[x3d-public] GeoPlanets, was: Invalid geoSystem values in X3D Resources BasicGeospatial examples.

[GPU Group]

Andreas,
Thanks for those links to those detailed Geospatial proposal. I hadn't see
that. I'll read it in more detail.

3TM is commonly used by cities around the world in their GIS department,
because the scale factor is close to 1 so when you take measurements it's
close to the surveyor's tape measure..

And it uses the same transverse mercator math functions as UTM, so just a
scale factor and 2 offsets different - a very easy thing for browser
developers to add to their UTM support capture the city GIS crowd.

> the underlying assumption is that in the scene regular (nongeo)
> coordinates are geocentric, eg. relative to the center of a single planet.

Not quite or that's not how freewrl is working, not my understanding. In
the Precision section of Geospatial, it talks about GeoOrigin (deprecated
in v3.3) or something done automatically in the place of GeoOrigin. That's
called the Local Coordinate System. Not at the center of the earth. Its
somewhere nearby to the scenery you have on the surface of the earth, to
reduce coordinate precision required (from double to float) as the specs
describe.
In freewrl, in place of deprecated GeoOrigin, freewrl uses the TCS
-Topocentric Coordinate System, as described for GeoLocation- of the first
GeoNode it parses, or more precisely, the TCS of the first GeoNode in the
planet, as the arbitrary LCS origin.
Then non-geo nodes are in this LCS by default.


> I think regular coordinates within a GeoPlanet node would be geocentric
only
> to that planet.
They are LCS. So one job of GeoPlanet is to convert from LCS to GC /
geocentric - or more precisely push onto the transform stack a transform
that will do that - so the parent scenery sees the planet in geocentric
coords like you were thinking.

> In a multiplanet scene it would be necessary to also
> specify the positions of the planets relative to each other.
Yes. Once each planet is wrapped in GeoPlanet then you can use normal
transforms to position the planets, and scripts to animate the planets.
The EarthRise_GP.x3d scene I mentioned does that - its got moon, sun, and
earth (although at shorter non-realistic distances).
I haven't done a formal release of a freewrl version that does all that,
but the develop version has been doing it all since March.


> GeoPositionInterpolator converts to GC
Really? Uh-oh. I missed that. I thought it was just working in double
precision coordinates, and left it up to you to decide GD, GC or XTM, but
without converting anything for you.
That's the main benefit I was targeting with the GeoConverter node
proposal. So if you do need conversion its available. And for browser
developers, its one of the easiest GeoNodes to add, because you already
have the conversion functions internally.


> relativeHeight: sounds interesting. Why use the highest point on the
> GEG and not the height at the actual location ?
It does/should use the GEG height at your location. But if you had
overlapping GEGS then takes highest one at your location.
That's handy if you have a low-resolution GEG covering for example the
whole planet, and want to put a high res GEG for a small local area of
interest. Rather than cut a hole in a GEG (awkward code), you can lower the
heights a bit, and your higher res GEG will take over.

> What if the registered
> GEG is somewhere else than the GL ? Just fall back to
> Geoid,eg.sealevel ?

Yes - fall back to relative-to-sealevel, for GL (or if you know you're not
going to be on a GEG, use absolute / leave .relativeHeight FALSE)
For GVP, a more common use case is starting out over a GEG and running off.
Lets say you start out on a small local GEG, with relativeHeight 100, and
you're in walk mode, and feeling the bumps in the terrain as you navigate
along. Then you hit the edge of the GEG and want to keep going. In freewrl,
with no GEG adjusting the height, you stay at the absolute (above sealevel)
height you were at when you hit the edge. If you turn and go back on at a
different place, the relative height you specified kicks back in, and
you're back at your 100m height walking on the GEG.








