '
| ' | '
| ' | '
| ' | '
||'||||||||||||| ' <- IMAGE_2 = bytarr(250,200)
''''''''''''''''''
Each, IMAGE_1 and IMAGE_2 has its own arrays of associated lats and
lons:
LATS_1 = fltarr(532)
LONS_1 = fltarr(532)
LATS_2 = fltarr(200)
LONS_2 = fltarr(250)
Notice that IMAGE_2 is
(1) not the saxe # pixels in X and/or Y and
(2) that its land coverage is not exactly the same as IMAGE_1, its
range covers an area slightly to the south and east.
In summary, IMAGE_2 has smaller size in array space but covers nearly
the same geographical space, just shifted a bit.
What I wish to do is create a new array IMAGE_3 that is the same array
size (532,532) as IMAGE_1 AND also has the same physical location as
IMAGE_1, but with IMAGE_2's data where IMAGE_1 and IMAGE_2 overlap
(the 2's in the graphic below). Where they do not overlap, I'll just
fill IMAGE_3 with nearest neighbor of IMAGE_2 (the 0's in the grahic
below).
0000000000000000 <- IMAGE_3 = bytarr(532,532)
0000000000000000
0022222222222222
0022222222222222
0022222222222222
0022222222222222
0022222222222222
0022222222222222
0022222222222222
So, it seems like I have to:
(1) find intersections of IMAGE_1 and IMAGE_2 using each's LAT/LON
arrays
(2) extract this data from IMAGE_2, call this IMAGE_2a
(2) congrid() this IMAGE_2a to the dimensions of IMAGE_1's subarea
that IMAGE_2a 'covers', creating IMAGE_2b
(3) Create final image and dummy fill for now: IMAGE_3 = IMAGE_1 &
IMAGE_3[*,*] = 255b
(4) Fill IMAGE_3 with IMAGE_2b, but only where IMAGE_2b overlaps (the
2's in the graphic above)
(5) Fill rest of IMAGE_3 with nearest neighbor from subarea IMAGE_2b
filled in STEP 4 (the 0's in the graphic above). Now final IMAGE_3 has
same LAT/LON bounds as IMAGE_1, the same array dims as IMAGE_1, but
filled with the data from IMAGE_2.
In thinking the problem, I believe my inexperience is making me
over-engineer this solution. I'm not even sure how to do this all.
Help is greatly appreciated. Thanks.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
