Camouflage Detection via Multispectral Imaging

[Music] there are very few subjects that capture the imagination and the billions of dollars in funding that the art of camouflage has encapsulated over the centuries military units all over the world have spent the past century developing the absolute best camouflage patterns fabrics colors textures materials and doctrine and not without good reason an enemy that cannot be seen cannot be targeted well in most cases anyway so today we're going to dip our toes into the ocean of camouflage once again and discuss a very specific field of technology camouflage detection and to begin this journey we have to make a pit stop in a farmer's field in africa over the centuries africa as a continent has suffered from widespread famine and crop failure millions of people have died from food insecurity that is directly related to various environmental effects which in turn is a major contributing factor for conflict so one very interesting bit of technology that can help us with this is what we call multi-spectral or hyperspectral imagery so what does this mean well again this is a gargantuan field of technology so wide-ranging in its applications that it's almost easier to show what it can do r oes this mean well again this is a gargantuan field of technology so wide-ranging in its applications that it's almost easier to show what it can do rather than define it with words but in this case when it comes to our specific topic today multi-spectral imagery is imagery photographs or video but usually still images that are taken using optical sensors like a camera but instead of capturing visible light like your ordinary handheld camera multi-spectral imagery captures different wavelengths of light that cannot be seen with the human eye so how does this help with famine in africa or really anywhere else in the world well let's find out way back in 1972 nasa launched the first landsat originally called erts or the earth resources technology satellite this satellite's mission was to study the surface of the earth using various sensors one of these missions was to use multi-spectral imaging sensors to study agricultural efforts and crop yields around the surface of the earth so let's dive in and see how this technology works to simplify things let's not look at a whole field of crops and not even an individual plant but a single leaf plants are fascinating organisms and one of th lify things let's not look at a whole field of crops and not even an individual plant but a single leaf plants are fascinating organisms and one of the things that most people know about plants is that the substance that makes the leaves green is called chlorophyll but in order to see that green color or really any colors at all we have to have light thinking back to third grade science class we remember that when it comes to colors the human eye is simply seeing the reflected light from an object that's what gives objects their color so when we see this leaf or when a human eye looks at it sunlight illuminates the leaf bounces off of it reflecting the color green from the chlorophyll to the human eye where it gets processed by the brain and interpreted as a color now if the plant is sick or dying the leaves will change color everyone that has tried to grow a bean plant on the windowsill of an elementary school classroom knows at least once that sinking feeling of seeing a brown shriveled stalk of what once was a bright green bean plant now this is a very oversimplified explanation of how humans visually see all objects so let's complicate it a bit instead of the human eyeball here nt now this is a very oversimplified explanation of how humans visually see all objects so let's complicate it a bit instead of the human eyeball here let's replace that with a multi-spectral imaging sensor that can see trillions of different shades of color that the human eye can't see and let's throw that image sensor that camera onto a satellite orbiting earth and let's hook it up to a wicked smart computer or more accurately a data processing center and let's add in something else have you ever gone to a home improvement store and stood in the paint section looking forlornly at a handful of paint swatches trying to discern the difference between satin white and eggshell white we have all done that at some point i cannot tell you how many times i've tried to visually match paint colors only to fail and have to look up the five or six digit number code on the back of the paint sample card a bit of interesting technology that has come out recently is that magic image sensor that can match the color of any object that you bring into the store quite a handy bit of technology that has become mainstream in recent years so an image sensor is way better than the human eye when it comes he store quite a handy bit of technology that has become mainstream in recent years so an image sensor is way better than the human eye when it comes to detecting colors or more specifically the differences between two colors we've all seen optical illusions like this one which make these two squares look like they're two different colors when in actuality they're the exact same color most humans are tricked by this simple optical illusion but computers and imaging sensors are not so we put one of these bad boys into orbit and we start collecting data when it comes to the human eye we are really bad at determining the different shades of green that plant leaves are made of