On Sun, May 13, 2018 at 1:50 PM, Andreas Plesch <andreasplesch at gmail.com>
wrote:

> Hi Doug,
>
> here is my unfinished response to your earlier posting:
>
> there may be planetary geoSystems or spatial reference frames proposed
> in the enhanced geospatial component proposal.
> http://www.igraphics.com/Standards/EnhancedGeospatialComponent_
> 2007_10_30/Part01/components/geodata.html
> http://www.sedris.org/wg8home/Documents/18026_FDIS/C030811e_
> FILES/MAIN_C030811e/ISOIEC_18026E_TOC.HTM
>
> But a more streamlined proposal seems very useful.
>
> I think geospatial nodes can only deal with scenes on a single planet
> since the underlying assumption is that in the scene regular (nongeo)
> coordinates are geocentric, eg. relative to the center of a single
> planet. Your proposal seems to address somewhat that question. I think
> regular coordinates within a GeoPlanet node would be geocentric only
> to that planet. In a multiplanet scene it would be necessary to also
> specify the positions of the planets relative to each other.
>
> An expansion may be a solarcentric mode, where the center of a solar
> system is the origin. Each planet would have its center given relative
> to a central sun (ignoring precision issues for the moment). I think
> astronomy also uses galactic coordinates.
>
> And here additional feedback:
>
> 3TM and 10UTM seems to be only used in Canada ? 3TM seems to have
> meridians based on province ? If we allow additional projections, they
> should be truly global and well defined. I do not see it as a big
> burden to convert to GD or UTM. The other route to is allow all epsg
> codes which can be converted by a library such proj which means
> essentially all possible projections.
>
> GeoPositionInterpolator converts to GC but there is no way to convert
> to UTM or GD which may be useful for reporting coordinates.
>
> relativeHeight: sounds interesting. Why use the highest point on the
> GEG and not the height at the actual location ? What if the registered
> GEG is somewhere else than the GL ? Just fall back to
> Geoid,eg.sealevel ?
>
> -Andreas
>
> > Date: Sun, 13 May 2018 12:10:21 -0600
> > From: GPU Group <gpugroup at gmail.com>
> > To: X3D Graphics public mailing list <x3d-public at web3d.org>
> > Subject: Re: [x3d-public] Invalid geoSystem values in X3D Resources
> >         BasicGeospatial examples.
> > Message-ID:
> >         <CAM2ogRfXeWAaqvS9WvLRrYUoGXMNHb1NWyR7+aFDTos_-mp3zg at mail.
> gmail.com>
> > Content-Type: text/plain; charset="utf-8"
> >
> > John, thanks for checking.
> > Hypothesis: Maybe I forgot to check the [x] public checkbox when making a
> > spec comment? I meant for it to be public, And I hope the recommendations
> > are adopted in some future version of the specs, and implemented by
> browser
> > developers.
> > I'll make it public here and do another submission.
> > -Doug
> > Here's what I submitted Mar 26, 2018
> >
> > ==============================================================
> >
> > GeoSpatial Component Comments Mar 2018
> >
> >
> >
> > Acronyms:
> >
> > GL - GeoLocation
> >
> > GVP - GeoViewpoint
> >
> > GEG - GeoElevationGrid
> >
> > GPS - GeoProximitySensor
> >
> > GP - GeoPlanet
> >
> > XTM - UTM or 3TM
> >
> > 3TM - like UTM, except 3 degree zones, .9999 scale factor and no false
> > northing / false easting
> >
> > user coords - coords matching geosystem, authored in scene file, could be
> > GC, XTM, GD
> >
> > TCS/LGS - Topocentric Coordinate System aka Local Geodetic System, as
> > described in specs for GeoLocation:
> >
> > -- tangent to ellipsoid at a given geo location, -Z north, Y up
> >
> > LCS - Local Coordinate System - as described in specs for numerical
> > precision
> >
