Glenn,

I meant simple orthogonality relation between column vectors of localToWorld 
transform. First column represents local east vector, second local north, and 
third local up. computeLocalUp returns the same vector as third column of 
localToWorld transform. If computeLocalUp method gets changed, shouldn't 
computeLocalToWorldFromXYZ, computeLocalToWorldFromLatLonngHeight, be changed 
as well ? And then if we want to mantain column orthogonality it would require 
changes in computation of local north vector(second column). 

Cheers,
Wojtek Lewandowski

  ----- Original Message ----- 
  From: Glenn Waldron 
  To: OpenSceneGraph Submissions 
  Sent: Wednesday, July 01, 2009 12:24 AM
  Subject: Re: [osg-submissions] EllipsoidModel::computeLocalUpVector


  IMHO: because getLocalUpVector() is a member of EllipsoidModel, it should 
respect the ellipsoid settings and return a vector orthogonal to the surface.

  Wojtek, I am not sure what you mean about killing the relationship b/w the up 
vector and the east/north vectors. Can you clarify this?


  Glenn Waldron : Pelican Mapping : http://pelicanmapping.com : +1.703.652.4791



  On Tue, Jun 30, 2009 at 4:55 PM, Wojciech Lewandowski <[email protected]> 
wrote:

    Hi Ronald and all interested,

    I think you have touched a very delicate subject... But since You started 
it  I will add my 2 pennies  

    I agree that local coordinate system produced by EllipsoidModel has a bit 
unusual orientation. I did some investigation one day and it looks like this 
local coordinate frame is not tangent to ellipsoid surface. Instead local 
coordinate orientations correspond to tangent surface on a perfect sphere. But 
at least these orientation vectors in such spherical coordinate system are 
consistent. Local up vector is orthogonal to east and north vectors. If you 
change upVector method you will kill this relationship.

    I would be happy if we could use local space that would be tangent to 
ellipsoid but this would require not only changing localUp method but also 
changing worldToLocal and localToWorld methods. However, this sounds to me like 
problems with former osgDEM and VPB generated databases.... 

    Frankly, I don't know what is best solution here. Maybe someone with GIS 
experience could provide some guidance. I think that we should ask for the 
answers to few simple questions : how the "up" vector should be defined ? As 
vector opposite to gravity vector ? Or as vector orthogonal to local east and 
north vectors ?  Should the local north and local east vectors be tangent to 
the ellipsoid or not ? 

    Cheers,
    Wojtek Lewandowski


    From: Maarseveen, R.A. (Ronald) van 
    Sent: Tuesday, June 30, 2009 8:07 PM
    To: [email protected] 
    Subject: [osg-submissions] EllipsoidModel::computeLocalUpVector


    Hi all,

    In OpenSceneGraph-2.8.2-rc2 in the file include/osg/CoordinateSystemNode

    Computing an up vector on an ellipsoid should take the geographic latitude 
into account.
    The geographic latitude is the angle between the normal to ellipsoid 
surface and XY-plane.

    The original code computes an upvector for a sphere.

    Best regards,
        Ronald van Maarsveen
This e-mail and its contents are subject to the DISCLAIMER at 
http://www.tno.nl/disclaimer/email.html


----------------------------------------------------------------------------


    _______________________________________________
    osg-submissions mailing list
    [email protected]
    
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org


    _______________________________________________
    osg-submissions mailing list
    [email protected]
    
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org






------------------------------------------------------------------------------


  _______________________________________________
  osg-submissions mailing list
  [email protected]
  
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to