but a multi-spectral imaging sensor is not for instance the human eye can look at two different leaves and say that they are the same color but a fancy satellite in orbit with just a really awesome sensor can tell that this leaf is just a few thousandths of a shade darker than this leaf and he can realize that there is something going on there with that particular leaf so some of you might have started putting the pieces of this puzzle together and realized that some very interesting folks might be interested in eaf so some of you might have started putting the pieces of this puzzle together and realized that some very interesting folks might be interested in this technology a great open source use of this kind of technology is the detection of marijuana growing operations the dea and other agencies have invested quite a lot of money in this technology because it turns out that while the human eye just sees green leaves on plants every single plant on earth is actually a slightly different color again it's far too small of a difference to see with the human eye in most cases but to a sensor and a computer which has access to a database of trillions of leaf colors all calibrated for different times of year and the life cycles of each plant and calibrated to the exact light reflectivity of a particular time of day it's really easy to see the difference between say deciduous tree leaves and marijuana leaves so you see how this starts coming together one can easily then make the assumption that this technology is very useful for analyzing the light reflectivity of other substances not just leaves and you would be right if you have a database that contains the magic color classification of say ht reflectivity of other substances not just leaves and you would be right if you have a database that contains the magic color classification of say various white powders or explosives you can see how the ministries of defense around the world might be interested in this technology and if your military is cripplingly dependent on the utilization of unmanned aerial systems in combat well you'd be highly interested in this technology indeed enter the field of camouflage detection as one has hopefully guessed by now multispectral imagery is a technology that is very well suited to detecting when a leaf is not a leaf but a piece of camouflaged fabric that is dyed painted or otherwise colored in such a way as to blend into its environment and now all of the sudden it makes sense as to why camouflage patterns colors dyes paints and fabrics are so difficult to make not only does it have to be resistant to infrared sensors or night vision devices both of which we could do entire videos on but camouflage has to mimic the natural environment that it's in almost to a molecular level you have to have green colors and materials that almost perfectly reflect light in the same way that chlorophy hat it's in almost to a molecular level you have to have green colors and materials that almost perfectly reflect light in the same way that chlorophyll does or any other number of materials found in nature so when it comes to camouflage there really is no hiding from multi-spectral sensors when the army announced in 2018 that they would be replacing all of their camouflage netting with a new program the ultra lightweight camouflage net system or ulcans it was done with multi-spectral camouflage in mind but like all camouflage attempts it was more or less doomed to fail from the start with nation state espionage at an all-time high and with private companies leaking defense technologies left and right it would not be inaccurate to assess that russian and chinese agents obtained samples of this new ulcans fabric almost immediately probably before even the dod got theirs so to continue our analogy all russia and china have to do is put this fabric in their million dollar home depot paint sampler and boom they've got the exact reflectivity of the us army's brand new multi-million dollar ultra secret camouflage netting and they can even tailor it to different lighting patterns differen ty of the us army's brand new multi-million dollar ultra secret camouflage netting and they can even tailor it to different lighting patterns different sun angles and can even write software packages that can be used on aircraft sensor pods and adjusted for ambient lighting conditions so that if a sensor sees that particular reflected color it will alert the sensor operator or even the pilot himself pretty cool technology that while maybe not widely fielded right now is definitely a possibility and will be fielded much more in the future so why even bother with camouflage anyway if multi-spectral imagery is so great that nothing can hide from it why even bother with developing and spending a lot of money on camouflage as with any technology there are a lot of downsides so let's go down the list of problems which can be used to evade surveillance if your adversary is known to be equipped with this technology and that's really the first point the adversary has to have it your adversary actually has to physically have the technology let's face it not many entities around the world have the funding and capabilities to actually develop this technology sure people like us can easily expl s face it not many entities around the world have the funding and capabilities to actually develop this technology sure people like us can easily explain it in excruciatingly oversimplified detail but when it comes