[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20011030: GINI conversion to AREA (cont.)

>From: "Jason J. Levit" <address@hidden>
>Organization: Center for Analysis and Prediction of Storms, University of Okla
homa
>Keywords: 200110222135.f9MLZn102781 McIDAS-X ADDE GINI AREA

Jason,

>  I'm not sure when you'll get this e-mail, since our campus e-mail
>severs have been very flaky yesterday and today.  No incoming or
>outgoing mail for about 48 hours.

OK.  If this bounces back to me, then I will know it failed.  If it doesn't,
I will assume that it failed (and retry) if I don't hear from you within
a day.

>So, here goes:

re: use the SSIZE= keyword
>  This worked perfectly!!!!  Good find in the manual.  I didn't see that
>particular option available.

I never had to use this option before, so I really didn't understand it
until now.

>I tried this on some of our data and the
>image remap worked like a charm.

Excellent.

>  About the only item that worries me with this remap is a topic you
>mentioned a few e-mails ago, concerning choosing a good image to re-map
>into.  What I'm concerned with is the that the numbers within the GVAR
>navigation block change from scan to scan, so, how do I pick an image
>with good map registration?

I would pick an image; load it centered on several LAT,LON locations
where identifiable landmarks can be seen; draw a VeryHigh (VH)
map on top of it (MAP VH); and see how well the drawn map matches the
underlying landmark.

>I also suppose the other item I need to
>check out is how our particular re-mapping code uses the GVAR navigation
>block.  If it's heavily dependent some of those values in the block,
>then I'm concerned that we might be introducing too many errors when
>remapping from the GINI data "into" a GVAR navigation block.  Do you
>have any more thoughts on this?

I think that the "proof" is in the registration of identifiable landmarks
in the remapped image.  You should be aware that the navigation coming
down with the GINI imagery is not perfect.  You can demonstrate this
to yourself by:

o creating a dataset that has several/lots of a particular kind of GINI
  image in it (or point to adde.ucar.edu and use the GINIEAST dataset)
o put up a loop of images loaded over a particular LAT,LON point
o animate the images in the loop and see if the images "bounce"

Here is what I have in mind:

1) start a McIDAS-X session with at least 15 frames.  Depending on the
   resolution of the X server you are running, I would make the frames
   at least 600x800 (LINes x ELEments).

2) load a loop of GINI images in your session specifying a particular
   center point (LAT,LON or STAion ID) (assumes that you have at least
   15 frames):

   IMGDISP GINIEAST/GE1KVIS LATLON=25 80 MAG=1 EU=IMAGE ALL=1 15 REFRESH='EG 
(GRA);MAP VH GRA=(GRA)

   - or -

   IMGDISP GINIEAST/GE1KVIS STATION=KMIA MAG=1 EU=IMAGE ALL=1 15 REFRESH='EG 
(GRA);MAP VH GRA=(GRA)

   LB 1 15
   DR 14*1 20
   TERM L ON

>Thanks again for all of your help. 

No problem.

>We're almost there, I think!

Yes, I think so.

Tom