Artem vs. Predator | Hacker News Reader

Artem vs. Predator

ribbonfarm.com

651 pointssharpn67 comments

A cool property of NIR filters is that they are able to produce a relatively recognisable picture of the world with all screens (TFT, projected) filtered out. I've used this in the past for computer vision to do object recognition on top of a projected display without having to programmatically filter out the projection.

It's also pretty easy to make a rudimentary NIR filter by layering a red, green and blue filter on top of each other: the resulting filter will not allow any visible light through (e.g. red, green or blue), only the "rest".

edit: and they make your eyes look really creepy too (not me in the picture btw): https://flic.kr/p/6CYzDZ

Another easy way to make a NIR filter is to get a couple of black frames of a regular photographic film - it's transparent to NIR, but blocks all the visible light.

I assume the distortions at the edge of the 'photos' is due to the X Y scanning rig. Moving the sensor around the edge of a sphere (on a curved X Y (and a little bit of Z) rig, the focal point being the center of the sphere) would sort that right out.

I expect the 3D printer could make a curved rail for the rig.

Or you could point the sensor at the center of the pin hole or lens using a mechanical linkage, then use some software to warp the resulting image.

No it's a property of pinhole diffraction: https://en.wikipedia.org/wiki/Airy_disk

The buildings at the edge of the photos look warped. Like an extreme fisheye lens would do.

Airy disk affects brightness and sharpness, not warping.

You could put a camera lens where the pin hole is to negate this warping, but that defeats the point of a pin hold camera.

Yes, it is proably more petzval field curvature than diffraction limiting: https://en.wikipedia.org/wiki/Petzval_field_curvature

About half of the pictures in the article are taken with the lens assembly. Pinhole is good for outdoors and sunshine, but is useless for MWIR/SWIR and indoors scenes, even with my 600W Metal Halide lights.

The lenses give the radial blur at the edges effect, pinhole gives the darkening at the edges effect.

I am intrigued by your ideas and would like to subscribe to your newsletter... Joking aside interesting projects! Could you send me an email? Perhaps I can convince you to do some consulting work. Haha, the world needs more predator vision! Though ya née gavayete pa Ruski.

Newsletter might be too big of a word, but i do publish various stuff i work on at http://orbides.org/

You can also find the e-mail over there in the bottom-right corner.

Not just heat, there is a lot of interesting spectral data we don't see. Hyperspectral cameras are rare and expensive right now - but I'm really looking forward to the day they are cheap and commonplace

https://www.youtube.com/watch?v=OcAstSnbVX4

http://www.specim.fi/

https://www.ximea.com/

Forgive my ignorance, but if this system would require liquid nitrogen cooling to detect heat from a human, how do commercial (now almost cheap) thermal cameras work? Are they not sensing IR? The can certainly distinguish the different temperatures of different people, without needing to be cooled themselves.

How is this system better than a commercial IR camera -- besides the very obvious and valid reason that this is DIY, and started long before those were cheap?

> "The most amazing part is not that it glows, but that it glows brightly enough to illuminate the stand. It’s not just the “temperature mapped to an image” of a regular heat vision camera, we see the actual long-wave light being emitted and reflected – a soldering iron turned into a lightbulb!"

How is that different from this image, which shows a duck's IR being reflected by water, made by a cheap-o "regular heat vision camera"? http://thermal-imaging-blog.com/wp-content/uploads/2012/05/d...

As I understand it, uncooled commercial IR cameras aren't photodetector-based: their sensors are actually heat-sensitive and rely on heating from IR to produce an image. This limits sensitivity and resolution and increases noise.

Higher-quality IR imaging equipment is photodetector based (as in, they operate on the photoelectric effect); these require very low temperatures to operate and need the sort of active cooling the author refers to.

See also https://en.wikipedia.org/wiki/Thermographic_camera , which goes into detail on the different technologies involved.

You know those black and white photos? I thought Russia looked like that with regular cameras :P.

Just kidding, this has to be one of the coolest projects I've ever seen. To make all this by yourself is just ... awesome.

Here is one of the more apocalyptical ones (i call it "The storm is coming"): http://i.imgur.com/zFD4znl.jpg

Eerily similar to the clouds in the original Doom! Luckily real Russia is only partly apocalyptic looking... Though I'm from Wyoming which boasts our own post-apocalyptic scenes. ;)

  This was my first encounter with ITAR – the “how dare you 
  want interesting stuff?” restrictions. Before then I 
  never realized just how USA, uh…, loves the whole world.

ITAR is a US regulation, and the USA doesn't have jurisdiction on a sale from a Japanese vendor to a Russian customer. You're thinking of MTCR, "an informal and voluntary partnership" between 34 countries. Both Russia and Japan are members: https://en.wikipedia.org/wiki/Missile_Technology_Control_Reg...

"..when I started there were no 3D printers. Eventually I made one, and all the clumsiness of the old rig got replaced with modern, 3D printed, well-fitted parts."

When I read this I thought, "nah"... but some digging later and

http://orbides.org/page.php?id=703&lng=eng

I've seen a lot of hacky DIY 3D printers, but this is the roughest. Love it!

Great read! I'm curious about the last bit, saying you need liquid nitrogen cooling to detect the radiated heat from a human. I'd always thought some snakes have infrared-sensitive spots near their mouths that could sense mammals from a distance, are they just that much more sensitive than a photodiode? Or are they sensing something else?

Snakes are seeing in the closest to visible specter IR IIRC. Also they have evolved a lot of other sensory mechanisms to detect pray - sounds, vibrations, taste

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

They can sense radiation between 5 micrometers and 30 micrometers.

Wow, so instead of a photochemical or photoelectric effect, it's actually literally a heat sensor - as in, radiation from the environment heats the interior of the heat pit (which is cooled by the snake's blood to maintain a baseline) and it is this change in temperature that is detected?

I wonder if the ribbonfarm guy could use a variant of this for his pixel scanner thing?

That is the operating principle of a bolometer.

No idea about snakes, but a sensor will glow itself at these wavelength, since it has room temperature which is basically the same temperature as body temperature. So to stop the sensor from blinding itself, you need to cool it significantly below the temperature you want to see.

Nice way to start the day ;) Can't someone send this guy a box of gently used CCDs and some wideband lenses already or maybe an old copy of labview...

Oh and Artem, if you happen to see this, can you try getting out of the city on a cloudless night and taking a long exposure of the night sky? Thanks for your hard work and inspiring us all!

Artem here, that's a strange idea. Right now the sensor got a fixed gain, and so exposure time is meaningless. It won't be too hard to add, but what sort of image would there be, with the frame time of many hours, line by line as the planet rotates? Sounds curious enough to try. :) I anticipate a distorted starfield - every star as a bright point, but all of them shifted in position.

Just star trails, probably:

https://upload.wikimedia.org/wikipedia/commons/f/f8/Star_tra...

Yes, but distorted because of the way the sensor is also moving.

Ah, interesting. They'll be discontinuous dots along that line, true.

So is the future wavelength project something like radio waves? It would be fascinating to see an image lit by radio waves. Imagine a wifi router being lit up like a lightbulb, casting shadows around a room, or a room lit by FM radio passing through walls and windows, etc.

Walls are transparent to FM (~100MHz) radio, so you won't see anything.

WiFi is closer. High microwave is probably what you want here.

And of course, this already exists. It's called radar. Haven't seen any radar "photos" though. Interesting if that's indeed possible.

The size of the things you can see depends on the wavelength of the light.

FM Radio (100MHz) has 3m wavelength. You'll need a giant camera and will be able to see only very big structures.

Wifi/Microwave/Radar is in the Gigahertz range. 2.5GHz == 12cm wavelength. For radar photos, you'll need a very big camera, and a human would be only a few pixels big.

Some airport scanners are based on millimeter waves (teraherz radiation), and you can find sample photos online. They produce pretty blurry pictures, but clothes become partially transparent.

> And of course, this already exists. It's called radar. Haven't seen any radar "photos" though. Interesting if that's indeed possible.

Oh, right, synthetic aperture radar. Duh. Should have thought of that. :)

Unfortunately the resolution of your image will decrease with an increasing wavelength so you might end up with a pretty fuzzy radio-picture.

> But for our goal it’s important to remember, because infrared light does not pass through glass

I thought that was UV?! If that statement is true, how on earth am I able to busk in the heat of the sun behind a glass window, indoors?

The sun produces visible light as well.

Yes, I think this is a common misconception: infrared light is "heat radiation" in the sense that it's emitted from hot things, but any wavelength of light you absorb (not just IR) will heat you up.

> but any wavelength of light you absorb (not just IR) will heat you up

Thanks - this is the bit I was un-/miseducated about. When we were taught about radiation in high school, I suspect my young mind concluded that since infrared and visible light were mutually exclusive on visibility, the same applied to heating. To this day, I thought infrared → heat and heat → infrared.

I just happens that at everyday temperatures familiar to humans, the primary wavelengths emitted are infrared. Heat something enough, and that curve will shift to predominantly visible wavelengths; think about an electric oven.

IF is NOT heat. as a matter of fact, IF is a low energy radiation. So, very few heat.

It is what bodies at low temp (like a human being) radiates the most. Higher energie bodies also radiates higher energy light, such as visible light and UV.

Infrared film used to be available commercially with sensitivity up to 900nm. It was quite useful for aerial photography work – forestry, surveying, spying. Very little of it is still around. Kodak Ektachrome was what one could get for your high school dark room. Aerochrome for surveying work. It was discontinued in 2009. Ilford however still makes some. Be forwarned, finding someone to develop the stuff is a nightmare, the chemicals are toxic, and the shelf-life brief.

Scientific infrared films, such as special formulations of AeroChrome I, II, II, approached sensitivity up to 1200nm.

In surveillance work, objects which were painted to look like their natural environment using various organic or inorganic paints may show up quite differently in the infrared spectrum.

In forestry work, old growth tree populations could easily be distinguished from new growth tree populations and were one of the primary uses for Nasa's version of the U2 (ER-2) for identifying old-growth redwood populations in northern California. [1]

A lot of work was done in the 1970's and '80's by astronomers and physicists to 'hack' Eastman Kodak scientific film, or plates as they were called. (Once you move past "point and shoot" film, you get into the realm of plates, 4"x5" trays similar to old-timey 1880's cameras.) Things like Kodak I-Z. One technique was to hypersensitize the film by bathing it in Ammonium Hydroxide [2]. Lawrence Livermore had such an appetite for IR-sensitive film with their laser work that they set up their own production process for hypsersentizing Kodak scientific plates. Another was to supersensitize them with acetic solutions getting film sensitivity in the >1500nm range [3]. This seems to be the limit of our knowledge for traditional chemical film processes.

Modern DSLR's have sensitivity up to 1600nm. Nikon worked with NASA for some of their special DSLR's [4].

One of the cooler things I saw was a University of Florida paper in Nature that used IR-OLED's to upconvert IR to visible light through a lens adapter achieving sensitivity from 400nm to 2000nm [5].

Beyond 2000nm you get into the MWIR range and FLIR devices take over.

[1] https://books.google.com/books?id=HZUTCgAAQBAJ&pg=PT129&lpg=...

[2] http://www.osti.gov/scitech/servlets/purl/4442636/

[3] https://books.google.com/books?id=nlftCAAAQBAJ&pg=PA259&lpg=...

[4] http://eol.jsc.nasa.gov/Collections/NearIR/IR_Intro.htm

[5] http://www.nature.com/articles/srep05946

Do you have do develop those films in the freezer? A normal dark room is not really dark at those wavelengths, even if you turn the red lamps off, right?

Former B&W technician: All film is developed in complete darkness; it it sensitive to all wavelengths. The B&W photo paper is not sensitive to red, which is what allows you to have a dim red light on while printing enlargements.

And yes, it takes time to learn to navigate the room and work in pitch blackness. Gave me a lot of respect for the blind. At school, we could develop one roll at a time, allowing for the use of small containers which would let you work within a dark bag for a small period of time. At the photo shop, they had a machine for processing film in batches of 10+, which required the entire room to be dark while loading it.

Modern film processors frequently don't even require the dark box (unless the film canister is in really bad shape) - just stick it in, close the door, push a button. Remarkably slick.

Home processing of some regular B&W infrared film was tricky because the small tanks would sometimes be giving off IR, and would fog the film.

It is possible to see in near infrared - just use a IR filter (used for photography), and stick it to your eye. Everything should look dim and reddish, but what is characteristic (and different from a red filter) is that leaves are very bright. Also, many black t-shirts are not-dark.

Also, it's fun to make such photos:

http://migdal.zenfolio.com/ir-old

Just be really, really careful about sunlight! When you're using a filter like this your iris is wide open, and exposure to direct sunlight or reflected direct sunlight can damage your eyes.

Yes, probably the most comprehensive site about making and using IR goggles is here:

http://amasci.com/amateur/irgoggl.html

He has some very interesting photos taken with them. He also cautions about the iris being opened wide when wearing them, although he claims that he's suffered no ill effects after using them for years (avoiding looking at the sun, of course).

Funny thing is, i actually tried that as a kid, with the lids from some big remote controls. The world was deep red, and had an alien feel to it, but i had no idea what i was actually looking at back then. The foliage might have actually been bright, can't really remember by now.

> Heat vision. [...] MWIR is neither heat, nor is it light. It’s both

No, it isn't. The "infrared light is heat" statement is casually-useful but technically-false, which means it backfires and harms people when they start learning more about physics.

IR radiation is not the intrinsic heat-content of the matter emitting it. Infrared is simply a particular set of wavelengths which happens to be prominent for temperatures which are "hot" for us just because we evolved on this small rocky planet which has a certain baseline temperature.

Go ask some icy crystalline aliens or ones powered by low-grade nuclear reactions what they think about "heat radiation" and you won't get the same answer.

Have you tried zone plates instead of lenses? Feature sizes should be larger for IR than for visible light, so it's should be possible to fit enough zones to get some focusing.

To expand a bit on construction: it's possible to take a long tube, put a fresnel lens at front, pyroelectric sensor with rotating shutter at the other end. Scan by rotating it with reductors/steppers as a whole. It's also possible to cool tube with contents down to -70 with dry ice to cut on background heat radiation. Maybe it'll lessen noise in pyroelectric sensor, maybe not (and it's quite easy to buy dry ice in Moscow).

Incredible dedication and a fantastic project. I wonder if a setup made from a FLiR Dev Kit, a Raspberry Pi and the Pi camera would produce similar result?

This guy is spectacular. Check out his other work: http://orbides.org/ Here is a nice concept of an interstellar ship http://orbides.org/concepts.php?lng=eng

Cool stuff!

I think that regarding the near infrared light a regular CCD element could be useful as well, as those are sensitive to infrared too. The only problem is to filter the visible light to leave the infrared spectrum only. Also be sure that the ccd doesn't have a IR filter (some does).

For 21st century Predator vision, this is the closest to it. Thermal IR from FLIR - https://www.youtube.com/watch?v=L7URETAl75A. Someone should hack this into a pair of glasses.

Never Give Up.. This should have take SOOOOO long, with so many disaopointment ! Really nice job

I got the feeling this did take a long time, he mentions things becoming available, and technology changing.

"This project spans years, and when I started there were no 3D printers. Eventually I made one"

This is probably my favorite line from the entire post:

"Okay. The purpose of the problems is to be solved."

People who dig to understand and create stuff like this are just amazing.

Now that's real engineering.

This was a fascinating read. Thanks for sharing.

Great read. Tireless enhancements until goal achieved. Great job.

This is a great story, but the original title of "Artem vs. Predator" is much better than the submitted one.

The title is artem v predator, but gives zero information of the content. I clicked out of curiosity but usually avoid vague titled links.

Indeed, the title brought to my mind the Alien vs. Predator movies, so I clicked to see what it meant when you replace Alien with Artem in that context...

I didn't click on this for ages because the title contained not a clue that this was such a cool multi-year exploration of infra-red radiation. I only clicked eventually because it stayed on HN front-page for so long.

I amuse myself with the thought of this contraption being fitted inside the Predators face mask. I'm easily amused.

Reply on news.ycombinator.com