Re: Gridded2DSet

Hi Desiree,

Desiree Hilbring wrote:
> Hi Ugo,
> thanks for your answers I will think about them and try, but first can
> yout clarify the difference between Gridded2DSet and Linear2DSet?
> I am not really sure which is better to use:

Both sets form a valid grid. A Linear2DSet forms a GRID like one knows
from ArcInfo or ArcView, that is, a such a set describes the domain set
of, say, a jpg image. A gridded set has the same topology, but the
distance between adjacent points is not constant.

To put it in other words, suppose you go out to make a barbecue. While
your steak, bread or whatever you fancy, grills over the (linear
sampled) grid, you think about VisAD Data types...and sees that you have
a Linear2DSet in front of you. As the food is not ready yet, and you
spent energy thinking about data objects, you get even hungrier you put
more coal on the fire to speed up the grilling. Now, the temperature
went out of control, your grill (graticule, lattice) has somewhat melted
and warped, but you realise that the intersections are still visible
(i.e. they have not melted into a single piece). In fact, you still
recognise your grid as such, but the distance between the intersections
is not constant anymore. You then realise, that if you were to model
this new grid, you couldn't use a Linear2DSet anymore, only a
Gridded2DSet (because the distance between the metal rods are not
parallel anymore). You realise further, that your initial grid could
ALSO be described by the Gridded2DSet. Then you go look up in the VisAD
Javadoc and see the Linear2DSet extends the Gridded2DSet, that is, the
linear set is an specialisation of the gridded set, in that the samples
are constantly spaced relative to each other.
When you come back to your barbecue, the grid has definitely melted, but
you still see SOME intersections. So I guess you could only model this
new melted grid with an Irregular2DSet.

> I do have a DEM with the constant distance of 1m between the points, but
> it is not rectangular, it looks like this:
>     . . .
>   . . .
>     . . . .
>     . . .
>     . .
> So my first assumption was to use a Linea2DSet and fill the not used
> points, which worked with Bills help.
> Bill told me I could use Gridded2DSet as well, so I am trying this now.
> I am not sure which one is better?
> And the real problem, i did not mention before is the following:

Well, this looks like a Linear2DSet. The missing points are made
invisible by setting the range samples (of the FlatField) to NaN.
A Linear set will do the hard work for you, so you don't have to set
every (x,y) value yourself. However, in view of the grid-like topology
of your DEM it is perfectly acceptable to use a gridded set.

> I can load my dem into the ArcView 3D-Analyst which results in a rough DGM
> and gets finer while zooming in.
> The same dataset loaded into my application GISterm looks like the
> resolution not that good as in ArcView,
> you can see the screenshots in my homepage:
> What can cause this effect?

I can't tell you. My first question is, how did you read the values?
That is, are you using the same resolution in your viewer as you have in
ArcView? (It looks like you aren't.)

My second guess is, is there a call like

GraphicsModeControl dispGMC = (GraphicsModeControl) 

> Thanks Desiree

You're welcome.


PS.: sorry about the barbecue metaphor. The summer has settled in this
part of the planet and I'd rather be outside, taking care that barbecue
grills do not become irregularly sampled ;-)

  • 2001 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the visad archives: