Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> broken and smashed Flock cameras

I wonder how resistant the cameras are to strong handheld lasers. I suppose they could harden them against some common wavelengths with filters, but that'd affect the image clarity in normal use.

 help



I have worked with watt class lasers before and I implore you not to do this. Even if it's tempting. Most places where there are surveillance cameras are places where there are also people, and unless you want to hand out OD5 goggles to everyone in eyeshot... a pellet gun would be safer.

My friend in college did an internship on high frequency, short pulse beams (I wanna say violet and picosecond? Which I still think was exotic at the time).

Most of his work was dealing with and accounting for reflections that left the machine. If you have a prism that’s sending 95% of the light where you want it to go, when it’s a multi watt laser you can’t just let that 5% go wherever it wants. You will blind someone. So his job was getting black bodies in all the right spots to absorb the lost light.

His safety goggles looked like even more expensive Oakleys of that era and they were (much more expensive).


The amount of safety when working with lasers is ridiculous. And for a good reason, you can get permanent eye damage faster than the blink of an eye.

Please, don't play with lasers. At all. Even supposedly "safe" lasers can output far more light than expected.


Not to mention the ones that have peaks in invisible parts of the spectrum.

Another friend’s favorite saying is, “do not look into laser with remaining eye.”

Unrelated, but I really want to take the opportunity:

How can one know what is dangerous for the eyes or not? Years ago I got an "IR illuminator" (from aliexpress, probably) that I wanted to use with my raspberrypi NoIR camera, for fun. Say filming myself during the night to see how much I move while sleeping, or making my own wildlife camera trap.

But I was scared that it could be dangerous and never used it (I tested it in an empty room, but that was it).

Is there a safe way for a hobbyist to get an IR illuminator and be sure that I won't make somebody blind with it?


Buy from a reputable dealer. I don’t buy batteries, lasers, or items I ingest from locations lacking any repercussions.

Is it just a bunch of IR LEDs? Surface mount or through-hole? What's the module power rating? What's the power supply power rating? Are there any secondary optics like lenses over the LEDs? Is there a diffuser of some kind?

If it's a cluster of garden variety through-hole LEDs with domed tops (like you would see on a TV remote), they're necessarily low power on account of having poor thermal performance.

Another way to tell is if nothing gets warm at all. It's pretty hard to hurt someone with an emitter that both doesn't have a focusing optic and doesn't get warm.

Let me be clear - you're still responsible for verifying the safety of your stuff, and I am in no way assuring you that the device you have is benign, because I can't do that without inspecting it directly.


Yeah, let's say something like this: https://www.instructables.com/DIY-IR-Infrared-Illuminator-Ni...

> Let me be clear - you're still responsible for verifying the safety of your stuff

Obviously yeah. I was just wondering if there were known rules like "these wavelengths under this power are fine for humans and wildlife, even if they put the LEDs right in front of their eyes", and also if you have an array of such IR LEDs, how they cumulate.

And curious about things like: if I don't see it, can it hurt my retina?

I probably will never do it: I wouldn't want to blind a fox just because I wanted to make my own wildlife camera :).


Invisible wavelengths can absolutely hurt your retina, but as wavelengths go farther beyond visible, your eye begins to not focus them properly on the retina, so the risks change. E.g. with 1550nm IR (common in telecom, sometimes in LIDAR) the risk of eye damage is to the surface of the eye rather than to the retina. Short wavelengths like UV will be absorbed by the lens at near-UV, and then eventually just be absorbed at the surface at shorter wavelengths.

I think it would be a cool exercise to figure out how much optical power you would see at, idk, 5cm from your illuminator. I assume it's a shortwave IR close to visible light, so you can assume it will focus like visible light, more or less.

Ideally you'd use an optical power meter but you could get a first pass by looking at the circuit and seeing how many mW pass through each LED, applying a conservatively high efficiency factor of W_optical/W_electrical, projecting that into a radiated cone for each LED and multiplying the power received on a dilated pupil sized spot at 5cm by the number of emitters.

Then you have to work out what the irradiance at the retina is once the light is focused. The hazard criteria include a time factor, so you'll have to decide if you/foxes would like to stare directly into the beam for 10 seconds? Or for the entire duration of your meditation session.


Very interesting, thanks a lot!

I assume CCTV surveillance cameras are usually fine because people cannot go and stare at it from very close then :-).


i went down a rabbit hole doing max permissible exposure calcs for the light you linked and basically i personally wouldn't worry about it. the energy is low and distributed among many emitters. by the time you're far enough away from the light to focus on it, you're receiving max a couple hundred microwatts/cm^2 at the cornea from each led.

I don't know if you ever write blog posts or anything like that, but I would love to see explanations about how you made those estimations. I honestly wouldn't know how to even start :-).

IR illuminators are not lasers. Their purpose is to cast light across a broad area, not to deliver it all to one point. They should not be harmful to vision.

> Most places where there are surveillance cameras are places where there are also people

I assume you're concerned about reflections from the camera lens or housing? In my mind, the archetypal camera is mounted on a nice tall pole, silhouetted against open sky, and painted matte black.

> watt class lasers

Surely those would be excessive for someone attacking the sensor, unless they want to remotely sear some graffiti by burning away paint.


Hitting the lens at an oblique angle won’t fry the sensor though? You have to get close to the cone of visibility which is then within the bloom area.

Please do not encourage people to go shining bright lasers at small targets from long distances right next to busy roads.

This is a nerd fantasy thing, but it's a really bad idea. It's hard to hit a tiny lens from a distance and it only takes one slip of the hand to shine it straight into traffic or someone walking down the sidewalk.


Last I recall they’re just a crappy 5 megapixel Arducam camera module based on teardowns.

https://www.cehrp.org/dissection-of-flock-safety-camera/

https://www.arducam.com/product/arducam-ov5647-noir-camera-b...


Lol that's almost literally the cheapest possible option. You can get these for $3-4 (on a board and with a mipi cable and everything) from China - I have a dozen in a box that I bought to test out a camera array idea before shelling out for nicer sensors.

The best part is seeing someone tear a Flock camera apart, see the camera, and immediately go slap it on their 3D printer and hook it into their Pi and just have it work out of the box ;)

Maybe pick up one [1] and experiment with it. If I had some spare change I would love to grab one just to hack it.

[1] https://www.ebay.com/itm/297938376075


Do not do this.

Comments in the sub-$200 LiDAR thread suggested those would play merry havoc with a camera too.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: