[LandXML] Re: Topcon landxml flawed?
- From: "Shane" <shane@xxxxxxxxxxxxxxxxx>
- To: <landxml@xxxxxxxxxxxxx>
- Date: Wed, 31 Oct 2007 08:52:42 -0000
Patrick,
This looks very much like a file generated to test the data format
rather than a real survey. If you look at the CGPoints group called
received, this is pretty clear. I modelled the triangulation and it
certainly doesn't conform to a Delaunay Triangulation (see
http://en.wikipedia.org/wiki/Delaunay_triangulation), but the LandXML schema
makes no reference to a TIN surface having to be either well formed or
mappable in plan. I think that this is on balance a good thing, as there
are many TIN models out there that are not projected to plan, or even
mappable to any plane. For example, there is nothing to say you can't have
a group of triangles that fully represent a road with a bridge over it.
While traditional survey software might struggle to contour, most 3d viewing
applications will have no problems modelling it.
It might prove a useful addition to the LandXML schema to state whether a
TIN surface corresponds to a constrained Delaunay representation, is
mappable in plan (i.e. when viewed from above no triangles overlap), is
mappable projecting to another described plane (i.e. facade of a bridge or
building), is mappable to the inside of a described cylinder, (i.e. inside
of a tunnel), or is non-mappable. LandXML 1.3 perhaps?
In my opinion, the TopCon file still represents a very useful example for
software developers such as myself, but probably does not have much useful
value beyond that. Testing samples from different vendors is valuable for
demonstrating compatibility with those vendors as well as potential
weaknesses in ones own implementation. The LandXML definition is broad
enough that there are significant variations in how a given survey object
can be described. I wouldn't agree that the Topcon data is flawed, just
that it has a limited audience that will find it useful.
Best regards,
Shane
-----Original Message-----
From: landxml-bounce@xxxxxxxxxxxxx [mailto:landxml-bounce@xxxxxxxxxxxxx]On
Behalf Of Patrick Bergeron
Sent: 30 October 2007 20:17
To: landxml@xxxxxxxxxxxxx
Subject: [LandXML] Topcon landxml flawed?
Hello,
I am currently looking at a sample LandXML from TopCon, but particularly
the Surface definition (search for "TIN")
http://www.landxml.org/schema/LandXML-1.0/samples/Topcon/survey.xml
I see the triangle definitions pointing to these points: 1,2,3 4,5,6
7,8,9 10,11,12 etc.
One can also clearly see that point "3", "4","8" and "17", are in fact
duplicates.
Basically, this creates a logically disconnected (but physically
connected) triangle mesh. Isn't there be a requirement that each point
only appear once, and that triangle indices refer to the same point instead
of duplicated points?
Thank you very much
Patrick Bergeron
Other related posts:
