Use of Reflective Tape as an Enhancement for IR Cameras

[nbk2000]

As decribed in my Security Architecture PDF, the use of retroflective tape as an enchancement to the detection range of IR cameras, in this case a Sony Handycam using NightShot mode at a range of 40 yards.

Download the video, and check on the file Properties/Content/Description for details.

152KB download, Windows Media 9 format.

http://rapidshare.de/files/16599578/Reflective_Tape_Enhancement_for_IR_Cameras.wmv.htm l

[Kamisama]

maybe i don't get it, but could you simplify what this all is?
All i see is reflective tape in the background and something walking(?) past it.

[nbk2000]

All i see is reflective tape in the background and something walking(?) past it.

That IS the point.

The use of cheap reflectors can extend the range of detection of IR sensitive cameras by a factor of 10x or more. My handycam won't pick me up at more than about 5 yards, but it can 'see' my movement at 40 yards or more if I get between the camera and a retroflector.

Practical applications?

Say you want to monitor a large backyard. A videocamera with motion-detection (not PIR, but pixel) can only work at night if:

A) The yard is lit up bright enough to get the perimeter to above the minimum detectablility of the camera (usually over 1 lux for color),

OR

B) An IR illuminator is used, making the yard appear dark to intruders (not equipped with NVD's).

In A), since they already know they could be seen, the intruders will take other (less detectable) routes in.

With B), they'll very likely cross in the open, believing that they'll not be seen,thus being detectable to the cameras motion-detector.

But an IR illuminator that can project enough light to provide the needed S/N ratio for the detection circuit to work is going to cost several hundred dollars (per camera=$), as well as be visible as a red-glowing spot. Unless you go with laser illuminators ($$$), you'll not get the range to detect an intruder until they're almost at your door!

But...

by using retroflectors, such as safety tape, bike and car reflectors, glass bead paint, etc., you can use cheap LED illuminators that will provide enough illumination to detect an intruder when his body blocks the highly efficient light reflection (>90% in some cases), rather than trying to detect his body reflecting the light (<10%) at great distance.

Even if the intruders have NVD's (*SNORK* ;)), they'd think the illuminator too weak to detect them at such range, not realizing that the fence and trees behind them is betraying their presence by reflecting their shadows.

This is almost like an IR break-beam detector, only you use a video camera instead of a photo-cell, as the detection means. Which also allows you to discriminate between a human intruder and animals, as humans are bipedal and there's no wild animals in the US that walks on two legs (unless you have feral ostriches!), so by simply setting up two parallel horizontial reflectors, knee and chest high, you can tell posture/size/direction/speed/possible weapons/etc.

In the demo video, you can tell direction and speed with no problems, and that it's obviously a human and not a dog, all with a single reflector stripe.

How about watching doors and fences? Put a reflector on them and the detector will sound if they're opened, with no wiring to be bypassed. Yes, they could remove the reflector from the door, but they'd have to be real careful not to move it out of line, tilt it, or block it, otherwise the detector goes off. Epoxy it on and its almost fuck-proof.

I've got several bike reflectors hanging off the cable I put up across the driveway at dusk, and they look like bright flashlights in the pitch-black to the $20 videocamera I have watching the driveway, and this at over 20 yards away. I can stand right under the (roof mounted) camera at night and not be seen by it's weak built-in IR illuminating LED's, but I can tell if the cable is up/down, moving, or if someone walks in front of it, all because of the reflectors.

Rain or dense fog could be a problem, but this is just part of the 'security in depth' strategy.

[bipolar]

This is a very good idea. I have also been thinking about using camera's as motion sensors also. This really makes it more effective especially for a large area, and harder to defeat than a normal motion detector.

Use a convex mirror above one camera and get a 360 view or motion detection capability. You can also take a picture or video with one and convert it to a panorama picture with free software for photoshop.
Check out this site: http://www.crosus.de/panorama/pancam8_eng.html

Also, NBK do you know of a good standalone cheap circut to detect motion straigt from the NTSC output or straight from the imager that would flip a relay or something. For something that doesn't involve a computer. I've seen some commercial products, but they seem way overpriced.



I've found out that most or a lot of non professional digital camera don't filter out IR. All digital CCDs see Infrared, most of the professional use camera's have filters that block it out for a better picture.

The way to test if your camera does not filter infrared is look in the view finder of your digital or video camera and point your remote control at it and press the buttons. I was surprised to see a white fast blinking light on my cheap video camera.

You can either buy a filter that ONLY lets infrared in and blocks out all other light or make one yourself. You can make one out of a film strip negative as I've seen some tutorials on the web for doing.

Being able to view the Infrared spectrum is important and can be very revealing. You can spot infrared illuminators for security camera's that may be hidden. You can see through things you wouldn't normally be able to, like polarized and other sunglasses you can see right through.

I haven't tested but I've heard some states have infrared security measures on State DL's. I've seen pictures of infrared security on US and foriegn currency.

Check out http://homepage.ntlworld.com/geoff.johnson2/IR/ for some infrared images and how to convert a webcam to infrared.

Sorry I got a little off topic here. Great work. Is this Security Architecture a PDF you are putting out, or is it available for download anywhere? Sounds cool.

[nbk2000]

The jews are using convex panoramic cameras on their newest tanks to provide 360&deg; FOV. Combine this with video motion detection and automatic slewing of weapons towards the movement (might as well make them self-firing too) and you've got some lethal shit!

http://www.defense-update.com/products/o/omni-panoramic.htm

All the stand-alone VMD's I've seen are several hundred dollars, but I'd be using a DVR with built-in motion detection that'd cost just a little more, but have multi-camera digital recording ability too, so why buy a stand-alone?

For a 'ghetto' VMD, I've seen a circuit in a "Engineer's Mini Notebooks" Radio Shack used to sell (still does?) that uses a photocell attached to the monitor right over where the target of interest is, and it sounds off if the state changes. Would only cost a few dollars to make.

I've got a composite picture of the backs of US dollar bills, $1 to $100, showing the IR-opaque bands on the back, that I made using my camcorders NightShot function. Quite interesting.

The 'Security Architecture' PDF was originally a stand-alone PDF that I posted years ago on another board. I'll re-post it on Rapidshare.

[nbk2000]

As promised:

http://rapidshare.de/files/16923530/Security_Architecture.pdf.html

~520Kb download

[Jacks Complete]

That's a great pdf. The bed is great! Of course, if your bedroom is already secured, how the heck did anyone get in without you waking up? If you don't bolt the door (locks can be picked) then why did you spend the money on a fancy door?

Oh, and the "roof security" section is still a placeholder!

I'll add a few things to it, too.

Since video cameras and screens are so cheap now, you would do well to set one up instead of a peephole in the door, to eliminate the weakpoint totally, and a second one further away in the garden, that shows the area around the door from the outside. This removes any chance of being stormed by the police, or shot through the door. Wired to a video intercom or computer system, you could check who was at the door from your living room, which would also get rid of Jehovah's Witlesses.

Also, the radar/mm wave gear would pick up the bars in the walls, and so warn the offensive team, so the first step would be to foil line the wall before starting to install the reinforcement. Then they would be "worryingly opaque" rather than "obviously a trap".

You can buy small alarms now for a few pounds that use a battery and magnet to alarm doors and windows in a few seconds. Very simple, but they will alert you to anyone entering the door, even with the key, while you are there, the moment they open it.

The original topic of extending the range of the IR illuminators is brilliant, and better than my original solution, which was to have small single LED sources in the field, pointing back towards the camera. This active solution needs power, sadly, or it would be better than the reflective tape, as the ranges achievable are incredible.
The answer, however, is clear to me - buy some solar powered garden lights from the DIY store/garden center. Open them up, and remove the LED, and change it for a cheap IR LED (& possibly a resistor). Now plant them in the garden like normal lights. The solar cell with recharge them every day, and they will glow all night. They will even glow for a lot longer than the white or blue LEDs used previously, since they require far less power - low power versions use about 2mA! - whereas white LEDs use near 60mA. Anyone looking will think they just don't work!

[akinrog]

The way to test if your camera does not filter infrared is look in the view finder of your digital or video camera and point your remote control at it and press the buttons. I was surprised to see a white fast blinking light on my cheap video camera.
Being able to view the Infrared spectrum is important and can be very revealing. You can spot infrared illuminators for security camera's that may be hidden. You can see through things you wouldn't normally be able to, like polarized and other sunglasses you can see right through.

