Infrared photography of banknotes: anybody tried this?

I was having a look at some videos of a banknote I have ordered, and in one part of the video they show the note under 870nm infrared.

Of course, I like this!

But I want to know if anyone has already done this photography and if so, what kit they used. 

Just bumping this to see if anyone is doing IR photos.

I have ordered a Swiss Franc note which has interesting IR appearances and I will need to photograph it in IR.

I'm going to put the infrared info from the other thread in here, as it is where people will look.

I bought two devices from China via AliExpress. Surprisingly, the service was very good.

The first one is the Xindabill V100. It has a combined watermark and infrared function:

That unit does IR, UV and sidelight for relief. It has a magnetic detector also but I have not tested that yet. It cost me £137.

The bad part about that unit is that it wasn't designed to photograph notes. It is for checking in the shop whether the note is counterfeit. There is only one output which is a composite video signal (low resolution 625 lines). If you want images saved, you have to grab the video feed. I use a USB-based grabber:

Photographing the LCD screen on the unit results in too much Newton ring artefact, you have no choice but to grab the video. Another bad thing is, it has a small field of view, you can't get a view of the whole note unless it is small.

See this thread how the V100 helped me image two unusual watermarks on older Italian Lire notes:

Watermark question on P#111b [solved] – Numista

The second device I got is much more useful, it is a stripped down android device with a low res camera and a small IR emitting LED:

It will save images in slightly better resolution, about 1000 pixels across. It can take a microSD card for extra storage and can write directly to that. That cost me £61.

There is one bad thing, one interesting thing and one lucky thing:

1. The little IR LED that is on the unit is not very strong. It is enough to check real time whether a note is fake though.

2. I bought an IR flashlight off Amazon and even on its lowest setting it completely washes out the image. It is too powerful, so don't bother going down that route.

3. Luckily: ordinary sunlight is perfect for this device, it contains enough IR light to get a nice picture. You can do it indoors with sunlight coming in through a closed window.

And that's how I got these two images:

You will need to raise the contrast in Photoshop a little otherwise the images are a bit flat. It also does not do that nice combined IR and watermark image. I still need the V100 for that (so far).

They are by no means ideal for producing images but for a start they are good enough. You can definitely see where the IR-reflective ink is and where it isn't.

Your alternate is to buy a full spectrum modified camera off eBay, get an IR lens for that, and then photograph notes in sunlight.

I may do that at some point, but so far I can at least find out which notes have interesting IR features and which ones don't.

I'll add some of them here in this thread.

Bermuda 2 Dollars (2009), plain scan left and infrared right:

Australia 50 Dollars (2018):

G&D “10 ECU” ATM test note (1992):

Aruba 10 Florin (2019):

I like how one serial number disappears on the reverse.

What a fascinating idea, I love all the work you're doing with UV & Infrared.

Not much to add, just 👍

Now do the ultraviolet ones and magnetic field properties (don't know if publishing these features might have consequences for you though 😅).

Don't give them ideas! 🤣

I did indeed look at analysing the magnetic properties of these notes but the machine is very expensive. Someone else can do that…

As far as the UV is concerned, I check every note I acquire under UVA and UVC. I check for phosphorescence too. I have a nice UV collection now, not as large as @gyoschakbut I have a decent selection of notes that have different appearances under UVA vs UVC. I regularly add UV images to the catalogue.

The infrared is interesting to me. It will result in a lot of work and a lot more photographs of my collection. I think I'm one of the first guys adding IR images to the catalogue.

I screengrab the notes while they are still in their sleeves, using the V100. The sleeves cause a reflection of the IR LEDs, but I can at least work out whether a note is worth photographing before I take it out of the sleeve.

I don't want to handle the notes unnecessarily. They have already been scanned and photographed under UV. Just picking a note up off the glass of a flatbed scanner requires some care: I have a thin piece of flexible plastic I use to lift one edge of the note up. 

Here are two video grabs from the V100 unit where the note has not been taken out of its sleeve:

If I decide to photograph a note, I have to wait for decent sunlight on the weekend because I cannot do this in working hours. The infrared photos are done on a piece of black acrylic plastic and that is a much easier process because I can just slide the note off the plastic. But then it has to be done out of the sleeve.

odd job

I don't want to handle the notes unnecessarily. They have already been scanned and photographed under UV. Just picking a note up off the glass of a flatbed scanner requires some care: I have a thin piece of flexible plastic I use to lift one edge of the note up. 

Great work on this!

Btw, I use a (non collection) polymer note to lift the edge of an UNC note from the scanner.

