| Re: Drawing satellite pixels on maps? [message #7452] |
Tue, 19 November 1996 00:00 |
Robert.M.Candey
Messages: 23
Registered: June 1995
|
Junior Member |
|
|
In article <3291C91F.7F816972@oma.be>, Philippe Peeters
<Philippe.Peeters@oma.be> wrote:
> I've already posted this question some time ago but did not get any
> answer. Here I go again:
>
> I want to draw satellite data on a map. Each ground pixel is defined by
> the latitude and longitude of the 4 corners. It is not a regular
> rectangle or square and depend on the viewing geometry of the satellite
> instrument.
> I have tried a simple polyfill,long,lat but I have serious problems with
> orthographic maps when the pixel is partly off the map. I got strange
> filled polygons from the edge of the map to the corner of the window.
> Someone on the net advice me to use a new polyfill routine which checks
> polygons boundaries before drawing it. Though slower than the original
> polyfill, it solved the problem.
> But I still have another problem with several maps when the pixel to be
> drawn is at the edge, i.e. when one or several pixel corner is on one
> side of the map (lon > -180) and the other on the other side (lon< -180)
> of the map.
> example longitude=[-179,-181,-179.5,-180.5] or [179,181,179.5,180.5]
>
> The pixel is drawn from one side to the other of the map which is pretty
> ugly. Obviously this is a 'normal' way of drawing that kind of pixel,
> polyfill is not supposed to know that it has to cut the pixel into two.
> Does anybody know how to solve this problem.
>
> And now a question related to the same topic. How can I resample the
> irregular ground pixels onto a regular (square or rectangle) grid?
>
> Philippe Peeters
I have struggled with similar problems and recently posted my attempts at a
solution (auroral_image.pro; email me for a copy). To get around the
problem of polyfill across the whole map, I checked to see if the endpoints
of the polygon to fill are nearby in normalized coordinates and then I
skipped plotting it if it was not nearby; you could split the polygon in
two and plot each part, I suupose. The code is like this (for a list of
triangles, with Lat1 and Lon1 defined at each corner):
pAll = convert_coord(Lon1, Lat1, /data, /to_normal)
pLon = pAll(0,*) & pLat = pAll(1,*)
for i=0L,n_elements(triangles(0,*))-1 do begin
tri1 = triangles(*,i)
Lon3 = Lon1(tri1) & Lat3 = Lat1(tri1)
;### which average scheme is best?
; Zb1 = avg(Zb(tri1)) ; average byte values
; Zb1 = Zb(tri1(0)) ; faster than avg and as accurate?
Zb1 = doByteScale([avg(Z2(tri1))],minZ1,maxZ1,Zsize,wBad,flipColor Bar,1)
; doByteScale is my routine for scaling the data values between minZ1 and maxZ1
; with some other constraints
if (total(abs(pLon(tri1)-shift(pLon(tri1),1))) lt 0.1) and $
(total(abs(pLat(tri1)-shift(pLat(tri1),1))) lt 0.1) then $
polyfill, Lon3, Lat3, color=Zb1(0), noclip=0
; you could do a "where" here instaed of "if" to find which corners are out of
; bounds
endfor
As for resampling, you could use triangulate and trigrid with the sphere
option (see the Map_image method in my auroral_image.pro for an example);
but you are probably better off (more scientifically accurate) plotting the
original polygons and not resampling
--
Robert.M.Candey@gsfc.nasa.gov
NASA Goddard Space Flight Center, Code 632
Greenbelt, MD 20771 USA 1-301-286-6707
|
|
|
|
| Re: Drawing satellite pixels on maps? [message #7455 is a reply to message #7452] |
Tue, 19 November 1996 00:00  |
davidf
Messages: 2866
Registered: September 1996
|
Senior Member |
|
|
Philippe Peeters <Philippe.Peeters@oma.be> writes:
> [Question about rendering split pixels that I can't answer without research.]
>
> And now a question related to the same topic. How can I resample the
> irregular ground pixels onto a regular (square or rectangle) grid?
The easiest way to resample irregular ground pixels to a regular
grid is to use the TRIANGULATE and TRIGRID routines that are
built into IDL. The method is straightforward: the TRIANGULATE
procedure takes your sample locations and returns a set of
Delaunay triangles constructed from the locations. Passing the
locations, the data itself, and the triangles to the TRIGRID function
results in a gridded data set. You can specify the grid intervals.