to actually writing the code and funding and building the sensors things get exponentially more difficult so chances are your local police department or even local national guard units don't have this technology next is the adequate knowledge of implementation even if an entity like a nation state or theater level command have access to this technology they have to use it the right way think about it you can put a 10 000 dslr camera into the hands of a complete novice and they might be able to figure out how to snap a couple of pictures but you will nowhere near approach the capabilities of the actual sensor so you've got to have the right sensor operator behind the lens getting the absolute most out of the technology that's possible also along these lines an adversary that possesses this technology must actually have the freedom to operate and use it for instance you can get a professional photographer and give them a great camera but only tell them to take pictures of reedom to operate and use it for instance you can get a professional photographer and give them a great camera but only tell them to take pictures of a garbage can and nothing else sure you'll probably get some really good photos of a garbage can but i honestly don't know of any garbage can that is a threat to national security so if your adversary's command structure is incompetent in that they do not know how to best utilize the assets that are given to them this is a huge advantage to anyone who is looking to evade surveillance next is appropriate usage like we mentioned earlier this technology has extremely wide ranging uses from agriculture and famine prevention to counter narco missions but very few assets that are in current usage around the world are actually dedicated specifically to camouflage detection so in other words if the dea is hunting for marijuana grow operations in your area chances are they aren't going to find your observation post unless your camouflage reflects light in the exact same way that marijuana does and even so a lot of camouflage detection efforts are still done by humans the satellite might take the image but in almost all cases it's up to the hum d even so a lot of camouflage detection efforts are still done by humans the satellite might take the image but in almost all cases it's up to the human analyst to look at it so for example if you are setting up an observation post in a remote area you might have a satellite or drone pass overhead most satellites and drones alike take images with a wide variety of sensors all at once but most analysts just stick with the normal electro optical sensor i.e the normal daylight camera or a thermal option to see heat even if the aerial platform is capable of taking multi-spectral images it doesn't matter if the analyst isn't looking for camouflage at the right moment as a side note most sensor operators around the world have gotten quite lazy because they've been following targets in the desert that don't exactly have first-hand knowledge of thermal imaging in the way that many americans do in other words it's really not that hard to use an infrared thermal sensor to track an insurgent who thinks he's hidden under a bush or something so the analyst gets sloppy and only requests platforms that can only do fleer or thermal imagery at best and they ignore the other capabilities like multi- e analyst gets sloppy and only requests platforms that can only do fleer or thermal imagery at best and they ignore the other capabilities like multi-spectral imaging because they think they don't need it next is sensor availability as scary as multi-spectral imaging technology is when it comes to camouflage detection we have to realize the size of these sensors they are massive right now most multispectral sensors capable of camouflage detection operations are really only on satellites or large fixed-wing platforms we expect this to change in the next few years with camouflage detection sensors being used at lower levels of command and not simply being theater level assets but for now there are an awful lot of hoops that have to be jumped through for a military unit to be able to detect camouflage in their area next up is sensor resolution sort of related to the last point sensor resolution is a huge issue for an adversary that might have this technology everyone likes to think that government satellites operate like the movie enemy of the state and ignoring the fact that near nadir imagery or imagery that is looking almost straight vertically down is pretty much useless in the in of the state and ignoring the fact that near nadir imagery or imagery that is looking almost straight vertically down is pretty much useless in the intel world we most always laugh at the scenes that show these government satellites having incredible resolution it means the nsa can read the time off between wristwatch no they can't even unmanned aerial platforms which can obviously get closer to the target than say a satellite cannot see anyone's wristwatch right now there are only a couple of satellites that are even capable of taking full motion video anyway and it doesn't take a russian spy to figure out what the resolution is for most u.s military imaging satellites spoiler alert it's atrocious and the russian and chinese satellites are a lot worse that's why the army's new ulcans camo netting program makes a lot more sense if the resolution for a particular satellite sensor is three or four meters per pixel well that's enough