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

[McIDAS #MMD-610472]: 20140718: Problem matching projections between McIDAS and Terascan



Hi Greg,

re:
> The problem with changing the reference latitude is that those images
> don't map properly in OpenLayers, Google Maps, etc. APIs that require
> Mercator projected images.

OK.  I don't know what OpenLayers requires, but I do know that Google
Maps requires a close variant of the Mercator projection:

http://en.wikipedia.org/wiki/Google_Maps

re:
> The reference latitude must be the equator
> according to these vendors. We now use these tools extensively in our field
> project support. The Terascan image uses the equator for the reference
> latitude - that's why I was doing the same for the McIDAS image.

OK.

re:
> I think the McIDAS image is a Mercator projection, but I don't believe
> the resolution is the same. In the Terascan image, each pixel is exactly
> 1.0 km high and 1.0 km wide.

The location of cloud and features looks to be more-or-less consistent same
between the remapped Terascan and remapped McIDAS images.  I say more-or-less
because the to pictures you sent me appear to be for different times:

19:09:04.266 - Terascan image
19:30        - McIDAS image

The biggest difference is that the McIDAS image has less areal coverage
than the Terascan image ... as if it has been magnified/stretched a bit.

re:
> While I set the resolution in McIDAS to 1, I
> don't believe the pixels in that image are 1x1.

In the example remap that I ran, the remapped McIDAS pixels are
about 0.77 km in resolution in both the horizontal and vertical.
This can seen from an IMGLIST FORM=ALL listing and from measuring
the distance between widely spaced points (using DIST) and dividing
by the number of pixels between those points.

Aside:

- the ratio of 1 km to 0.77 is 1.3 which appears to account
  for the apparent magnification of the McIDAS picture over the
  Terascan picture

re:
> They seem substantially larger given the different plot areas.

They are actually smaller (0.77 km vs 1 km).  The big difference is the
areal coverage.

The question then boils down to whether or not the remapping in McIDAS is
correct or not. Here is what I did to convince myself that it is:

- display the original EAST/CONUS image in one frame and the IMGREMAPped
  image in a second frame centering both images displays on LATLON=39.25 105

- interrogate McIDAS for the location of identifiable cloud features across
  the full areal coverage of the image(s)

  I get consistent Lat,Lon locations for each feature I examined which I
  interpret as the remapping was done correctly.

So, I think we can say that the RES= keyword in the IMGREMAP command
only approximately represents the actual resolution of a pixel in the
remapped image.  I do _not_ know why this is!

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: MMD-610472
Department: Support McIDAS
Priority: Normal
Status: Open