you know the military is well into camoflage, such as DPM.

https://en.wikipedia.org/wiki/Disruptive_Pattern_Material

just thinking,
this is made from material off a roll,
i.e.. the pattern repeats .

AI is very good at finding known patterns ,
seems like youd stand out to an AI drone / missile more than if you did not have camo!

even with random parts of the pattern on different parts of the uniform, the AI could be programed and recognise sub bits of the pattern,

Id imagine , even mud etc woukd only attenuate the pattern , it woukd still be there.

any thoughts ?

 

I saw a report that the Russians were trying to use "penguin" camouflage, but it apparently made them more visible not less.
Here's one
https://www.thesun.co.uk/news/38060507/russian-penguin-soldier-blown-up-camo/

 

Don't the miltary drones use thermal imaging?

 

Don't the miltary drones use thermal imaging?

thermal is certainly used , but it can be countered with " tin foil" based capes ,
optical has come back into prominence ,

 

It is best to "fade" into the background preferably under the overlying foliage and in the shadows. "Camouflage" only helps somewhat. The real trick is to be motionless. Movement against the background is the real giveaway (along with thermal). I have spotted many hunters in tree climbers wearing the "best" camouflage who stood out like a sore thumb and any small movement is an instantaneous giveaway. Even highly trained snipers often fail field trials due to movement even when their Ghillie Suit perfectly matches the surrounding foliage.

 

Modern detection is often multi-spectral looking at visible, near-IR, and possibly mid- and deep-IR. It can also involve radar using UWB (ultrawide band) imagining. The image processing is also very sophisticated looking for anomalies in many aspects, including motion and non-motion (in a field with a breeze, things that otherwise look like logs shouldn't move, but when something that otherwise looks like a bush isn't moving....). Exactly what technologies can be brought to bear in a given scenario vary a lot, but I would not want to be hiding from a modern military that is motivated to find me. Having said that, it is still a safe bet that those modern military units would sure love to have the capabilities that are routinely depicted in the movies.

 

 

It is best to "fade" into the background preferably under the overlying foliage and in the shadows. "Camouflage" only helps somewhat. The real trick is to be motionless. Movement against the background is the real giveaway (along with thermal). I have spotted many hunters in tree climbers wearing the "best" camouflage who stood out like a sore thumb and any small movement is an instantaneous giveaway. Even highly trained snipers often fail field trials due to movement even when their Ghillie Suit perfectly matches the surrounding foliage.

Agreed , for humans looking .
The rules we were taught was
Stay in shade
Watch what's behind you, I've stay off ridges
Move as little as possible if trying to stay hidden
Try not to look human, curl up , arms and legs in etc.

But
AI , is very good at spotting patterns.
It could easily look at a scene , and detect where a pattern of dark light matches that if camo.

 

Btw
We were also shown the video of good looking teenageres chucking a ball around , and asked to count the throws.
None of us spotted the gorrila walking through the scene . For humans , it's easy to be deceived .

 

Don’t we have the technology to produce non-repeating camouflage fabric?

 

Don’t we have the technology to produce non-repeating camouflage fabric?

exactly,

currently we dont , the material comes on big long roles with the pattern repeating every few meteres, dependent on the size of the roller .

The thought is , that things are going to have to change.

 

exactly,

currently we dont , the material comes on big long roles with the pattern repeating every few meteres, dependent on the size of the roller .

The thought is , that things are going to have to change.

This is certainly something that you can bet is constantly being assessed -- just consider the publishability of research papers that claim to show that the pattern repetition is an exploitable vulnerability. My understanding is that, currently, the pixelation (which has evolved over time as a result of advances in detection algorithms) is a net gain and reduces detectability via other measures, such as edge detection and motion detection, by more than any amount that the pattern detection exploits it. That, of course, is always subject to change.