to hide a small vehicle inside just one pixel now granted most satellites are a lot better than this being far less than a meter in resolution on a good day right but when it comes to imaging sensors with a lot better resolution it can be detected by mult g far less than a meter in resolution on a good day right but when it comes to imaging sensors with a lot better resolution it can be detected by multi-spectral imaging sensors but that's a lot harder to see than the previous camo nets which really only worked against the unaided human eye so that's why the system kind of makes sense and when it comes to an individual human being or small observation post or foxhole or something trying to conceal that from a drone or even better a satellite that's really easy to do as we have pointed out in the right circumstances it is impossible to fool the sensor but what about the analyst as i mentioned camouflage detection is solely dependent on a series of circumstances which all have to line up perfectly for a human analyst to even realize that deception or camouflage is taking place many marksmen and snipers and militaries around the world have discovered that the absolute best camouflage on the planet is the planet itself natural foliage will always be better than something man-made at least for the foreseeable future and even though fresh foliage will wilt and have to be replaced over time a pile of wilted leaves is still much more natura for the foreseeable future and even though fresh foliage will wilt and have to be replaced over time a pile of wilted leaves is still much more natural looking than fresh-looking fake leaves to an image sensor anyway so even if camouflage efforts are automated within a computer which is rare right now but even so natural foliage has a higher chance of altering the data that the sensor is seeing and placing the burden of camouflage detection on the analyst now the analyst has to make a judgment call is that dead foliage from something natural like a parasite infestation or a drought or is this a deliberate attempt to camouflage something you multiply that up over a 12 hour shift and hundreds of miles scanned and you come up with nothing right remember sensors cannot make judgment calls nor can they make mistakes because they cannot make decisions but humans can so by creating that little bit of doubt within the data that that sensor is collecting you are forcing the decision onto the human analyst who is far more likely to make the mistake so as one can see the stars have to align perfectly for camouflage detection efforts to actually work well enough to be considered a significant take so as one can see the stars have to align perfectly for camouflage detection efforts to actually work well enough to be considered a significant threat a nation-state has to have the interest in the first place and fight through the bureaucracy and politics of higher military command structures in order to allocate the right sensor platform to the exact right place at the exact right time and be led by the right commander with a sensor operator that has the right amount of experience using the right sensor mode and who is on top of his game at the exact moment the sensor sweeps over you oh and the sensor operator has to even notice an anomaly to point out to the analyst who then has to have the interest skill level time and make the judgment call to look into it further and even then the analyst has to see through the deception to realize they are looking at a deliberate camouflage attempt and if any of these circumstances are out of alignment you will not be detected by your adversary a lot of these intelligence collection platforms and capabilities and sensors are pretty scary stuff especially considering the world we live in today but one must remember the tragically hilari orms and capabilities and sensors are pretty scary stuff especially considering the world we live in today but one must remember the tragically hilarious irony that goes into a lot of these intelligence collection platforms that a billion dollar sensor more often than not can be defeated by something simple a tree limb cut the right way a cave in the ground or even something as simple and uncontrollable as the weather it doesn't take a billion dollars worth of technology to combat a billion dollars worth of technology all it takes is a cloud in the right spot to negate billions if not trillions of dollars worth of effort and decades worth of time so hopefully this was a good primer on multi-spectral camouflage detection providing various little bits of knowledge that you can use to counter these systems if you need to as much as we'd like to write up a concrete list of things you can do to evade this or other types of surveillance the stark reality is that it is far better to learn about the technology so that one can apply to their specific situation and environment the earth has a pretty diverse landscape so by understanding the overall doctrine and larger themes one can develop pecific situation and environment the earth has a pretty diverse landscape so by understanding the overall doctrine and larger themes one can develop their own practices that work in their specific areas which is far better than having a pocket full of raw tactics but no understanding of how to implement them in each specific situation so thanks for watching everyone and we will see you next time and always remember fight in the shade or maybe in the bushes [Music] you