Check out http://homepage.ntlworld.com/geoff.johnson2/IR/ for some infrared images and how to convert a webcam to infrared.

With PC it's very easy to detect motion programmatically using web cams, I started a project (after being robbed) and attained a great progress. And by means of a USB extension cord and a small (laptop style) high resolution (1.3 MP) cam, I monitor the office throughout the day and if any motion occurs it simply takes a snapshot. In future I'm planning to add capability to the app to control multiple webcams (at least one for the office and one for the door). I shall test the IR thingy on a old webcam (low resolution CMOS type) and see if it works.

Developing such an app is simple, you shall use VFW (Video for Windows) API and pick three random lines on the captured image and compare them with the corresponding lines of the previously captured image. If there is a certain amount of change in the lines, then you trigger snapshot feature (it can also trigger a voice alert on your computer to draw your attention). You could even use Indy components to stream video (or still images) to a remote computer.

However recently my workload immensely increased and I cannot continue the development process. Shucks :mad:. Regards.

[akinrog]

Since I'm a little bit (?) obsessive person, last night I abandoned all the jobs I had and started working on this conversion of web cam into IR sensitive one.

And I managed it. In my previous post I said I shall use an old (and obsolete) webcam. But when I dismantled and assembled the thing. I noticed it's dead (RIP). Anyway, it was already half dead before dismantling.

For this reason, I sacrificed my cheap (1.3 MP) laptop style web cam which I use as security cam during the office times.

What I should emphasize those small thingies (the square ones) are so simple to dismantle and their lenses are so simple that their IR filters are removed very easily.

Entire procedure of opening the casing, removing the lens (by screwing out), removing the IR filter and assembling the entire assembly together took less than two minutes.

I played the thing almost till the dawn by using remote control IR LED to illuminate in the darkness and see what happens. And what must be emphasized here is that the non-glowing remote control IR leds are much brilliant and provide more (IR) light than the red glowing ones.

But there is a MAJOR problem with removing the IR lens, which I discovered during the daylight. The web cam lost almost all color tints (I mean red colors are white, even black shirts, etc. seem quite whitish) and consequently the contrast it normally provides; which in turn rendered my crappy motion detection software inefficient. In addition, since IR filter is removed the picture became very brilliant which makes it difficulty for one to distinguish faces under the broad daylight. I had to work on it to render efficient again. So far I cannot manage it very well.

In addition, you must be aware that when you remove the lens you shall most probably deteriorate its focus which is pain in the arse to fix (needs two persons to fix it one to look at the monitor the other to screw lens assembly in and out).

So, if you modify a web cam to render it IR-sensitive, then you must keep in mind that you shall most probably render your web cam useless under daylight.

I'm now planning to purchase small health heat lamps (IR lamps that are used for physiotherapy) to illuminate dark yard during nights. How about this idea? Are those IR lamps provide the right IR spectra for IR enabled web cams? Regards.

[bipolar]

The Security Architecture PDF is very good and informative. It has a lot of stuff I haven't thought about before. Good work.

akinrog:
Thats cool you did it. After you remove the filter that blocks IR, replace it with a filter that only lets IR through. This filter will block the visible light and only let Infrared light through. You can either buy one or make an improvised one out of a exposed film negative. That should solve your problem.

Also, you can buy cheap filters that fit on flash lights or other lights that only let IR through to be used as an illuminator.

Here are some links on improvised IR filters:
http://www.rit.edu/~andpph/text-infrared-filter.html
http://members.misty.com/don/irfilter.html

[DirtyDan]

It might be tricky, but heres a simple fix to the day / night problem. You could save the IR filter and also get some exposed film, as said; then, if you hooked up an armature to move them back and forth with an actuator of some type, you'd simply need a signal to flip it up (visible light) or down (NV). You would want to do it this way be if you were operating remotely. Or, you could even make it automated, controlled by the lighting. You just need some signal from the computer (parallel port signal?)

If you wanted to get fancy, you could probably rig some circuit up that would cause the armature to move everytime the camera turned off and on (through the LED indicator? It turns off and on again in Linux if you load / unload the driver module.) That would avoid any extra cables.


Might be easier to just get an expensive camera with pan/tilt/zoom. I'm sure some of those have a NV mode now.

