experiments with video image capture
using a near infra-red light source
by Dave Walker
If you have a black and white video camera for microscopy it may have a significant sensitivity to near infra-red light. This can be exploited for simple near infra-red microscopy by using a near IR light emitting diode for the light source.
The author's first venture into transmitted near IR microscopy is simply described below and inspired by Tony Dutton's reports (see Acknowledgements and Refs.). I was quite intrigued by the results and hope this article will encourage others to try it.
Near infra-red microscopy has potential benefits for capturing images of insect parts and other subjects such as tiny shells, horn and some minerals which in visible light are almost opaque. A bonus is that a significant increase in depth of field is achieved with IR (typically a 70% improvement) which is especially useful for the 'thicker' subjects which benefit from this type of lighting.
The blue hypertext links to the more technical aspects in an Appendix.
Many amateur microscopists use a black and white camera for video microscopy as they are relatively inexpensive and offer good resolution. The typical CCD sensor in the modern solid state B&W camera has a significant sensitivity to infra-red (more correctly near infrared), especially if designed for low light or night-time surveillance use. For visible light video microscopy this sensitivity is detrimental and an IR filter is essential to cut out the IR component emitted by a microscope lamp.
However, this IR sensitivity can also be exploited for video microscopy by using an inexpensive IR light emitting diode (LED) as the light source. The eye isn't sensitive to IR so a TV monitor connected to the camera is used for visual studies with images stored as stills/video on a PC or video recorder. The LED's are available for less than a pound (or a few dollars) from electronic hobby shops.
What do you need?
- Compound microscope (the subjects described below are best viewed at low power i.e. 3x-10x objectives.)
- Black and white video camera without a lens (they usually have a so-called 'C' mount). Note that the camera must be a security type with a removable lens. A domestic colour camcorder will not work because they have IR filters. (Note added Sept. 2000. Some camcorders and digital stills cameras now have a 'night mode' where the IR sensitivity of the sensor is exploited. Maybe these also work if attached to a microscope to capture IR microscopy images. I'd be interested to hear of readers who have experience of this).
- Infra-red light emitting diode (See sources).
How do I know if my black and white camera is sensitive to IR light.
Point a remote controller for a TV or video recorder straight at the video camera with the camera hooked up to a monitor. If you get bright flashes from the remote controller when you press the buttons it should work for IR microscopy. If you don't, it may mean the camera is designed for visible light use only and has an IR filter incorporated on the CCD sensor which cannot be removed. (Some cameras have the IR filter on the 'C' mount lens hence try again with the lens off).
Preparing the IR light source
Most of the IR emitting LED's can be powered from a single 1.5V battery. From the LED's stated specifications a simple circuit can be devised to drive the LED correctly (ask a friend with electronics knowledge if unsure). The LED should have a suitable resistor in series to drop the current to that recommended. I used a variable resistor (available from hobby shops) which acted as a lamp dimmer. Thanks to a reader, who suggested it is advisable to have a 100 ohm resistor in series with the variable resistor so that the LED is protected when the variable resistor is set to zero.
The crude (but effective!) set-up I used for simple trials is shown right. I've simply friction fitted the LED in the centre of a piece of card that fits the condenser filter holder and the card is held in place with 'Blu-Tak'. (The filter holder with LED would of course be swung in for microscopy).
If you don't have a condenser, the LED can probably be mounted somehow in the optical axis below the microscope with the LED 'lens' pointing upwards. Some of the LED's have quite wide beam angles so a condenser may not be necessary for even lighting.
Preparing the video camera
Mount the video camera in the way you normally would on the microscope without a video camera lens. I do all my studies with the camera directly mounted into the eyepiece tube with an adaptor, some prefer to use with an eyepiece and an appropriate adaptor or camera support. All adjustments to the lighting, microscope and camera have to be carried out by inspecting the image on a TV/video monitor (or a PC image capture board may give a good enough preview image to use).
Centring the LED
Put a slide on the stage and focus with a low power objective on the subject using the available IR light. (This is to first ensure the objective is focused).
