h a l f b a k e r yCeci n'est pas une idée.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Tree Recognition
individual tree recognition as opposed to tree type recognition | |
The linked image is (not i now realize) of the Nigerian
girls being held
hostage (but) there is (still) a tree in the background.
It was making me think that there should be tree
recognition systems, the way there is face recognition.
So this could be a job for all of those Amazon drones,
as
they are delivering your groceries they could also be
getting data about the trees they pass and loading it into
a
tree recognition database so that among other things,
photos like this one could be identified by location.
It would also be an interesting problem, recognizing a
tree
from multiple angles.
Treezilla
http://www.treezilla.org/ [JesusHChrist, May 13 2014]
image with tree
http://www.bbc.com/...rld-africa-27383865 oops, it turned out not to be an image of the bad guys but an image of the good guys [JesusHChrist, May 13 2014]
How to Recognise Different Types of Trees From Quite a Long Way Away
http://en.wikipedia...ite_a_Long_Way_Away Prior Art [8th of 7, May 14 2014]
The video therof
http://www.youtube....watch?v=ug8nHaelWtc [not_morrison_rm, May 14 2014]
Tree Data Structure
http://en.wikipedia...28data_structure%29 This could be used for the AI [the porpoise, May 14 2014]
[link]
|
|
Pre-link, please don't tell me we have AI that can do this. I'm all for hunting down those bastards, but can we leave AI out of it? |
|
|
The image turned out to be an image of the good
guys. I'm still for tree recognition though. |
|
|
Tree recognition: it's byte is worse than its bark. |
|
|
You could start a company that could do this, with branch offices. |
|
|
+ I like this. Once whilst boating in a Louisiana swamp
/ lake basin, we got a bit lost except for the
*recognition* of certain swamp trees that led us back
to our houseboat! |
|
|
Very useful for recognising, for example, the giant redwood, the larch,
the fir, the mighty scots pine ... |
|
|
Agreed, the larch. The larch. etc See link. |
|
|
If we're going to be able to recognize all trees, we might as well name them all. |
|
|
They have names. For example: the larch. |
|
|
//please don't tell me we have AI that can do this// |
|
|
Not yet, perhaps, but we have the perfect tool to build such an AI [link]. |
|
|
That certainly gets to the root of the problem. |
|
|
You keep branching off of the actual idea. These
names would be like: the oak with the 2 low
branches, the pine with the broken top, the birch
full of worms, etc. (Something to identify the
individual tree.) |
|
|
My main question would be how specific is the form
of an individual tree. |
|
|
Annual growth, changes in leaf pattern, animal
action, even shed bark (does a paper birch grow back
the same pattern every time), etc. can all make some
significant differences in the tree's appearance. |
|
|
I''m not saying it's impossible, but I would question
how up to date your database could actually be. |
|
|
good for puns, but I don't think you can recognize a
tree. |
|
|
You can recognize it from it's GPS location, as trees
don't move too fast. |
|
|
But, trees change with seasons & years & pruning &
storms, & other trees shading it, & animals of all
types copulating among it's branches, & everything
else. |
|
|
Our faces can be recognized because the distance
between our eyes & nose & mouth don't change
when our eyebrows make odd shapes like they do
when they read my anno out loud. |
|
|
Actually, I don't think it would be that hard to do.
If the tree is reasonably mature, the pattern of
the major branches is unlikely to change
drastically. |
|
|
You could probably do this along similar lines to
fingerprinting. Tree 151171773, for instance has
its first major branch at 3m, a fork at 4.5m, then
the next five branches are at heights P,Q,R,S and
T with elevation angles U,V,W,X and Y and
headings of A,B,C,D and E. That's 15 pieces of
data from five branches. If each piece of data has
10 possible values (to be error-tolerant), you've
got 10^15 distinguishable fingerprints. |
|
|
Even if the tree loses one or two branches, the
pattern should still be recognisable. |
|
|
Eminently bakeable, I'd say. Needs a very fancy
iPhone app to extract data from multiple photos
of the same tree. |
|
|
This could be a GPS replacement, too, in reasonably wooded areas. |
|
|
Trees get a brand new face each year or anytime they are pruned. And the Aspens change in the slightest breeze. |
|
|
This is a GREAT idea. Not only trees, but any background. You
would need a massive database of images of everywhere -- like a
google maps on steroids, but it could in fact be done with drones. In
fact, in a case like the kidnapped girls you could go get those
images after the fact; drone around the Nigerian forests looking for
a similar tree pattern. Narrow the list to the most likely matches.
Check them all out. This has to beat manual searching. Hmmm ...
to my mind this is so close to something we could do today that I
wonder if it isn't secretly being done already. This sounds like an
NSA kind of capability to me. (P.S. I hate those <dirty word here>
NSA snoopers, but I can see where that capability can be used for
good as well as evil.) |
|
|
Anyway, isn't there already pretty good face recognition software for
human faces? How much different is tree recognition? (Probably a
lot, I don't know; hence the half-bakededness.) |
|
|
//Trees get a brand new face each year or anytime
they are pruned. And the Aspens change in the
slightest breeze.// |
|
|
As [MaxwellBuchanan] pointed out in an earlier anno,
that is simply not the case. The disposition of major
limbs would be enough to identify a mature tree,
and will not change over many years. |
|
|
If you can't identify it, it's a missed tree. |
|
|
If find it where people enter the forest, it's an ent
tree. |
|
|
If you are looking for a 100 trees, & only find 3 left,
that's a pal tree. |
|
|
If you identify them having dignity & grace & male
gender, they are gent tree or gallant tree. |
|
|
If you don't like puns, well, I didn't start this mess. |
|
|
You just couldn't leaf well enough alone, could you. At least you left AI out of it - thank you. |
|
|
I've been reading the title as a bit of skiing nomenclature. . . |
|
|
Treecognition: enough time to admire God's creation in the form of a spruce tree; enough time to do an inventory of body parts for later comparison; not enough time to successfully swerve out of the way. |
|
| |