Re: Thoughts on GALEON Phase 2

NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.

It' interesting to see the example of WCS return in GML.  In the WCS RW we 
had been discussing how to use GML RectifiedGrid in WCS getCoverage request 
and/or getCoverage response.  I had difficult time to understand the 
example in page 221-222, section 19.2.2 Rectified Grid (gml:RectifiedGrid), 
ISO/CD19136, N005r3, dated on 2004-02-07.  I attach this example in this 
mail and hope that someone in this group will be able to help.  What I 
don't understand are the three values in the gml:position element and in 
the gml:offsetVector elements.  Shouldn't there be only two values in each 
of these elements because the grid dimension equals to "2"?  I remember 
that I had asked at least one of the GML project team experts about this 
question but I didn't get response.



Another question concerns using of the gml:RectifiedGrid is the grid 
coordinate values defined by gml:low and gml:high.  The WCS return of the 
example getCoverage request posted by Rudy bellow:


<http://webapps.datafed.net/dvoy_services/ogc.wsfl?SERVICE=wcs&REQUEST=GetCoverage&VERSION=1.0.0&CRS=EPSG:4326&COVERAGE=THREDDS_GFS.T&FORMAT=GML&BBOX=-180,-90,180,90,1350,1350&TIME05-12-06T12:00:00Z&WIDTH=900&HEIGHT=400&DEPTH=0>


is a grid, in GML, with grid origin at (-90,-180), i.e., 
<gml:pos>1133870400000 1350 -90 -180</gml:pos>  (note that I only focus on 
the lat/lon dimensions for simplicity) and grid coordinate values being 
<gml:low>0 0 0 0</gml:low> and <gml:high>0 0 72 72</gml:high>.  It seems 
that the returned grid has 72 rows and 72 columns, which is not consistent 
with the "width=900 & height=400" in the request.  The "width=900 & 
height=400" means that the returned grid should have 900 columns and 400 
rows.  I don't know how the values in gml:low and gml:high were determined.



I try to ask my question regarding the gml:low/gml:high by using examples 
and hope that someone can help me out. I'll just list partial elements 
rather than a complete gml:RectifiedGrid for simplicity.  I also omit the 
crs name and assume it's in EPSG4326:



Assume there is a WCS returned grid, named GRID-A, described by the 
following gml

Axes: <gml:axisName>Lat</gml:axisName>  and  <gml:axisName>Lon</gml:axisName>
Envelope: <gml:low>0 0</gml:low> and <gml:high>10 10</gml:high>
Origin:  <gml:pos> -90 -180</gml:pos>

Offset: <gml:offsetVector>1.0 0.0</gml:offsetVector>  and 
<gml:offsetVector>0.0 2.0</gml:offsetVector>
This gml shows a WCS returned grid with origin at (-90, -180), having 10 
rows and 10 column, with cell spaces of 1.0 degrees in latitude direction 
and 2.0 degrees in longitude direction.  Thus, the grid covers from 
(-90,-180) to (-80,-160).



Now, consider another WCS returned grid, named GRID-B, with everything the 
same as the above GRID-A, except that the gml:low/gml:high values are 
different, which are <gml:low>10 10</gml:low> and <gml:high>20 20</gml:high>:

Axes: <gml:axisName>Lat</gml:axisName>  and  <gml:axisName>Lon</gml:axisName>
Envelope: <gml:low>10 10</gml:low> and <gml:high>20 20</gml:high>
Origin:  <gml:pos> -90 -180</gml:pos>

Offset: <gml:offsetVector>1.0 0.0</gml:offsetVector>  and 
<gml:offsetVector>0.0 2.0</gml:offsetVector>



Question: what are the spatial coverage and origin of this GRID-B?  I am 
not sure which one of the the followings is correct:



1. GRID-B covers from (-80,-160) to (-70,-140) because it starts from 
<gml:low>10 10</gml:low>.
If this is true, then the 10-row by 10-column grid's origin is NOT at 
(-90,-180).  The origin should be at (-80,-160).



2. GRID-B covers from (-80,-160) to (-70,-140) because it starts from 
<gml:low>10 10</gml:low>.  It's origin is at (-90,-180) but the origin is 
outside the 10x10 returned grid.
If this is true, what is the usefulness of this origin information?  What 
is it different from saying: the grid's origin is at (+10,+10), or anywhere 
else?