5 Euros (Europa Series) - Eurozone – Numista

I recently got some nice commemorative notes. Amongst them was this one:

20 Zlotys (Our Lady of Częstochowa) - Poland – Numista

Before the note goes in a sleeve I am doing plain scans, UV and IR photos. One thing of interest is three different ways I acquired the IR images.

The images that follow are all of the reverse of the note. The only image manipulation I have done is rotate, crop and lens correct. I didn't do any contrast, brightness or sharpening adjustments.

First, three cropped images from screengrabs of the video output of the V100 device:

Now the stripped Android device. Firstly, using sunlight:

And now the same device but this time the note was lit by a halogen lamp at an oblique angle:

From what I see, the halogen light is good enough. It means I can do these IR photos any time I like instead of having to wait for decent weather.

By the way that bottom left serial number is supposed to fade into view under IR. This is what it looks like on a standard flatbed scan (this one has been made brighter because my scanner makes the image a bit dull):

That green rectangle on the reverse absorbs IR, but doesn't do anything special under UV:

Good work on these.

The IR activity on some of the portraits is interesting, and cleverly implemented.

I am writing a fourth article on LinkedIn to do with banknotes, and part four is about IR. It is proving to be harder to do than the other three parts.

I am finding it hard to pin down a starting decade for these IR features on banknotes. The examples in this thread so far are easy: there is an obvious abrupt change in appearances when comparing the white light photo and the infrared.

But in a lot of cases, whole features have disappeared in circumstances where it isn't clear that this was intentional. A typical example is red serial numbers and pale red or pink features that disappear.

If the note goes back to 1915 you can be fairly sure they didn't have infrared features in mind. But what about the 70s and 80s?

Even on some later notes it is tough to work out whether the infrared appearances were deliberate or coincidental. Here's an example from the 1990s:

When I first saw the note I thought that it had deliberate IR features on it. But the problem is, the areas that disappear are in red ink. It might well be that those colour choices resulted in that appearance under infrared. The same thing happens with some other notes that have a pale pastel background: they can disappear under IR.

I'm finding that many colours that are near infrared in the spectrum (such as red, pink and tan) disappear under infrared. Those colours must naturally reflect IR or at least not absorb it. Those features disappear under reflected IR and transmitted IR imaging.

I think regardless of the intent of the bank, the IR features can be used to verify whether a note is genuine or not. For example if somebody faked that 500 Tolarjev note and the red 500s appeared under IR, you could identify it as counterfeit, even though the original printer/bank never intended the note to have IR security features.

It reminds me of this 1982 Allied Irish banks £1 note under UV:

It has a segmented security thread, very subtle even under backlight. On the reverse, it passes over the rear of the ship.

That thread fluoresces under UV.

When I tried to make a catalogue modification request with that info, my request was rejected with the reason being “This is not a security feature of this note.”

If that is the case, then the UV appearances may not have been what the bank was aiming for, but nonetheless those appearances are expected to be present on a genuine note. It is a method that an ordinary person could use to see whether their note is real or note, or a certain type or not.

It may well be in the future we have some other kind of imaging that wasn't thought of when the notes were produced, but it may well be a useful modality to determine whether a note is real or not.

I think X-ray falls into that category, at least in some cases. X-ray imaging has been around for a long time, but not in a way that is useful to X-ray banknotes until the last 20 years. This is because a single sheet of paper requires certain technical factors and processing in order to discern differences in ink densities on paper. And those technical factors are a recent development in diagnostic radiography.

I don't know who else is X-raying banknotes but I can say I have made a start on it. Possibly they are doing it behind closed doors in forensic labs.

One advantage I have is that it is my job: I am a radiographer. It is easier for me to locate the appropriate equipment and interpret the images.

Some of the IR activity will be incidental. However, especially on later issues, it appears to be quite deliberate.
Your work on these has prompted me to investigate IR activity on Irish notes in detail, need to get access to a lab for that.
 
Many of the modern security features, UV, IR, holograms, and optically active ink appeared in the 1980s and later.

odd job

That thread fluoresces under UV.

When I tried to make a catalogue modification request with that info, my request was rejected with the reason being “This is not a security feature of this note.”

You got my attention! The thread is metallic. Is it fluorescence or reflection. 

I have this note, and had a look at mine under UV, didn't see anything special.

However, I stand to be corrected, and am always open to resubmission after discussion…

That's the P#137. 

I checked my P#138 and it looks the same under IR. I don't have a P#139 yet. I do want a P#139 because I found out that the P#137 and P#138 differ significantly under UV and I want a P#139 to check under UV.

Here's an example where transmitted IR is helpful. The left image is plain backlight, the middle one is reflected IR and the right hand one is transmitted IR:

Like the Italian note previously, transmitted IR removes most of the background and leaves watermarks visible.

This was a fun read and will really help with counterfeit detection. I think I will try it too. Unfortunately I also have not found a device with photography in mind. I have recently started to collect test (house) notes from different security printers and would like to see them under infrared. It would be cool to find intentionally placed activity…

There is a G&D test note that I bought off eBay with a special IR mark on it, this is the one:

Unfortunately that note never arrived and I was issued a refund. But certainly there will be test notes with IR features, after all the companies must showcase what they can do.

Oh thats a pretty one! I too also recently had one of my orders refunded after a similar incident. It was this Joh. Enschedé test note pictured below.

IR can make it easier to identify an iridescent ink pattern in circumstances where the background of the note makes it hard to discern. I recently bought one of these:

10 Rials - Haitham bin Tariq - Oman – Numista

On the reverse there is an iridescent ink strip made from two colours. One is a pinky-peach colour and the other is green. It is very difficult to photograph, even if done in pieces:

In the middle of the ink strip is the denomination of the note, a green 10.

Now look what happens under reflected infrared photography:

The iridescent ink reflects infrared light. A bonus is that the green ink reflects more than the peach/pink. So you get the clear geometry of the pattern and an indication of which parts are in different iridescent ink.

It is a shame I got this note after I wrote the LinkedIn article. I would have included this one, it is that good. 

This is a link to that article:

https://www.linkedin.com/pulse/banknotes-not-so-obvious-features-part-4-infrared-brandon-bertolli-fjane/

Very interesting articles.

This is a nice example of transmitted infrared helping visualise a watermark:

25 Denari - North Macedonia – Numista

This is the backlight scan:

And this is the transmitted infrared. It shows the watermark pattern nicely:

Under reflected infrared you get much less contrast:

The watermark in the transmitted IR is still discernible. Lack of contrast could likely be a significant issue on a well circulated note though.

I thought I would show you side by side the difference between my home-brew infrared photos and a photo taken on a VSC9000, courtesy of the Foster and Freeman forensic company. On the left is my infrared photo and on the right is the VSC9000 image. These are not the same note but they are both UNC:

My image on the left has an unresized horizontal pixel count of 1858. I have resized the VSC9000 image on the right to match that, but it was originally 3588 pixels across.

Another thing I noticed: there was no lens distortion on the VSC9000 image. But on my image I had to lens correct it because it had barrel distortion.

Though the image on the right is better, your home-brew image stacks up very well against it.

Here's an interesting one: a test note that has different appearances under different wavelengths of IR.

This is the Nature 2022 series of four notes released by Goznak. They all have a similar reverse image with a tiger on it, and this tiger has a different appearance under 850nm vs 940nm. 

This brochure describes the note:

GOZNAK | Products | Souvenir banknotes | "Nature 2022" banknotes

The V100 device that I bought has a mode where it pulses from 850nm to 940nm. Up to now I haven't seen anything change appearances on this pulsed mode, but I verified today that this nature 2022 note does in fact change. Look at the paw prints…

Here's 850nm:

Here's 940nm:

I'll duplicate my answer about this from another site so that Numista forum users can also find out.

This behavior can be accidental (when the manufacturer didn't pay attention to the different IR values) or intentional (in Goznak terms, this is called the Special Element M - "Спецэлемент М").

Goznak used different responses of inks in IR with different wavelengths (officially since the 1997 series, but I've also seen this behavior on banknotes from 1995).

The Special Element M is an image applied with a special ink that exhibits different behavior in the infrared spectrum at different wavelengths. Typically, a pair of sources with wavelengths of 850 nm and 940 nm is used in detector. When illuminated by one of the sources (usually 850 nm), this ink absorbs the light (and, as a result, the image created with this ink appears "dark" when forming the IR image of the banknote). When illuminated by the second source (940 nm), the ink reflects the light (and, accordingly, the image element appears "light").

I've seen this on Russian 1,000-ruble banknotes from 1995, 500,000-ruble banknotes with date 1995, the entire original 1997 series, and the 2001 modification. Starting in 2004 (including the 5,000-ruble note that was issued in 2006), this security feature was removed. However, it appears to still be used on advertising banknotes.

Left image of each banknote is IR850, right - IR940.

This is very interesting! I don't think I have checked my notes carefully enough with 940nm. I did a quick check now and here is what I can confirm with my notes. In each case the left image is 850nm and the right one is 940nm. These images are done in plastic sleeves, so there is reflection of the LEDs from the V100 device. I don't want to take them out of the sleeves yet.

