TransDEM Forum

Hello Roland,

I'm just wondering :

Will it be possible in a future version of transdem to do

a) In the section export to trainz Vector Data ( Tracks and other splines ) : possibility to add a kuid2 number

b) in the section route editor : make a copy of an existing polyline and be able to paste it a few centimeters to the left/right of the existing polyline.

This would make it possible to place a track using the export to trainz vector data option and also a catenary, and all in a fast way.

Now placing a catenary ( pylon+catenary+wire ) takes a lot of time.

Best regards

Kurt

I am currently finalizing the doc update for TransDEM 2.3 which will bring the new functions for GPS and for complex georeferencing using triangular networks.

Internally TransDEM processes KUID and KUID2 evenhandedly. You should normally not need to specify a KUID2 version number, since the latest is the one to go for. TransDEM 2.3 will fix a shortcoming, though, allowing the "-1" to be set as the User-ID.

You can already do that. Select the polyline, press [Ctrl] and drag a copy of the polyline to the position you like. However, this would be an exact copy of the original. What you may have in mind is a copy that observes perpendiculars and angle bisectors. Though not very complicated, I think that's beyond the scope of TransDEM. But with TransDEM's support of vector data file formats, it should be easy enough to accomplish such tasks with an external tool.

I wanted to make an exact copy of the route, however I overlooked one thing ( ), when using the section export to trainz vector data, both routes will be placed, for example as a track. It's not possible to get one part as track and the other part as catenary. Again, I have to go back to my drawingboard. ok - came back from the drawingboard - Saving each route separately is an option and does work. I manage to get a track and a catenary using the trainz vector data option


Best regards

Kurt

It's for distorted track schemas. I wrote about it in German:
http://forums.auran.com/trainz/showthread.php?92430-TransDEM-Vorschau-Labor

That should work, but I'll check again.

No, indeed. That would mean a new layer system. Something which is on my list for quite some time.

Hello Roland,




Hmm, a new layer system and it's on your list. Is this a tough element to incorporate?

Best regards

Kurt

Well, one mathematical principle is the Triangular Irregular Network (TIN). It has been implemented in TransDEM from the very beginning as it is the base math tool for most DEM editing.

The other principle is that of an affine transformation. As long as you use online geo data sources you may not be confronted with the transformation, but as soon as you start using the 3 (+1) georeferencing method, you will implicitly use it. Each reference point has two coordinate pairs, one is the pixel coordinates, the position within the image, the other is the geo coordinates of that same point. To determine the parameters of an affine transformation, you need three reference points, each with the two coordinates pairs. This gives you six equations with as many variables to find. These are: Translation: Shift of origin between pixel and geo coords (2 vars), rotation (1 var), scaling for x and y axis (2 vars) and shear (trapezoidal factor) for one axis (1 var).

Now, combine triangular network and affine transformation and you can allow many more than 3 georeferencing points, yielding an individual transformation for each triangle. The triangles are formed on the fly, simply taking the nearest reference points.

Where can you use it? The easier variant is to compensate slight distortions which happen when scanning a paper map. The other application field is the heavy stuff: distorted track plans, where scale is different along the track and perpendicular to the track. The new georeferencing will offer help here, but it remains a tough task which still requires accuracy, diligence and patience.

Problem is, there is a layer system already in place, but it's hidden under the surface and only one layer (and its sub-layers) is actually in use by TransDEM, but due to the current form of embedding this structure, other layers cannot be made use of. That needs some serious reworking and therefore gets postponed.

Hello Roland,

Thank you for some more explantion. Now, I went back in time, but I can't remember having see something about TIN during the maths lessons. I even had a look in my math geometry book that I used at the time. Nothing about TIN. Must be maths of a higher level.



Problem is, there is a layer system already in place, but it's hidden under the surface and only one layer (and its sub-layers) is actually in use by TransDEM, but due to the current form of embedding this structure, other layers cannot be made use of. That needs some serious reworking and therefore gets postponed.[/quote]

Absolutely a very logic point of view to postpone the rework.

Best regards

Kurt

No, it's probably not something you will be taught in school math. TINs are a variant of the ubiquitous triangles in computer algorithms and 2D and 3D graphics, a field of Applied Mathematics. There is a lot of science and academic research behind the computer games we are often not aware of. A keyword for TINs is "Delaunay triangulation", and you'll find articles on both topics in wikipedia.

Hello Roland,

Obiquitous triangles in computer algorithms. No, never saw anything about that. Thanks for the keyword Delaunay triangulation, which gives me the oppertunity to read more about it.

Best regards

Kurt