> > -- shared cartesian coordinate system used internally
> >
> >
> >
> > A. Proposed nodetype: GeoPlanet
> >
> > Implicit in the web3d geospatial specs: all geoNodes are on the same
> > planet.: earth That should cover 99.99999% of uses. And even if you have
> 2
> > planets, the transform stack will keep them separate for most use
> cases.  The
> > exception being node-node interactions such as GVP-GEG,  GL-GEG, and
> > GPS-GEG as I'll explain later for the relativeHeight feature.
> >
> > x3d usage:
> >
> > <GeoPlanet planetId='#'>
> >
> >  ... geonodes for planet...
> >
> > </GeoPlanet>
> >
> > where
> >
> > planetId - an Int32 field that can be used to discriminate between
> planets,
> > could also be called planetName or planet
> >
> > example scenefile:
> >
> > http://dug9.users.sourceforge.net/web3d/tests/geo/World/earthriseGP.x3d
> >
> > http://dug9.users.sourceforge.net/web3d/temp/freeWRL2018_GeoPlanet.mp4
> > feature details:
> >
> > 1. similar to GeoTransform (GT) it wraps geonodes, but instead of
> > converting to TCS/LGS it converts to GC
> > - allows orbital mechanics on multiple planets using planet centers and Z
> > as rotation axis
> >
> > - same job could in theory be done by modifying the specs for
> GeoTransform,
> > to allow converting to GC instead of TCS/LGS, and a field for planetId,
> but
> > adds a connundrum: 2 geoTransforms for the same planet could sport
> > different translation, rotation etc, which doesn't make sense.
> > 2. for node-node interactions such as GVP-GEG and GL-GEG helps keep the
> > planets separate
> >
> > - internally planetId is pushed and popped during scenegraph traversal,
> so
> > when visiting a the code can tell which planet it's working on
> >
> > a) for GVP-GEG:
> >
> > -- when visiting GVP to start the modelview matrix, store the planetId in
> > the GVP
> >
> > -- when visiting GEG, check the current planetId (applies to GEG) against
> > the GVP._planetId, use if same
> >
> > b) for GL-GEG:
> >
> > -- after loading scene/inline/externProto body, on first traveral add
> each
> > GEG to a per-planet list of GEGs
> >
> > -- when visiting a GL with relativeHeight TRUE, look through the current
> > planet's list of GEGs, and get the highest terrain height, if any GEGs
> > covering that location.
> >
> >
> >
> > GeoPlanet : X3DGroupingNode {
> >   SFInt32  []        planetId       0           [0,?)
> >   SFNode   [in,out] metadata       NULL        [X3DMetadataObject]
> >   MFNode   [in]     addChildren                [X3DChildNode]
> >   MFNode   [in]     removeChildren             [X3DChildNode]
> >   MFNode   [in,out] children       []          [X3DChildNode]
> >   SFVec3f  []       bboxCenter     0 0 0       (-?,?)
> >   SFVec3f  []       bboxSize       -1 -1 -1    [0,?) or ?1 ?1 ?1
> > }
> >
> >
> >
> > B. Proposed nodetype: GeoConvert
> >
> > Web3d specs allow a different geoSystem on each geo node. And each as
> > fields in user/geoSystem coordinates that can be routed to/from. But no
> way
> > to convert between geoSystems.
> >
> > Proposed GeoConvert node
> >
> > - allows you to convert between GeoSystems via routing and 2 GeoConvert
> > nodes
> > myGeoNode1 -> set_geoCoord (geoConvert1) gcCoord_changed -> set_gcCoord
> > (geoConvert2) geoCoord_changed -> myGeoNode2
> > where geoConvert1.geoSystem == myGeoNode1.geoSystem
> > and geoConvert2.geoSystem == myGeoNode2.geoSystem
> > http://dug9.users.sourceforge.net/web3d/tests/geo/World/
> World33geoConvert.x3d
> >
> >
> > GeoConvert : X3DNode {
> >   SFVec3d  [in]     set_geoCoords
> >   SFVec3d  [in]     set_gcCoords
> >   SFNode   [in,out] metadata       NULL        [X3DMetadataObject]
> >   MFString []       geoSystem      ["GD","WE"] [see 25.2.3]
> >   SFVec3d  [out]    geoCoords_changed
