3D surface modeling of a giant redwood trunk

General discussions of measurement techniques and the results of testing of techniques and equipment.

Moderators: edfrank, dbhguru

fooman
Posts: 100
Joined: Sun Dec 25, 2011 2:18 pm

Re: 3D spacial modeling of a giant redwood trunk

Post by fooman » Fri Dec 30, 2011 9:05 pm

All,

Somevery intersting stuff going on here, and it is starting to get very similar to the stuff I do for a living. I work as an engineer, and part of the work we do involves reverse engineering of reasonably complex engineering components (pressure casings, turbine scrolls) to help assess remenant service life. One way we capture the geometrical information is using a handheld 3d-laser scanner (http://www.zcorp.com/en/Products/3D-Sca ... spage.aspx). These can be used for scanning objects up to car-sized at quite high resolution. The resulting point clouds are then imported into reverse engineering software, and the analysis can be performed on the resultant 3-d models. The particular method of scanning used in this product is resonably manually intensive - reflective spots placed approximately 4 inches apart over the surface to be scanned is required. This is not our core business, but it is a valuable tool we can draw upon as required. The company producing this scanner advertises a variant that can be used for recording archeological geometry

I know of another engineering company in Europe doing reverse engineering of large power station turbine rotors (3 or 4 m in diameter, 10 m long) using (semi)portable structured light scanning (http://en.wikipedia.org/wiki/Structured ... 3D_scanner).

Another company I used to work for was developing another handheld laser scanning system, and was looking at room-sized objects (e.g. http://www.irl.cri.nz/newsroom/news/set ... revolution).

These techniques are all used to automate the manual acquistion process that Michael is performing. I do not know of any particular outdoor scanners, e.g. for surveying/mapping, but I know they exist, and are available at a price - our basic laser scanning setup at work is more than $100k, hardware plus software, but this is for the required resolution - tree measurement may not be as demanding.

Cheers,
Matt

User avatar
M.W.Taylor
Site Admin
Posts: 530
Joined: Sun May 29, 2011 11:45 am

Re: 3D spacial modeling of a giant redwood trunk

Post by M.W.Taylor » Sat Dec 31, 2011 3:07 am

Hi Matt,

I am in contact with several companies about Flash 3D LIDAR for tree modeling. That is the ultimate way to map a tree trunk but the range on these cameras is limited and shooting angles restricted. These cameras are improving rapidly though!

Each pixel has azimuth, distance and inclination embedded in the image so you can contruct a 3D image. The Swiss Ranger 4000 takes a 320 x 160 picture so that would be 38,000 pixels with angle, distance and inclination angle in one shot !!! Range is limited though...30 foot is longest range 3D flash LIDAR I know of and you can only point at two angles, 49 degrees and 63 degrees. These cameras are expensive too. They would work great for modeling interior hollows, cavities and burls.
drury4.JPG
drury4.JPG (40.62 KiB) Viewed 2075 times
Matt, what is the range on your company's 3D flash LIDAR ?

I attached the latest version of the Giant Redwood 3D Graph, but also with an unlocking Y-axis so you can fly over the tree and look down like on Google Earth. This new perspective is useful for identifying flying pixels for removal prior to volume solving.

Michael Taylor
Attachments
Forest Mapper Drury Tree- MacB.xls
Mapping A Giant Redwood- 3D Graph
(2.53 MiB) Downloaded 59 times

fooman
Posts: 100
Joined: Sun Dec 25, 2011 2:18 pm

Re: 3D spacial modeling of a giant redwood trunk

Post by fooman » Sat Dec 31, 2011 2:41 pm

Hi Michael,

The laser scanner we use is not a Lidar scanner - it does not use a reference position to generate the point cloud. It provides self positioning data via reflective spots that are randomly positioned on the scanned object. Once these points are registered the laser beam is scanned across the object, much like a spray can, "painting" the scanning beam across the object. This generates reasonably high resolution data (sub-millimetric) very rapidly - maybe a million points in an hour or so. But not really suitable for a large tree.

Another scanner I came across used self positioning in the magnetic field as some sort of golbal reference, but it could not scan metallic objects, so wasn't much use for us.

I've just remembered that our inspection division used to scan the inside of delayed coker drums (large refinery vessels - maybe 20-30 m tall, 4-5 m in diameter) using a laser process - a rotating prism was used to scan a beam around the inside circumference of the vessel, while being dropped down the axis, to pick up cracking/bulging in the vessel wall.