Newer webcams have really surprised me. This ip-cam is very good quality and streams at a surprising rate. The fountain is a great way to show it off as well. Too bad it's >$1000. Here (http://www.opentopia.com/showcam.php?camid=4023)

[akinrog]

akinrog:
Thats cool you did it. After you remove the filter that blocks IR, replace it with a filter that only lets IR through. This filter will block the visible light and only let Infrared light through. You can either buy one or make an improvised one out of a exposed film negative. That should solve your problem.
Also, you can buy cheap filters that fit on flash lights or other lights that only let IR through to be used as an illuminator.


I don't want to lose the cam's visible light functionality since I'm using it as a surveillance (sp?) cam during office hours.

Instead, I opted for programmatically changing gamma property of the cam in my app on the fly. In this fashion, the cam is capable of reducing and increasing its gamma value thereby increasing picture brightness or darkness based on the level of ambient light.

In addition, by accessing camera properties, I programmatically converted it into a B/W cam, which saves bandwidth of USB line. Now, I'm planning to modify doorbell button to conceal a small laptop style cam into it and ornament the edges of the button with transparent IR LEDs (which emits no visible light). Or I may conceal it behind the corporate sign just next to the door, again ornamenting it with transparent IR LEDs.

The advantage of transparent IR LEDs, if they are embedded into say plexiglass with only their heads are visible, they will look like glass beads embedded into plexiglass, very pleasant to eye and nobody suspects them being IR LEDs. I hope I have time to do this, because due to these operations, I neglected all jobs I have, which means losing cash. Regards.

[akinrog]

I finally reinstalled the damn IR filter into its place, since no matter how much I played with the settings of the cam, I cannot get rid of white ghosts appearance under broad daylight. Now, it operates normally again.

In short, I could not set the cam's whiteness, whitebalance, brilliance or whatsoever which are necessary to get its correct luminescence and I cannot find any reference how cameras do this automatically. So I gave it up for the time being.

But after making the hidden IR (web) cam gadget for installing around the door (of course whenever I have time), I shall use the used cam for the IR cam and purchase a new one. Regards.

[FUTI]

I have funny idea...does anyone know exact spectral sensitivity (Intensity of light needed to produce a same current response vs. wavelength) of web cam. Where to look for that info? I don't think I will find it through Google (but I will try anyway).

Maybe I can "calibrate" it myself using the two IR LEDs with diferent spectral distribution curves/maximum? But...I'm not skilled in electronic and related fields to know how, where and what to look to purchase at least two appropriate LEDs to do that experiment. In other words...HELP. I got an old web cam to waste and can't find better purpose for it.

[nbk2000]

If you can get the manufacturers name or code off the CCD chip (good luck!), then you could find the spec sheet for it which should have the spectral response curve that you want. :)

[Jacks Complete]

I'd suggest that the ghosting and over-brightness means that the CMOS sensor is simply getting too much light, and glitching. Try adding a filter to reduce the light coming in a bit, and see if it helps.

Most cheap cams just speed up the electronic shutter system to compensate for high light levels, but it can only go so fast, at which point it starts to ghost and mess up. The quickest test is to take it outside into sunlight, and see if it simply whites out.

[FUTI]

Thanks to NBK and Jack to advice. Unfortunate purchase in the old time of no-name brand webcam is painfull mistake. However search based on the chip that gathers the data from the CMOS give two possible CMOS sensors that can be used with it. It seems that search narrowed the hits but it won't give better choise then flipping of a coin.

What do you propose Jack for a filter? Do you think that it is to much IR that comes to the sensor? I had an impresion that relatively small amount of energy is in IR area compared with visible part of spectra. Are those sensors somehow more sensitive to those wavelenghts - what are those some kind of bolometers/thermocouples? Maybe it is just IR LEDs that make problems...and then the Gremlins start yelling "bright light";)

[akinrog]

What do you propose Jack for a filter? Do you think that it is to much IR that comes to the sensor? I had an impresion that relatively small amount of energy is in IR area compared with visible part of spectra. Are those sensors somehow more sensitive to those wavelenghts - what are those some kind of bolometers/thermocouples? Maybe it is just IR LEDs that make problems...and then the Gremlins start yelling "bright light";)

Actually with ghost image, I mean pictures are so bright that people's faces look very pale (like white ghost) and their complexion becomes barely identifieble.