3. GRID-B covers from (-90,-180) to (-80,-160), as does the GRID-A, and 
thus its origin is correctly stated as (-90,-180).  However, it has grid 
coordinate extends from (10,10) to (20,20).
If this is true, what does this "grid coordinate extents" means?  Is there 
any difference between referring a 10x10 array to being (10,10) to (20,20) 
and referring it to being (0,0) to (10,10)?


4. Any other difference between GRID-A and GRID-B?


When considering these questions, it should keep in mind that we are 
talking about using gml describing a grid already produced by a WCS server 
and returned to a WCS client.  We are not talking about subsetting from a 
source grid.  It's obvious that gml:low/gml:high can be used to specify a 
subset of a source coverage, i.e., from gml:low to gml:high.



Also notice that for in many data formats, there is no such thing as to 
saying a 10x10 array beginning from (10,10) and ending at (20,20), or any 
other values such as (r0,c0) to (r0+10,c0+10).  The offsets, r0 and c0 in 
terms of row and column numbers, are simply useless.  For example, in 
HDF-EOS, it specifies the geoinformation for the first and last row/col of 
an array.  There is no place to say that this array is starting from 
(r0,c0) to (r0+10,c0+10) because the (r0,c0) information is of no use.  I 
believe that it is similar in GeoTiff, where image array's row/col numbers 
are defined by imageWidth/ImageLength tags.  The tie points are defined in 
reference to the first row/col (e.g., geo-x/y values for first r/c, 3rd 
r/c, 10th r/c, etc), rather than in reference to absolute r/c numbers such 
as geo-x/y values for row 10001 and col 2001 of a 10r by 10c image with r/c 
ranging from (1000,2000) to (1010,2010).


Regards.

Wenli Yang
George Mason University

At 01:37 2006-2-27 -0600, Rudolf Husar wrote:

Gerry,


The use of multiple 'languages' e.g. GML and SensorML seems to me the only 
way to describe the different aspects of Earth Science monitoring and data 
systems. This may be naive, but my question is not whether but  how to use 
the combination of these standard languages to describe the entire 
elephant. Some early, thinking-fragments on this topic, related to air 
quality is here.

<http://capita.wustl.edu/capita/researchareas/GALEON/Reports/AQ_Pt_OGC.ppt>http://capita.wustl.edu/capita/researchareas/GALEON/Reports/AQ_Pt_OGC.ppt

Ron,


Expanding on Ben's comment, GML (rectified grid) is one of the WCS data 
formats returned from (some of the) test servers in GALEON, e.g.
<http://webapps.datafed.net/dvoy_services/ogc.wsfl?SERVICE=wcs&REQUEST=GetCoverage&VERSION=1.0.0&CRS=EPSG:4326&COVERAGE=THREDDS_GFS.T&FORMAT=GML&BBOX=-180,-90,180,90,1350,1350&TIME=2005-12-06T12:00:00Z&WIDTH=900&HEIGHT=400&DEPTH=0>http://webapps.datafed.net/dvoy_services/ogc.wsfl?SERVICE=wcs&REQUEST=GetCoverage&VERSION=1.0.0&CRS=EPSG:4326&COVERAGE=THREDDS_GFS.T&FORMAT=GML&BBOX=-180,-90,180,90,1350,1350&TIME=2005-12-06T12:00:00Z&WIDTH=900&HEIGHT=400&DEPTH=0 



All,

FYI, George Percivall, Chief Architect of OGC, has prepared this PPT on 
the OGC Networks, including the GEOSS Services Network, (GSN)

<http://portal.opengeospatial.org/files/index.php?artifact_id=13947>http://portal.opengeospatial.org/files/index.php?artifact_id=13947

Rudy
Professor of Mechanical Engineering, Director
CAPITA, Washington University
St. Louis, MO 63130
1 Brookings Drive, Box 1124
314 935 6099



On 2/26/06, Gerry Creager N5JXS 
<<mailto:gerry.creager@xxxxxxxx>gerry.creager@xxxxxxxx> wrote:

I'd like to offer the thought that we need to consider SensorML, as well
as GML, or potentially ESML, to allow description of datasets and
sensors involved in this work.

Gerry

Attachment:
rectifiedGrid.jpg

Description: JPEG image

 



   
  
   







   
 
   

    
    
     
  • 
      2006 messages navigation, sorted by:
    
      
      
       
    1. Thread
      2. 
       
    2. Subject
      3. 
       
    3. Author
      4. 
       
    4. Date
      5. 
       
    5.  ↑ Table Of Contents 
      6. 
      
     
     
    • 
     
  • 
      Search the galeon archives:
    
      
    
         
       
      
    
      
     
    •