I do think structured light scanning is a real possibility for large trees - all that is needed is a reference dimension (e.g. taped girth) or length between two features on the trunk, a method to project a light pattern (a large projector at night?), and two cameras recording the image at a known baseline plus the software. The guys at the company I referred to in my earlier post were just using normal cameras mounted on tripods a known distance apart. The resolution of a scanned tree would be lower than the smaller objects normally scanned (e.g. people) but I guess +/- 1 inch would be ok for volume, rather than +/- 1 mm?

Cheers,
matt

User avatar
M.W.Taylor
Site Admin
Posts: 530
Joined: Sun May 29, 2011 11:45 am

Re: 3D spacial modeling of a giant redwood trunk

Post by M.W.Taylor » Sat Dec 31, 2011 4:24 pm

Matt,

The 3D Flash LIDAR name is a misnomer. The company puts LIDAR in there for the name but technically speaking the pixels are not encoded with laser based distance returns in the same way as a typical LIDAR array on an airplane would work. which is one distance "Z" point return per laser emitted. For Flash LIDAR (and and I really don't understand the process) A laser is used for an initial flash and then some type of RGB color distortion for each pixel is generated as a result of the laser flash in a single image and it can be used to find the distance by some type of algorithm. Again, this stuff was designed by people way smarter than me. but I do believe it would work on a big redwood trunk..maybe. I think the Swiss Ranger 4000 is a little different than the camera you use. It snaps a 320 x 120 pixel image..this would be over 38,000 pixels with distance, angle and inclination for each pixel embedded into the image files. At 100k and up, the camera is a completely unaffordable dream however.

Have a happy New Year,

Michael
Last edited by M.W.Taylor on Sat Dec 31, 2011 4:42 pm, edited 2 times in total.

User avatar
M.W.Taylor
Site Admin
Posts: 530
Joined: Sun May 29, 2011 11:45 am

Re: 3D spacial modeling of a giant redwood trunk

Post by M.W.Taylor » Sat Dec 31, 2011 4:32 pm

Matt,

1 Inch accuracy for the entire surface of the trunk would be very useful !

Michael

fooman wrote:Hi Michael,

The laser scanner we use is not a Lidar scanner - it does not use a reference position to generate the point cloud. It provides self positioning data via reflective spots that are randomly positioned on the scanned object. Once these points are registered the laser beam is scanned across the object, much like a spray can, "painting" the scanning beam across the object. This generates reasonably high resolution data (sub-millimetric) very rapidly - maybe a million points in an hour or so. But not really suitable for a large tree.

Another scanner I came across used self positioning in the magnetic field as some sort of golbal reference, but it could not scan metallic objects, so wasn't much use for us.

I've just remembered that our inspection division used to scan the inside of delayed coker drums (large refinery vessels - maybe 20-30 m tall, 4-5 m in diameter) using a laser process - a rotating prism was used to scan a beam around the inside circumference of the vessel, while being dropped down the axis, to pick up cracking/bulging in the vessel wall.

I do think structured light scanning is a real possibility for large trees - all that is needed is a reference dimension (e.g. taped girth) or length between two features on the trunk, a method to project a light pattern (a large projector at night?), and two cameras recording the image at a known baseline plus the software. The guys at the company I referred to in my earlier post were just using normal cameras mounted on tripods a known distance apart. The resolution of a scanned tree would be lower than the smaller objects normally scanned (e.g. people) but I guess +/- 1 inch would be ok for volume, rather than +/- 1 mm?

Cheers,
matt

fooman
Posts: 100
Joined: Sun Dec 25, 2011 2:18 pm

Re: 3D spacial modeling of a giant redwood trunk

Post by fooman » Sat Dec 31, 2011 5:02 pm

Hi Michael,

Just doing a google search for "structured scanning trees" or "3d scanning trees" gives a number of interesting hits, e.g.http://www.coste53.net/downloads/Delft/ ... _Ursem.pdf

Cheers,
matt

User avatar
M.W.Taylor
Site Admin
Posts: 530
Joined: Sun May 29, 2011 11:45 am

Re: 3D spacial modeling of a giant redwood trunk

Post by M.W.Taylor » Sat Dec 31, 2011 8:46 pm

fooman wrote:Hi Michael,

Just doing a google search for "structured scanning trees" or "3d scanning trees" gives a number of interesting hits, e.g.http://www.coste53.net/downloads/Delft/ ... _Ursem.pdf

Cheers,
matt

Yep, those guys are doing just that. Modeling trees with 3D scanners. With small trees in open areas it works like a dream.

The biggest challenges for using this type of equipment for the big redwood example that I am modeling is the extreme forest clutter and few open view of the trunk anywhere. I am using the point red dot laser to get between these cluttered areas one scan one piece of the trunk's surface at a time. Takes time, but it is accurate and reliable and I am using inexpesive equipment and free software. Also, this tree is nearly 300 feet up. How would you model the mid to upper trunk ?

Also, my MS Excel spreadsheet volume solver and 3D grapher would probably have a seizure if presented with such a large data set as that which would be generated by the 3D scanners. I would need some animation software and a bigger-faster computer too...which again cost more $$. I am trying to measure this tree's volume accurately, but without breaking the bank in other words. Perhaps a company is willing to lease me or even better loan me a 3D scanner for a few days and I will test it for them on a big tree trunk.

Michael Taylor

fooman
Posts: 100
Joined: Sun Dec 25, 2011 2:18 pm

Re: 3D spacial modeling of a giant redwood trunk

Post by fooman » Sat Dec 31, 2011 11:16 pm

Hi Michael,

I can appreciate the issues in data acquisition. Unless your view points are rather small, then it may be possible to capture visible surfaces and interpolate between them. Either that or run a scanner up and down the trunk or along lines next to the trunk.

In terms of software, have you used MeshLab (http://meshlab.sourceforge.net/) - it is free software used for editing point cloud surfaces - I've had a play with it, but not much else. Don't know if there are any useful querying tools for the data.

Out of curiosity, has anyone had a play with submitting tree photos to Photosynth (http://www.photosynth.net/ and querying the resultant point cloud against known data to check accuracy? Lots of tree photo's already coverted to 3d point clouds there (e.g. http://photosynth.net/view.aspx?cid=3d7 ... 2367f10783).

Cheers,
matt

User avatar
edfrank
Posts: 4217
Joined: Sun Mar 07, 2010 5:46 pm

Re: 3D spacial modeling of a giant redwood trunk

Post by edfrank » Sat Dec 31, 2011 11:26 pm

Perhaps you could run the scanner up and down a fixed line suspended from tree and attached at its base.
"I love science and it pains me to think that so many are terrified of the subject or feel that choosing science means you cannot also choose compassion, or the arts, or be awe by nature. Science is not meant to cure us of mystery, but to reinvent and revigorate it." by Robert M. Sapolsky

User avatar
M.W.Taylor
Site Admin
Posts: 530
Joined: Sun May 29, 2011 11:45 am

Re: 3D spacial modeling of a giant redwood trunk

Post by M.W.Taylor » Sun Jan 01, 2012 12:31 pm

Hi Matt,


Meshlab looks like a powerful and open source 3D graphing program. I will give that a try for adding a texture and color to the trunk after it is mapped. I have been looking for something like that to display the mapped trunks...which is open code. Thanks for the link !

I think most of your ideas would work well for mapping tree trunks with open views, but I don't think they would work for Drury Tree or most other big redwoods in the temperate rain forest. It's litterally a jungle in there. Drury Tree has extreme clutter around its trunk from base to top and it also can't be climbed due to being in a restricted park. To get a permit to climb in Prairie Creek State Park is next to impossible so that options is out for me.

My main reason for posting the Drury Tree mapping project data was to demonstrate a new way to measure trunk volumes of trees with a irregular trunks that have no predicatble radii and also do so in-expensively and with tools that normal tree people have access to like MS Excel and Trupulse200 laser ......and Meshlab.

Instead of seeing Drury's trunk as having radii..I am seeing the trunk of Drury as form or structure in space to be mapped. and I will use MS Excel with VBA code to find the volume of this trunk by use of incremental "homothetic" slices. I will post the code when finished so others can use it and see the math theory behind how I did the volume calculations. It appears Meshlab will do most of this already.

As you can see by the 3D graphs I provided, I already have been successfully mapping Drury's trunk through the clutter with nearly 3,000 points collected in a 4 hour mapping run... I can tell I need to fill in some gaps by looking at the tilting graph of the bole. As soon as I get a few more points on the trunk for top closure I will estimate the volume using MS Excel and MapSmart (software from Laser Technology Inc. for volume calculation under a mapped surface). If all goes well I will display these incremental "homothetic" volume slices with using Meshlab.

Steve Sillett and Bob Van Pelt have tried ground based LIDAR to map redwood trunks by suspending a rotating spherical LIDAR array from ropes and they hoisted it up into the canopy. The mapping did not work very well due to clutter. Too many missing gaps to fill in. They gave up on the project as far as I know but perhaps they figured out a way around this problem.

I will post another update on Drury when get a volume estimate and at least 7,000 more points on the lower trunk.

Thank you again for the information,

Michael

Post Reply

Return to “Measurement and Dendromorphometry”