Damage QA of electro optical NV sensors :)

[gwenhastings]

due to the recent appearance of the A130 and A140 and preassembled hosts for same, the artic and I am sure others sure to follow I have decided to start studying the forensic effects on Electro-optical sensors I use as part of the security system(s) I design and install.

I suspect that the security industry will shortly see a rash of video cam blinding attacks and I want to design/fabricate a number of protective lens caps for 445nm after I determine what is needed by testing sensors to damage and beyond.

has anyone in the forums looked at this issue yet ?

gwen

 

[Morgan]

^Only by accident!

You can see plenty of CCD burns in photos posted here! you don't need an awful lot of power for that either.

This is an interesting avenue of research. I would think it should be relatively easy to protect a purely IR/low light camera from 445nm, (or other visible wavelengths), as it simply needs a filter for the other end of the spectrum but daylight cameras? And protection of IR cameras from IR lasers? hmmm, I can see these last two being tricky. Are you certain this will a problem in the future? There are many lasers, particularly green, that could render either type of camera useless in the short or long term. I would've expected this type of study or protection to already be in circulation if it was an issue.

Others may have specific info or experience but I, for one, will be interested to see what the results are. Sorry I can't help more.

(Welcome to the forum BTW! :wave

M

 

[gwenhastings]

Hi Morgan,

ir != low light..

I am interested in damage from both spectral ranges, I formerly was doing these tests with an uncalibrated 20 mw 532nm and a 5mw 805nm. I will be purchasing a laserbee or similar for power calibration and assist my building some much higher powered emitters.
One thing to note is the software based frame summation feature used in midrange passive NV CCTV sensors and the sensitivity to various infrared spectral ranges and whether the sensor incurs more damage if this feature is used and also autoiris and current shutter settings.

ie most of my Lowlight cam installs are in full color at night from very low levels of illumination only reverting to black and white at the very dimmest illumination.

I also would like to test attacks against bolometer type sensors used in thermal imaging and new sensors from china may make that practical financially.

 

[HaloBlu]

Please keep us updated on your research. We should be able to provide some helpful info if the right members come by.
I plan to test 500mW-1W 980nm on a few sensors. 445nm & 532nm of course will be more common.

 

[Morgan]

Hi Morgan,

ir != low light..

QUOTE]

Yes, I was aware of that. Hence the IR/low light... As opposed to perhaps FLIR cameras which do, occasionally, get used in bright light but work in a different way from my understanding.

You may already know this but remember that a DPSS green, without IR filter, will output 532nm, 808nm and 1064nm so be sure to add that into the mix.

All very interesting stuff and I look forward to future posts. Be aware that a Laserbee won't be able to measure below 5mW as standard although I have my doubts these would cause much risk but hey, you'll soon be able to tell us that!

Good luck with the testing.

M

 

[HIMNL9]

No need to use a 1W 445nm for ruin a security camera ..... i install them too, and know for experience that some was ruined from 100mW greens and 200mW reds, without too much problems .....

And, sure, there is the possibility to order to some specialized laser / optic factories a multilayer, multiband blocker dichro filter ..... but i just don't dare to imagine how much they can ask you for one of these filters that can block 635, 670, 532, 445 and 405nm (just for considerate the more available ones), and let pass all the rest ..... i mean, does it worth the expense, if for a filter like this, they ask you 20 or 30 times the cost of the whole camera ?

 

[gwenhastings]

normally I would agree with you if the sole purpose was to ruin NV sensors but as I stated this is a forensic study and I will be taking polairized light and possibly electron micro-photographs of the induced damage for comparison purposes for insurance and other industries.
And just for your information I dont just install them I also integrate them(nv sensor assemblies) and the software that goes into them for custom cameras with custom firmware, I have just started working with integrating the current crop of HD and larger sensors into security camera frameworks to start a line of lower cost LP capture cams.. but ALL my work I do at this level first. As the HD sensors cost quite a bit in comparison to normal resolution cams (See mobotix pricing on 3 mp cams) and I am ultimately aiming at the 8mp and up sensor class.

As far as dichroic multispectal filters go I am looking for cheaper "maker" type solutions.
and my belief is filtering WONT be successful to still have a cam image if all those frequencies are included.

Again this is NOT about destruction and instead is about classifying the nature of the exact damage incurred by such a device.

ie I have a couple of my custom astro grade NV cams in the 3k price range in the oakland riot area that appear to bear some type of damage(and they are mounted on rooftops) , that is why my interest is up.

 

[HIMNL9]

Oh, i understand what you mean, now ..... sorry, English is not my main language.

I agree, then, just filtering cannot be a solution ..... also cause there are too much ways for damage or "blind" a survey cam, if someone want to do it (just thinking the more easy one, a CREE P7 flashlight with a magnifying lens in front ..... clip it on something pointed at the objective and, "tadaah" ..... cheap long lasting blinder )

And yes, having an idea about the type of damace occurring in the chip, can give some suggestions about improving them against this type of damage, too, but i personally don't have too much hope, about this ..... megapixels and HD sensors MUST be very sensitives, for their own nature, so i think is a bit difficult to produce some HD chip with decent night sensitivity, like 0,01 lux, as example, that at the same time does not go at least in saturation when shooted with some high lightflow .....

I had an idea about this, times ago, and proposed it to a camera manufacturer, but it was like 15 years ago, the time was too early for the available technology, and the cost was absurd, so we left all drop ..... but the actual technology have increased a lot the possibilities, so if you want to take it in consideration, as suggestion .....

Basically, my idea was to substitute the front glass that protect the sensor chip with an LCD high speed "variable shutter", drived directly from the light level detected from the sensor (or also better, as in your case you use a software for manage it, from the software) ..... the main principle was that, if i change the transparence of the LCD plate proportionally to the detected light (after the maximum threshold set), it may act as an "high speed phisical autoiris", preventing the damage of the sensors ..... seeing that you use a software for manage the image processing, the natural evolution of the idea, can be to make a "grid type" lcd panel, so the software can detect the higher lights in the different zones of the sensor, and drive the grid for opacize only the zones hit from the high level, and left the rest clear, so the camera can still give some decent image ..... 15 years ago, a similar setup was a technical nightmare to build, but with the actual high speed LCD panels, it must be easy enough .....

EDIT: BTW, nice job you have, there ..... is interesting :beer: ..... and is interesting also to know the exact type of damage of the sensor cells ..... i always wondered about avalanche discharge due to the saturation, or pure and simple burning of the sensitive layer, but mine are pure speculations ..... What type of sensors you use for your setups ? .....