Page 6 of 9

Re: 3D spacial modeling of a giant redwood trunk

Posted: Wed Jan 25, 2012 5:53 pm
by Don
Ed/Mike-
I had similar thoughts...in many ways quite analogous to that employed by the Nat Geo photographer, that resulted in my all-time favorite magazine foldout!
-Don

edfrank wrote:Perhaps you could run the scanner up and down a fixed line suspended from tree and attached at its base.

Re: 3D spacial modeling of a giant redwood trunk

Posted: Wed Jan 25, 2012 10:02 pm
by M.W.Taylor
Don and Frank,

Yes, this is exactly what I had in mind. A complete tree scan is currently in the works. Climber deployed linear scanning up the rope is going to be an important part of the project.

Michael



Don wrote:Ed/Mike-
I had similar thoughts...in many ways quite analogous to that employed by the Nat Geo photographer, that resulted in my all-time favorite magazine foldout!
-Don

edfrank wrote:Perhaps you could run the scanner up and down a fixed line suspended from tree and attached at its base.

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 12:00 am
by mdvaden
Michael,

I'll probably need to have you show me in the field how you do this sometime. It's a lot to soak in all at one time.

It would be cool if a program could take some of the data you got, like on Drury Tree, then if a .jpg photo were imported, it could mould the image like a 3D skin.

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 3:35 am
by fooman
It would be cool if a program could take some of the data you got, like on Drury Tree, then if a .jpg photo were imported, it could mould the image like a 3D skin.
Meshlab will do this, like most other 3d modelling packages. Working out how it does it will be the issue!

Cheers,
Matt

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 2:46 pm
by M.W.Taylor
Matt and Mario,

The existing online manuals that I have seen so far tell me nothing about how to do that. There is a MeshLab blogsite/help forum for programmers. That may be the best place to get the nuts and bolts on picture importation and texture applications.

I attached the latest volume solution for Drury Tree. The default example solves for volume between 0 ft and 2 ft. You can change the ray angle, clice thickness and height interval. You'll notice an increase in volume for a smaller ray angle which is essentially the number of triangles the cross-sectional cookies are fitted to. The smaller the angle the finer the fit. For a ray angle of 1 degree that would equate to 360 sub-triangles to fit the polygon.

For 0-5th height solution it takes about 10 minutes to finish the calculation. Be patient or use Control-Alt-Del. User input ray angle of 10 degrees = volume of 1027 cubic feet. Ray angle of 5 degrees=1042 cubic feet. Ray angle of 2 degrees = 1050 cubic feet. Rayangle of 1 degrees = 1052. As you can see the gain in precision is not linear in proportion to the decrease in ray angle. One side of the tree has low cloud density so the software interpolates a few of the in-between points(using the most conservative curve fitting possible).

There is a point where there is little to be gained with a finer ray angle. For the Drury cloud set in the attachment example, it appears 1 degree is about as small as you would want to go. It would slow the program down tremendously and give no added value to go .5 or .1 degrees for this sparse cloud set.

Press Control-V to solve for volume. Be patient. The program takes about 30 seconds to display area and volume updates as it constructs the various polygons for each cookie wedge. Solution takes about 2 minutes per foot of height intervals. You can see the current working height interval status on row M7 to gauge time left for volume solution.
freformvolumesolver.JPG
The cloud set acquired from Drury is not yet dense enough to solve for volume above 10 feet. Not enough points to interpolate the cross-sectional wedges...yet.

Michae Taylor

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 4:28 pm
by edfrank
Mike,

The attachment was missing. If it is some odd extension, I can add it to the allowed extension listings.

Ed

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 7:07 pm
by M.W.Taylor
Ed,

I had some problems uploading the latest volume solver spreadsheet earlier today. But now it has worked.

Also, I deleted a few older attachments to free up more memory for future attachmets in effort to stay under the 15mb limit. This attachment replaces the older version of the volume solver with a more stable version.

Michael

edfrank wrote:Mike,

The attachment was missing. If it is some odd extension, I can add it to the allowed extension listings.

Ed

Re: 3D surface modeling of a giant redwood trunk

Posted: Sat Jan 28, 2012 8:41 pm
by edfrank
Mike, there is a limit to the size of an individual attachment, but there is not one for a topic or form, so you should not need to delete any older attachments. But it is good to go ahead and delete outdated or incorrect attachments and replace them.

Ed

Re: 3D surface modeling of a giant redwood trunk

Posted: Sun Jan 29, 2012 8:48 pm
by M.W.Taylor
Ed et al,

The attached is a more stable version of the Free Form volume solver spreadsheet. The default cloud is 198k points for the lower trunk of Redwood Creek Giant. I would recommend replacing the older version with this one if you already downloaded the older version. Be patient. The solution takes a long time, especially if you choose fine intervals. You can see the current working height interval to estimate time left for solution. You can see the current cross-sectional area being graphed as its upper and lower areas are calculated as well as volume.

I am working on a new revision of the Free Form Volume Solver that will allow solutions for trees with multiple iterations and any type of complex geomorphic structure.

Michael Taylor

WNTS VP
http://www.landmarktrees.net

Re: 3D surface modeling of a giant redwood trunk

Posted: Mon Jan 30, 2012 5:07 pm
by Amy69
Wow, that is incredible cool! And very useful I am sure as well!