Depending upon the geographical extent of your data locations,
you may prefer to perform spherical gridding (i.e. use spherical
triangles rather than flat, 2D triangles). Be aware that the
TRIANGULATE and TRIGRID routines are capable of this as
well, although different keywords need to be used (e.g, see
the SPHERE keyword to TRIANGULATE).
David
*************************************************
* David Fanning, Ph.D.
* 2642 Bradbury Court, Fort Collins, CO 80521
* Phone: 970-221-0438 Fax: 970-221-4762
* E-Mail: davidf@dfanning.com
*
* Sometimes I go about pitying myself, and all along my
* soul is being blown by great winds across the sky.
* -- Ojibway saying
************************************************
|
|
|
|
| Re: Drawing satellite pixels on maps? [message #7464 is a reply to message #7452] |
Tue, 19 November 1996 00:00 |
Andy Loughe
Messages: 174
Registered: November 1995
|
Senior Member |
|
|
Philippe Peeters wrote:
>
> I've already posted this question some time ago but did not get any
> answer. Here I go again:
>
> I want to draw satellite data on a map. Each ground pixel is defined by
> the latitude and longitude of the 4 corners. It is not a regular
> rectangle or square and depend on the viewing geometry of the satellite
> instrument.
> I have tried a simple polyfill,long,lat but I have serious problems with
> orthographic maps when the pixel is partly off the map. I got strange
> filled polygons from the edge of the map to the corner of the window.
> Someone on the net advice me to use a new polyfill routine which checks
> polygons boundaries before drawing it. Though slower than the original
> polyfill, it solved the problem.
This is NOT a perfect solution, but it may give you some
ideas on how to proceed...
;
; Name ............ my_polyfill.pro
;
; Description ..... Before executing polyfill, check that
; the (x,y) coordinates are valid.
;
; Parameters ...... x,y contain vertices of the polygon.
;
; Keywords ........ See documentation for polyfill.
;
; Originator ...... Andrew F. Loughe, NOAA/ERL/CDC, 21 MAY 1996.
pro my_polyfill, x, y, _EXTRA=e
on_error, 2
; Convert the x,y coordinates to the normalized system.
result = CONVERT_COORD(x, y, /data, /to_normal) ; DEFAULT:
From /data
; Check that /normal or /device might be set.
if ( KEYWORD_SET(e) ) then begin
names = TAG_NAMES(e)
for i = 0, N_TAGS(e)-1 do begin
if ( names(i) eq 'NORMAL' ) then $
result = CONVERT_COORD(x, y, /normal, /to_normal) ; From
/normalized
if ( names(i) eq 'DEVICE' ) then $
result = CONVERT_COORD(x, y, /device, /to_normal) ; From
/device
endfor ; i-loop
endif ; KEYWORD_SET(e)
; Check that the normalized coordinates are within the bounds
; set by !x.window and !y.window
out_of_bounds = 0
if ( MIN(result(0,*)) lt !x.window(0) or $
MAX(result(0,*)) gt !x.window(1) or $
MIN(result(1,*)) lt !y.window(0) or $
MAX(result(1,*)) gt !y.window(1) ) then out_of_bounds = 1
if (out_of_bounds eq 1) then print, '(x,y) coordinates are out of
bounds!'
if (out_of_bounds eq 0) then polyfill, x, y, _EXTRA=e
end
> But I still have another problem with several maps when the pixel to be
> drawn is at the edge, i.e. when one or several pixel corner is on one
> side of the map (lon > -180) and the other on the other side (lon< -180)
> of the map.
> example longitude=[-179,-181,-179.5,-180.5] or [179,181,179.5,180.5]
What if you add 360 to your points vector and
make your map go from 180 to 540?
map_set, 0, 360, /cont
Does that work? |
--
Andrew F. Loughe afl@cdc.noaa.gov
University of Colorado, CIRES http://cdc.noaa.gov/~afl
Campus Box 449 phn:(303)492-0707 fax:(303)497-7013
Boulder, CO 80309-0449 "He who laughs last thinks slowest!"
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