And I suspect this (i.e. extremely bright pictures under sunlit conditions) is due to overflooding IR spectra of natural sunlight. When I close the curtains, the cam is getting very good image even if the ambient light looks quite dim to my eyes.

Anyway this was a good test to demonstrate conversion of an ordinary webcam into IR sensitive one. And whenever I have time I shall make a new fancy ;) corporate sign to provide surveillance.

And NBK's idea even inspired me. If one can write an app, which shall observe certain dots (maybe small non-glowing IR-leds, I referred to above) on the picture taken by the web cam, one can even make motion detector which covers really large areas.

When an object (human body, etc.) moves before certain dots (thereby making them invisible to web cam), the app simply triggers an alarm.

I believe it would be a very interesting project. Regards.

[FUTI]

I found a spectral sensitivity for one of two posible CMOS used in my webcam. I will cut it out in a day or to and attach to this thread as an example. The other CMOS sensor datasheets didn't give those information but I won't give up that easy:). Thanks akinrog for clearing up the confusion I was (and maybe "infect" other minds with it). I didn't read carefully enough to understand that ghousts were appearing while webcam work in daytime. I wasn't myself that time I wrote...which is obvious by number of typoes and grammar errors - my regrets to NBK.

For motion detector why wouldn't you use LED IR and set of mirrors and prisms around room projecting net that you would monitor.

[FUTI]

I hope that I will get it right, and that the file I attached with this post will get approved by mods.

Good side is that it explain some of the effects mentioned in the thread, bad side info it provide is somewhat confusing to me. First diagram shows spectral sensitivity up to 700nm, then the "monochromatic" (meaning black-white or 8 colors of grey probably) give another distribution far above previous wavelength. Can I asume that removing IR filter and placing VIS one instead will produce camera whose spectral sensitivity will be that region above 700nm?

Has anyone idea how to calibrate that CMOS chip if needed? :confused: Only thing that crossed my mind is usage of some reference material with known temperature of phase transition for example.

[akinrog]

Today I programmatically played with adjustments of the webcam and determined the reason for why I cannot change brightness, gain etc. is that the webcam's driver simply do not support changing them.

But I could change exposure setting. And I remove the filter again play with the exposure rate during daytime.

BTW, I am impatiently expecting the FUTI's attachment's approval by the mods. Regards

P.S. I determined that the best method to install the secondary webcam around the door is the peephole. But I'm not sure if the peepholes plastic optics shall pass IR information.

[FUTI]

I forgot to add link I find usefull

http://en.wikipedia.org/wiki/Infrared

akinrog you didn't answer me about posible filter for visible light...are you planing to use it or you completely ignore the influence of visible light? I see that you hope to circumvent the problem on software level so I asume that you plan to use it in night regime only.

I found the spectral distribution for other CMOS sensor...but this time producer didn't incorporated IR data or monochromatic. However it stated the type of filter and light source characteristics used when they recorded visible wavelength spectrum. If you need that one too...ask and you shall recieve.:)

[nbk2000]

A BIG webpage about using door peephole viewers as fisheye lens for digital imaging. :)

http://aggregate.org/DIT/peepfish

And any plastic that's clear to visual light will pass IR light as well, unless specifically designed not to.

[akinrog]

I forgot to add link I find usefull
akinrog you didn't answer me about posible filter for visible light...are you planing to use it or you completely ignore the influence of visible light? I see that you hope to circumvent the problem on software level so I asume that you plan to use it in night regime only.


No Sir, I don't plan to use a visible light filter for such a simple reason that removing IR filter makes the cam more efficient. The cam can take pictures (snapshots) even when the ambient light is quite dim to eyes. As you know, web cams cannot get proper pictures when the ambient light is low. This is a plus. But it has a disadvantage, you lose (or misrepresent) the color information. :(


I found the spectral distribution for other CMOS sensor...but this time producer didn't incorporated IR data or monochromatic. However it stated the type of filter and light source characteristics used when they recorded visible wavelength spectrum. If you need that one too...ask and you shall recieve.:)

Thank you. But for the time being, I don't need it. Because I optimized the algorithm to determine brightness of the snapshots taken by the cam, so I can adjust cam settings more efficiently now. No more excessive fluctuations any more.