Russia 5 Roubles 1997:

Russia 5 Roubles 1997 (2022-2024 print):

Russia 100 Roubles (Sochi Olympics) 2014. The regular and the replacement issue do the same thing:

This 100 Roubles “FIFA” note from 2018 is strange. Ball pattern is lighter at 940nm and on the other side half of the 2018 date disappears. Might be non-intentional:

Russia 5000 Roubles 1997 (2010 modification). This could be by accident, the red serial number disappears under 940nm. I have two of these notes and they both do this.  

I checked a 500 Roubles from the same year and the red serial number also disappears:

I checked three 1000 Roubles notes - 1997, 2004 and 2010. There is no difference between 850nm and 940nm.

I also checked four 10 Roubles notes from 2004. There is no difference between 850nm and 940nm.

I also checked three 50 Roubles notes from 2004. There is no difference between 850nm and 940nm.

I also checked two 100 Roubles notes from 2022. There is no difference between 850nm and 940nm.

I also checked three 200 Roubles notes from 2017. There is no difference between 850nm and 940nm.

That's what you said would be the case in the post on the other forum.

odd job

Here's an example where transmitted IR is helpful. The left image is plain backlight, the middle one is reflected IR and the right hand one is transmitted IR

 

Like the Italian note previously, transmitted IR removes most of the background and leaves watermarks visible.

I've found a similar security feature on banknotes issued by the Banque de France.

There's a watermark beneath the portrait: several low-contrast stripes that aren't noticeable in a plain backlight. But because the portrait becomes invisible under IR, these stripes are easy to spot. Me and my frend (French banknotes collector) call it the "Mysterious Gradient Square", as no collector of French and colonial banknotes has been able to confirm whether anything is known about it. We also couldn't find anything about this security feature in advertising brochures or public releases. When I asked directly, the Banque de France remained silent (as expected). But this feature has appeared on banknotes printed by the Banque de France for West and Central Africa, Djibouti, and other countries since the early 2000s (I'm still searching for the earliest banknote with this feature).

(I have no scan of exactly this ^^^^^ banknote, but this is 2000 francs for Central African States N#205066 )

I suppose its appearance with the surge in counterfeit 10,000 franc West African notes in the early 2000s, but I don't have any evidences, just guesses.

That is interesting, I noticed it on several notes also. The one that caught my eye was the 500 Francs from the 1990s, which is significant to me because on there is Marie Curie and her X-ray van.

That is a very nice note for me, because I am a radiographer so it is in my area of interest.

That is one of the first notes I X-rayed and of course a lot of features including those bars are showing up on X-ray:

Looking at banknotes with an X-ray... an extremely unusual approach.

But it's important to me - I see new timelines in this security feature ))

I am building quite a nice X-ray database of banknotes. There are some things that are visible on X-ray that you can't see with backlight, UV or infrared, mainly on the security threads. The magnetic components of a thread will be dense on X-ray, for example. 

It's a long side project I am doing. If I had the money I would have my own machine - the one I use for the banknotes can be installed in your house, it doesn't need special shielding and it is smaller than a half size refrigerator.

Amazing!

I'm trying to build magnetic scanner for magnetic inks and threads as a hobby... for several years now )) There are few industrial solutions, but they are expensive.

I was lucky: I was shown a top line magnetic detector at the Foster and Freeman Forensic company earlier this year.

I also want one of those but it is beyond my reach.

Here is a screengrab from their VSC9000 interface. This is part of the magnetic security thread on the Ukraine 50 Hryven note from 2024:

And here is part of the radiograph, flipped vertically to match the magnetic map:

You can have some fun, trying to pick up that sequence on backlight, UV or infrared…

Can VSC9000 take a full magnetic image of a banknote? Or does it only show a specific section?

It is a small area about 2cm x 2cm in area. Here is the brochure:

VSC Magnetic Security Feature Imager

They image the entire security thread by recording the output in real time as the note is moved across the device. They did that for me and saved it as an MPEG.

In practice they can prove a note is counterfeit just by showing one small area, so I guess that is the reason they haven't tried to make an expensive plate to image an entire banknote.

I would like that though!

It's very sad... I'd like to see the entire magnetic image.

I understand that one of the goals of manufacturers of such devices is to prove the authenticity of the document, but in my research I use known-to-be-authentic banknotes, and I'm more interested in the overall visualization. In any case, thanks for the information.

I got this 1998 France 500 Francs note recently, in a PMG slab. There are IR850 vs IR940 differences, not sure if these were intentional though: