Thursday, 19 August 2010

Fluorescent lamp + IR filter =?

A quick diversion from infrared photography to Infrared Luminescence (aka fluorescence). This can be used to reveal faded writing on old and damaged documents, and for other forensic things. I recently discovered a somewhat elderly but fascinating article by the legendary Andrew Davidhazy (of Rochester Institute of Technology in the States) explaining a do-it-yourself approach to this. Usually you need an expensive infrared-blocking filter over the broad-spectrum/white light source and an infrared-pass filter over the lens. Also usually, if you try and take an infrared photo under fluorescent lights you get almost nothing but noise.

Andrew points out in this paper that the average fluorescent light bulb (which presumably includes a lot of those low energy bulbs now prevalent) has a strange spectrum, attuned to our vision, which includes (virtually) no infrared. So if you illuminate your document with this and have an infrared-pass filter over your camera lens, all that will get through is infrared resulting from fluorescence in the document. The light, which includes no infrared, causes some of the document to fluoresce at a longer wavelength and some of that gets through the infrared filter.

Something to try I think.


  1. According to your view point it is possible to acquire face images using Fluorescent light and SWIR filters. My question is how we acquire images using fluorescent light which have range only upto 800nm and and we are using filters like 1400, 1500 etc?

  2. This technique will only work if the spectrum of your fluorescent light and the band-pass of your infrared filter did not overlap ... and that the object would itself luminesce within the band-pass of your filter. It is inherently problematic since luminescence is usually quite 'dim' and the sensitivity of a digital camera over 1400 nm would be, at best, poor. Argon peaks in the output of a typical tube are very slight over 750 nm so I'd suggest trying an 830 nm filter initially.There's some interesting fluorescent light spectra on the Wikipedia page:

  3. what do you think that Fluorescent lamp are better then led tubes??

  4. I Am also here to know that how we acquire images using fluorescent light which have range only upto 800nm and and we are using filters like 1400, 1500 etc?

  5. Fluorescent tubes will work for this because they have a specific spectrum which does not include infrared wavelengths. The same would be true of LEDs so they should also work. (Check the spectra of your lights.) What happens is that the photons of visible light excite the material and it fluoresces at a longer wavelength (lower energy). What wavelengths are given off will depend on the exciting wavelength and the properties of the material.

    But to answer your specific question (if I understand you correctly) you illuminate your material with your fluorescent tube and use a filter that excludes the tube's wavelengths completely. If your material is excited by the tube and if that fluorescence is longer than 1400 nm.

    However, when you get to wavelengths as long as 1400 or 1500 nm the sensitivity of the CCD or CMOS sensor is not as good.

    Maybe someone who understands fluorescence better than me can chip in.