h a l f b a k e r y"Bun is such a sad word, is it not?" -- Watt, "Waiting for Godot"
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,
|
|
|
Gaborflage
now you see it, now you're trying to see what it is | |
During WW1 several ships were camouflaged using a technique called Dazzle to confuse and deceive enemy submarines as to the orientation of their target vessels. In a mark of the ever inventive British, this was known as Razzle Dazzle.
This worked very well apparently, so well in fact that the process
has been recently deployed again. I am proposing that this technique is greatly enhanced by the use of Gabor Patterns as a type of camouflage.
For anyone not knowing what these are, it’s best to look up the links then imagine trying to focus on something, within possibly only a few seconds, that is in motion and displaying a Gabor Pattern as a form of camouflage.
see links to fully understand
Razzle Dazzle
https://www.smithso...m-dazzle-180958657/
[xenzag, Jul 31 2021]
Gabor Patterns
https://www.syfy.co...is-circle-illusion?
[xenzag, Jul 31 2021]
Extensions to get rid of those pesky straight edges
https://www.google....36&bih=750&dpr=1.25
Disguise the shape of the ship as well as the pattern. [AusCan531, Aug 03 2021]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
I can't see why static zebra stripes would be better than Razzle. And if you can make the appearance change there are more effective ways to actively camouflage. |
|
| |
They're another step up re creation of visual
confusion. Have you looked at them? Circles
become squares. Parallel lines are not parallel etc. |
|
| |
The reason dazzle camouflage became obsolete was the
increasing use radar, to which it offers no defence.
I can't see how gaborflage would differ in that respect. |
|
| |
It's not meant to defeat radar which also cannot be
used in many scenarios. Visual camouflage is still
widely deployed and Gabor patterns extend the
possibilities of confusion and deception regarding
the use of scopes and optical focusing devices. |
|
| |
I wonder whether there's a version that can confuse automated image recognition via radar and thermal. |
|
| |
Because it can be hard to judge distance at sea, I came up
with a different form of camouflage: paint your enormous
battleship to look like one guy in a little dinghy, fishing.
Also, this idea is awesome. Those Gabor illusions break my
brain. |
|
| |
Of course! (The stripes looked like zeds). |
|
| |
[Voice} As a rule, automated image recognition isn't going
to be fooled by a lot of things that trick our eye. Among
other reasons, because we're both better and worse at
taking in the entire image at once. (That is, our eyes
can't capture the whole image at once like a camera can,
so it makes a lot more assumptions in processing it). |
|
| |
As far as visual camouflage, I like this idea, but my
concern is that it doesn't handle edges well. If you look
at the header image of the syfy article, the illusion
disappears once you no longer have full circles. |
|
| |
To address [MechE]'s concern about straight edges, it should
be a trifle to change that with extensions [link] which
change the regular shape into anything you like. Perhaps
spring loaded nylon sheets which snap taut. |
|
| |
Alternatively, a battleship-sized inflatable bust of Gabe
Newell, giving the impression of a giant computer
programmer wading through chest-deep ocean. That'll
cause confusion. |
|
| |
I first read this as "Garbaflage" e.g. being camouflaged as
city garbage. Now there's an idea. |
|
| |
[MechE] automated vision systems often also break an image apart for the purposes of parallel processing. And such a system could be vulnerable to difficulties humans don't face due to humans' vastly superior tertiary processing of the image. A contemporary image recognition system is severely limited by the images it was trained on. It has no imagination, no awareness of the item it's looking at, and very little cross-association.
For example if you make a modern war-ship of a wholly new shape but with individual parts about where they're supposed to be by function and shaped about how they should be by function a human gunnery officer could suss out what it is and even some of its capabilities. The officer would certainly recognize it as a warship, and probably identify its primary and secondary functions. An image recognition system would throw up its digital hands and release the magic smoke. Or take a thermal image of a known warship, but overlayed with insulation on certain parts, heat emitters on others, and heat spreaders on still others. A human could figure out what had happened. An automated image recognition system might gleefully report, "There is a piece of a dock floating in the water. 95% confidence" |
|
| |