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

Re: Newbie needs help...

karmic confusion reigns supreme... 

David Fanning wrote:
> Bruce Bowler (bbowler@bigelow.org) writes:
> > I'm willing to try this "thinking outside the box" thing for a while,
> > but I can't see the box.
> There is no box. (But see more on spiritual development, below.)
> > Now that I have my image mapped on to a lat/lon grid (see, I'm learning
> > already :-), how to I access the data by lat/lon?
> You are going to love this! :-)
> You have set up the map coordinate space with Map_Set.
> You have placed your image on the map with Map_Image
> (or something similar). You put your map grid and
> continental boundary on your map with Map_Grid
> and Map_Continents.

now we maybe be getting somewhere...  I have this big huge array (here
after referred to as a BHA) of data (BHAd) an equally sized BHA of
latitudes (BHAlat) of each point in BHAd and another BHA of longitudes
(BHAlon) of each point in BHAd.  NB, row x of BHAlat is not constant,
neither is column y of BHAlon

I understand map_set, I understand map_grid, I understand map_continents
(well, not in the tao-ist sense of the word).  map_image is where I'm
getting confused...  how does map_image know about BHAlat and BHAlon
when they are not inputs to the process? 

To add a little to the confusion, and to give some scope to the problem,
BHAd is an array [1354,2030] and covers lat/lon
[34.9,-78.33,56.58,-41.68].  I'm only interested in lat/lon
[41,-71,45,-66] (it's a MODIS swath but the data's an experimental
product so none of my "normal" MODIS tools work), we're doing sea-truth
work for this product in the Gulf of Maine.

> Now, you get lat/lon value from the user by just
> having them click on the map! Too easy!
>    Cursor, lon, lat, /Data

no, they're going to tell me the lat and lon from a data file, but that
part I can handle.
> Now, what you do next depends on you. If you
> have an image data set in which each pixel
> has an associated lat/lon coordinate, you can
> go pull out the closest pixel value from that
> data set.

I suspect once I figure out the map_image part, this part MIGHT fall out
on it's own...

> If you don't have such a data set, you might
> have to get the value from the warped image.
> That value, of course, was created by smushing
> (technical term) several real values together
> in the warping process. It may not be what you
> want. (How this is done is fairly complicated.
> I'd explain it, but I'm pretty sure it's not
> what you want to do anyway.)

Ah shucks, go ahead, explain it.  It can't make my head hurt anymore
than it already does...