> >   SFVec3d  [out]    gcCoords_changed
> > }
> >
> >
> > C. Proposed field for GVP, GL and GPS: .relativeHeight
> >
> > Sometimes you don't know the exact height of something above the
> ellipsoid,
> > but you know you want it at ground level, and you have a GEG in the scene
> > covering the area. It would be convenient to put 0 for the
> geoCoords/center
> > height, if they are GD or XTM user coords (not GC), and have the browser
> > adjust the height to terrain level.
> >
> > For GL and GPS if relativeHeight = TRUE, then on each frame when visiting
> > the node, check the list of GEGs registered for the current GeoPlanet,
> and
> > add the highest of those (if any) to the .geoCoords / .center height if
> in
> > GD or XTM, to convert from user to internal GC, and subtract if going
> from
> > GC back to user.
> >
> > For GVP, the behaviour is similar, but only for initial positioning,
> > because browser-supported navigation modes likely want to adjust the
> > relative height:
> >
> > - WALK + Collide:  maintain relative height to terrain, by
> climbing/falling
> > while navigating on GEG
> >
> > - FLY + Collide: maintain height relative to ellipsoid
> >
> > - FLY (no collide aka FREEFLY): allow height to be adjusted
> >
> >
> >
> > Example scene (GVP and GL):
> >
> > http://dug9.users.sourceforge.net/web3d/temp/freeWRL2018_
> relativeHeights.mp4
> >
> > http://dug9.users.sourceforge.net/web3d/tests/geo/World/
> World33GLrelativeHeight.x3d
> >
> >
> >
> > Field descriptions:
> >
> >
> >
> > GeoLocation  : X3DGroupingNode {
> > .. as per specs ..
> >   SFBool   []       relativeHeight FALSE
> > }
> >
> >
> > GeoViewpoint : X3DViewpointNode {
> > .. as per specs ..
> >  SFBool    []       relativeHeight FALSE
> > }
> >
> >
> > GeoProximitySensor  : X3DEnvironmentalSensorNode {
> > .. as per specs ..
> >   SFBool   []       relativeHeight FALSE
> > }
> >
> >
> >
> > D. Proposed geoSystem additions
> >
> > http://www.web3d.org/documents/specifications/19775-1/V3.3/Part01/
> components/geodata.html#Spatialreferenceframes
> >
> >
> >
> > D1. A, IF, F, R
> >
> > An assumption of the specs: all geoNodes are for one planet: earth, and
> the
> > list of allowed ellipsoids are for earth.
> >
> > Online samples include a Mars scene, and it uses default earth ellipsoid:
> >
> > http://www.web3d.org/x3d/content/examples/Basic/#Geospatial
> >
> > Here are some additions to geoSystem to allow users to specify the
> > ellipsoid parameters:
> >
> > # - a number like a 6378137 or 298.257223563
> >
> > "A#" - semimajor axis
> >
> >  "IF#" - inverse flattening
> >
> > "F#" - flattening
> >
> > "R" - radius
> >
> > coherent combinations include:
> >
> > A and F
> >
> > A and IF
> >
> > R
> >
> >
> >
> > D2. 3TM
> >
> > 3TM is like UTM, except with (built-in) central meridan scale factor
> .9999
> > (instead of .9997 for UTM) and no false easting or false northing (so
> > coordinates can be negative). Typically used for cities, and GIS packages
> > export 3TM often without mentioning the zone or central meridian. In that
> > case you can set Z0. Otherwise its generally 2x the UTM zone or 2x- 1.
> >
> > coherent combinations include
> >
> > 3TM
> >
> > 3TM and Z
> >
> > along with other UTM-compatible options except no FalseEasting or
> > FalseNorthing.
> >
> >
> >
> > ================================================================
> >
> > and April 9, 2018 addendum:
> >
> > ======================================
> >
> > (refering to a previous spec submission from the same commentor which
> > included new nodetype GeoPlanet)
> >
> > GeoPlanet - there's a preference among freewrl users that instead of
> > numerical planetId='#' it would be an SFString ie planet='earth', which
> > would be easier for scene authors to keep straight in their own minds.
> >
> > SFString planet "earth"
> >
> > ========================================
> >
> >
> >
> > On Sun, May 13, 2018 at 10:58 AM, John Carlson <yottzumm at gmail.com>
> wrote:
> >
> >> You may want to resubmit your 3TM and Planet changes to Mantis.  I
> >> searched for 3TM and Planet in Mantis (All Projects) and got nothing
> >> significant back.
> >>
> >>
> >>
> >> However, I was denied access to:
> >>
> >>
> >>
> >> http://www.web3d.org/node/1694/submission/1671
> >>
> >>
> >>
> >> Why?
> >>
> >>
> >>
> >> and there?s:
> >>
> >>
> >>
> >> http://www.web3d.org/member-only/mantis/view.php?id=1215
> >>
> >>
> >>
> >> Not that I want to imagine writing schema for such a thing?that?s why I
> >> hope to get geoSystem into more reasonable shape in the Unified Object
> >> Model (we need volunteers, I think) before trying to do it in schema.
> >>
> >>
> >>
> >> John
> >>
> >>
> >>
> >> Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986> for
> >> Windows 10
> >>
> >>
> >>
> >> *From: *GPU Group <gpugroup at gmail.com>
> >> *Sent: *Sunday, May 13, 2018 9:16 AM
> >> *To: *X3D Graphics public mailing list <x3d-public at web3d.org>
> >> *Subject: *Re: [x3d-public] Invalid geoSystem values in X3D Resources
> >> BasicGeospatial examples.
> >>
> >>
> >>
> >> I think N should be OK, just as you say a redundant default.. I proposed
> >> several new things in the geoSystem field, submitted to spec comments a
> few
> >> months ago, so I hope your processor works on the new things if ever
> >> adopted.
> >>
> >> -Doug
> >>
> >> R,A,B,F,IF - ways to drectly specifiy the shape of the ellipsoid so
> other
> >> planet shapes and sizes could be modeled
> >>
> >> 3TM - similar to UTM (except 3degree zones, no false northing or
> easting,
> >> different central meridian scale factor .9999)
> >>
> >> ...
> >>
> >> I also proposed a Planet node - allows multiple planets in one scene
> >>
> >>
> >>
> >> On Sun, May 13, 2018 at 6:14 AM, John Carlson <yottzumm at gmail.com>
> wrote:
> >>
> >> The object model is expressive enough to list conforming values, yet
> >> accept others, but as far as I know (I may be wrong here), it is not
> >> expressive enough to express supported geoSystem values.   This is a
> >> problem.   How do we resolve it, and should we, if we are accepting
> other
> >> values?   Is it resolved in X3DJSAIL?  Should I get rid of Roy's JSON
> >> subschema for geoSystem?
> >>
> >>
> >>
> >> John
> >>
> >>
> >>
> >> On Sun, May 13, 2018, 7:43 AM John Carlson <yottzumm at gmail.com> wrote:
> >>
> >> I believe the purpose of schema is to support minimally acceptable
> files.
> >>  It should flag files which are questionable.   If a file can be brought
> >> into conformance easily, shouldn't it be?
> >>
> >>
> >>
> >> Yes, you should be able to ignore the results of the schema
> >> validation...at your own risk.
> >>
> >>
> >>
> >> What is the purpose of the X3D resources examples, but to show a good
> >> example practice?  Or are we testing tools to make sure certain values
> are
> >> acceptable?   If that's the case, then the schema should be updated for
> ALL
> >> versions.
> >>
> >>
> >>
> >> John
> >>
> >>
> >>
> >> On Sun, May 13, 2018, 7:31 AM John Carlson <yottzumm at gmail.com> wrote:
> >>
> >> Then the version of the document should be upgraded to 4.0 or above, and
> >> the corresponding schema updated.   I can upgrade the schema for all
> >> versions, since it is hard coded into the schema generator.   But
> really we
> >> need support from the object model if possible.   Right now, I have to
> >> explicitly allow it for all versions, since I use stdin/stdout instead
> of
> >> files.  Does the unified object model specify a version # in its
> contents?
> >> When the unified object model supports "N" in a usable fashion, then I
> can
> >> code something into the schema, or delete the requirement to check
> >> geoSystem.
> >>
> >>
> >>
> >> Thanks, away from computer presently, or I would check.
> >>
> >>
> >>
> >> I don't see why Roy went to all the effort to create a schema, if we are
> >> going to ignore it?
> >>
> >>
> >>
> >> This applies in many areas...if we merely have supported values, yet
> >> others must be accepted, why is there a standard?
> >>
> >>
> >>
> >> I don't believe it's ambiguous in this case of X3D.   A missing "N"
> means
> >> Northern Hemisphere.
> >>
> >>
> >>
> >> John
> >>
> >>
> >>
> >> On Sun, May 13, 2018, 12:46 AM Andreas Plesch <andreasplesch at gmail.com>
> >> wrote:
> >>
> >> http://www.web3d.org/specifications/19775-1/V3.2/
> >> Part01/components/geodata.html#Specifyingaspatialreference
> >> <http://www.web3d.org/documents/specifications/19775-1/V3.2/Part01/
> components/geodata.html#Specifyingaspatialreference>
> >>
> >>
> >>
> >> lists the supported strings for the geoSystem MFString field.
> >>
> >>
> >>
> >> The prose choosing 'supported' over 'legal' or 'conforming' could be
> taken
> >> to mean that other than listed strings may be allowed to be supported as
> >> well by some browsers.
> >>
> >>
> >>
> >> However, previous discussions indicate that schemas are not sufficiently
> >> expressive to describe unknown but conforming string values.
> >>
> >>
> >>
> >> One resolution was to explicitly allow 'N' in upcoming X3D version, and
> >> perhaps silently allow it for 3.3.
> >>
> >>
> >>
> >> Outside of X3D UTM zones often have the N hemisphere identifier to avoid
> >> ambiguity.
> >>
> >>
> >>
> >> Andreas
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >> Date: Sat, 12 May 2018 19:54:17 -0400
> >> From: John Carlson <yottzumm at gmail.com>
> >> To: Don Brutzman <brutzman at nps.edu>,  X3D Graphics public mailing list
> >>         <x3d-public at web3d.org>
> >> Subject: [x3d-public] Invalid geoSystem values in X3D Resources Basic
> >>         Geospatial examples.
> >> Message-ID: <5af77ea8.1c69fb81.7da1f.13e3 at mx.google.com>
> >> Content-Type: text/plain; charset="utf-8"
> >>
> >> Don, These files contain "N" in geoSystem, which I believe is not valid,
> >> and should be removed (default is Northern Hemisphere,  in the
> standard, I
> >> believe. "S" can be specified?not in this case).
> >>
> >> Is there a tool which hasn?t been changed which is
> >>
> >> _______________________________________________
> >> x3d-public mailing list
> >> x3d-public at web3d.org
> >> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
> >>
> >>
> >> _______________________________________________
> >> x3d-public mailing list
> >> x3d-public at web3d.org
> >> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
> >>
> >>
> >>
> >>
> >>
> > -------------- next part --------------
> > An HTML attachment was scrubbed...
> > URL: <http://web3d.org/pipermail/x3d-public_web3d.org/
> attachments/20180513/8defb5db/attachment.html>
> >
> > ------------------------------
> >
> > Subject: Digest Footer
> >
> > _______________________________________________
> > x3d-public mailing list
> > x3d-public at web3d.org
> > http://web3d.org/mailman/listinfo/x3d-public_web3d.org
> >
> >
> > ------------------------------
> >
> > End of x3d-public Digest, Vol 110, Issue 31
> > *******************************************
>
>
>
> --
> Andreas Plesch
> Waltham, MA 02453
>
> _______________________________________________
> x3d-public mailing list
> x3d-public at web3d.org
> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20180513/75e39192/attachment-0001.html>