A BIG webpage about using door peephole viewers as fisheye lens for digital imaging.

http://aggregate.org/DIT/peepfish

And any plastic that's clear to visual light will pass IR light as well, unless specifically designed not to.


Thank you for the inspiration, Sir. I believe web cams have so many improvized uses. You have mentioned about a panaromic camera that Israelis are using on their tanks. I guess that kind of thing uses a tapered (or conical) mirror just placed above a high resolution camera whose objective is vertical (i.e. facing upwads). This setup may have many uses in surveilance.

I even imagine a camera setup which was installed upon a very high (say 10 m) flag pole, comprising of a concave mirror and a high resolution camera beneath it.

With this setup one can attain panaromic view with a very long range. This option can be very useful for monitoring around if certain individuals (guess who) are approaching. :D Regards

[megalomania]

I don't know if this is relevant to the thread since it is for a digital camera, but recently I was talking to a fellow who told me about his website to attach a webcam to a digital camera. My Rebel is just high enough on the copystand to make focusing on a document a bit of a chore. Anyway, he has an excellent tutorial on how to remove the IR filter from a Rebel XT.
http://ghonis2.ho8.com/rebelmod.html
He has some suggestions about correcting the color balance and using filters that could be helpful.

[akinrog]

I don't know if this is relevant to the
....
http://ghonis2.ho8.com/rebelmod.html
He has some suggestions about correcting the color balance and using filters that could be helpful.

Certainly helpful Sir. Instead of programmatically changing color balance, brightness, etc. This one is more helpful. I am checking the document and convert it into PDF for keeping it for future reference.

Now, I am struggling with the computer threads. They give me a lot of trouble. :( Regards.

[sparkchaser]

So if the camera has just become overly sensitive to light in general, why not use an automatic irising apperture like that used in so many high end security cameras? I can't imagine it being too overly difficult to find one in one of the many surplus electronics wharehouses on the net these days. Less concealable, true, but it would be worth it for a dedicated home defense system IMHO. While you're at it you could use tracking and firing coding on your computer like these guys!:D
http://www.hackaday.com/2005/09/21/robotic-sentry-gun

[bowhuntress]

I've got one better, It's called microrefective prismatic banding. It's quite simple to make these and they work better than anyhting I've tried. Simply band together 24" long fiber optic strands (10 to 20 work quite well) by wrapping then tightly around a 2" wide x 2-3" diameter clear plastic or glass tube. Use clear adhesive to tack the ends to the tube. Wrap the strands around the tube until you have only enough to wrap one more time. Glue one strand down at this point with a small drop and trim all but 1/2" of the strand. Then wrap the next stand 1/2" beyond the previous glue bead and glue down and repeat until there are no more strands. After everything is holding tight bend the loose 1/2" strands up so they may catch more light. Cut two pieces of reflective tape into a circle to fit the inside diameter of your tube and stick together and glue into the center so there's a reflective surface on both sides. It's hard to explain but you're making a light collector that looks like a hair brush. The rounder the cut ends of the fiber optic the more light. Hope this helps.

[sparkchaser]

This sounds similar to a stack of brewster windows, but in reverse and for ambient light. The only problems I see are very low resolution at long distance, difficult if not impossible to focus a camera through, and not very efficient for any thing but laser light, unless of course you are using it as a light filter for a *sensor* that's only looking for a certain narrow band of frequencies.

To use it for a camera would require precise grinding equipment, and to use it for light filtering would require a knowlege of what lengths/widths of fiber would be needed for your particular desired band. Unless I'm looking at it the wrong way. Do you have a reference or website about it?

[Jacks Complete]

bowhuntress, that's a nice idea for a retro-reflector. Seems like a lot of effort when you can buy a load of cheap retroreflectors for pennies, though, or buy the blaze tape on construction site jackets for very little. Heck, a bit of roadsign will do the trick, and they are free!

Along a similar line, however, is the light-up sights you can get for next to nothing. Take a plastic rod with a circular diameter. Cut to length, and paint one end with silver paint. Polish the other end, then point the rod at the camera. A huge amount of incident light is trapped, and it all comes out the end! I'm not sure how well this would work late at night in near-pitch black conditions, but in low light it sure does glow.

Yes, this is exactly what you see on the muzzle of a modern shotgun, rifle or bow sight.