Move the subject out of the field of view and inspect the light quality on the monitor. The LED's are quite small and may not be centred properly or the LED aligned quite vertically, so a bit of adjustment may be needed (this is easier if you have a centring condenser). If a condenser is being used, check the centring by racking down the condenser so the LED comes into focus. The high power LED I used had a four segmented array (shown right above). After centring, rack the condenser back up towards the slide to obtain an even light field. I was quite surprised how even a light was achieved from such a small source (for my LED the black cross was invisible after refocusing the condenser).
What subjects to try
It's best to start with prepared microscope slides. If you have a selection pick some which have some detail that is very dense i.e. almost opaque to visible light that in IR light may be more transparent.The present author tried a selection of subjects including those demonstrated and reported by Tony Dutton (ref.1, and asterisked below) e.g. chitin of insect parts, horn etc. Subjects that may work well are:
- insect bodies which are almost opaque in visible light *
- insect legs or jaws with semi-opaque features *
- semi-opaque shells e.g. foraminifera
- some minerals e.g. biotite *
- horn sections *
- most exciting of all is live subjects, e.g. where the IR may penetrate the exoskeleton of live insects, aquatic larvae to reveal internal features in motion (I've no experience of this area yet, let me know if you have tried this).
(Out of interest I have tried a variety of prepared subjects e.g. stained botanical specimens, histology to see if IR lighting offered any benefits. I obtained pale ghostly images for most of them as perhaps expected, let me know of other subjects that may benefit).
Comparison of video stills taken in IR and visible light
In the table below I have shown pairs of video stills captured for subjects taken in both IR light and visible light. The IR images were captured using the set-up described above. For visible light I used the same video camera with a 240V / 75W tungsten photo-enlarger lamp with an IR filter (available from photo-outlets) in front to cut out the IR component emitted by the lamp.
I wasn't too sure how to compare the images fairly, so have presented a selection of unretouched images and some post processed if the originals were too 'flat' (low contrast) or too dark. Compare the images more in terms of the relative detail conveyed and the depth of field as post capture brightness and contrast adjustments will often be carried out to obtain pleasing results without losing detail.
|Comments||Image capture in visible light||Image capture in IR light|
|Part of diving beetle middle leg with wing case below it. Note improved d.o.f, and clearer detail in dense chitin bottom left Objective 3.5X.Unretouched images.|
IR reveals detail in areas almost
opaque to visible light capture & with greater d.o.f. Objective 3.5x. For both images brightness increased 15%, contrast not altered.
|Blowfly leg (Calliphora). Visible light image brightness increased 30% as leg detail very dark in raw image. IR image capture brings out detail in thicker hairs especially the roots. IR image post processed to improve contrast.||
Click to see raw image
|Cinnabar moth larva skin detail. Visible image brightness increased 20% as raw image so dark. IR (unretouched image) penetrates the dense cuticle and gives greater d.o.f, and shows the fine hair patterning well. Objective 9x.||
Click to see raw image
- The depth of field (d.o.f) is directly proportional to the wavelength of light used, so a 70% improvement is calculated using near IR light cf. visible light. For a 3.5x objective this equates to a depth of field of ca. 95 micrometres for IR cf. 55 micrometres for visible.
- Resolution is also directly proportional to the wavelength of light used so there will be a loss of resolution when viewing in the longer wavelengths of near IR. However, the subjects that benefit from IR lighting tend to be viewed at low power so the resolution loss is not noticeable.
- IR seems to work particularly well in showing clearer detail and structure in image capture of 'muddled' subjects like dense insect leg hairs or folded areas of chitin etc.
- The simple video surveillance camera used here (and often used by other amateurs) seems better able to cope with flatter contrast image captures as the camera doesn't have the backlight compensation feature of more sophisticated cameras. As an aside I did try using my colour video camera with and without the backlight compensation feature and then converting to greyscale. Visible light images did improve with compensation but couldn't reveal detail in the semi-opaque areas transparent to IR.
Thoughts on initial trials
Note that the comparisons above and comments on the potential benefits for some subjects are for image captures in infra-red light cf. visible light. I also inspected the subjects above by eye (visible light, 7x eyepiece) and attempted to subjectively compare the detail seen to the image captures. The subjects available were semi-opaque giving high contrast images but not sufficiently so that the eye was not able to pick out most of the detail revealed by IR image capture. The true test of the potential of IR will be to try some subjects too opaque even for visible study but potentially transparent to IR and see if this lighting can reveal detail.
For the present though, where certain subjects need to be image captured, infra-red microscopy does seem to be useful for both revealing detail in subjects almost opaque or too contrasty to visible light capture with a bonus of increasing the depth of field. (The visible improvement in d.o.f may to some extent be subjective, even if predicted, as comparison of images of such different contrasts can be misleading.) If you have a black and white video camera, why not see if it is sensitive to infra-red and have a try at the technique if you haven't already. Let us know how you get on, especially if you try live subjects.
I was inspired to have a go at near IR microscopy myself from the notes of Tony Dutton in the Quekett Microscopical Club Bulletin (Ref. 1) and comments by Tony and colleagues by email on the IR sensitivity of CCD cameras (see earlier article). Hopefully my modest trials may also encourage others to have a go.
Valuable comments on the article by Tony Dutton are also acknowledged, although any errors are all mine!
I would particularly like to thank a colleague who generously donated the black and white surveillance camera which I used for these studies (he wishes to remain anonymous).
The slides were prepared by the author but with the expert help of Eric Marson of Northern Biological Supplies (UK) at one of his excellent Belstead House courses on slide-making.
Comments to the author sent via our contacts page quoting page url plus : ('dwalker','')">Comments to the author welcomed. Note the author has no previous experience of this area and is on a steep learning curve(!) so any comments and suggestions for further studies would be greatly appreciated.
1) 'Constructing and adapting sub-miniature CCD TV cameras to the microscope' by J A Dutton. Bulletin of the Quekett Microscopical Club, 1997, 30, p. 21-22. Also the record of a demonstration by J A Dutton at the QMC Annual Exhibition (1997) on p.41-42 of the same issue.
A colleague has kindly informed me that two articles have been published below which the author hasn't been able to access to date.
2) 'Infrared Light in the Microscope: History, Theory & Practical Aspects' by Tim Richardson. Proceedings of the Royal Microscopical Society, Vol 32, Part 4 December 1997, pp. 229-235.
3) 'Practical Infrared' by Tim Richardson. Proceedings of the Royal Microscopical Society, Vol 32, Part 4 December 1997, pp. 236-242. This informative article deals in depth with photon detection and includes a simple DIY circuit and drawing to enable a detector to be constructed. Both articles are aimed at the professional reader but there are useful graphs and definitions included.
4) 'Photography through the Microscope' by J. Delly. Kodak 1988, 9th Edition., page 15.
Notes on infra-red
- Strictly speaking 'near infra-red' is being used i.e. just beyond the visible spectrum.
- The term 'light' is often used for IR and UV radiation that extends beyond the visible spectrum (as I have done for convenience), although strictly speaking the term should be limited to the spectrum visible to the human eye.
- The IR emitting LED's described typically have a wavelength l = 850 - 950nm. (nm=nanometres).
- Visible light wavelengths extends from ca. 400 - 760nm, infra-red extends from ca. 760 - 10 000nm.
- Glass transmits IR with little loss up to 900nm, for greater wavelengths calcite or quartz optics are needed.
- CCD's or charge couple devices (i.e. the light sensor in the camera) is sensitive up to ca. 1000nm.
- I believe the older vacuum tube (e.g. Vidicon) B&W cameras are also sensitive to IR. I'd be pleased to hear of anyone's experiences with this type of camera for IR.
Return to article top.
Purchasing an LED
The electronics catalogues e.g. Maplin in the UK have a wide range of IR emitting LED's, most cost less than a pound. Not knowing what would be needed I opted for a more expensive high power LED made by 'Temic' but the cheaper ones are probably just as good as I needed a 1kohm variable resistor to drop the light intensity. The current drawn was 1-10mA for imaging using a 3.5x to 9x objective without eyepiece.
Depth of field
Various equations for depth of field in photomicrography are reported, (see ref. 4 which has a good discussion on d.o.f). Shillaber's equation for dry objectives is as follows.
d = l x square root [(1- (NAxNA)) / (NAxNA)]
l = light wavelength (use micrometres to obtain d.o.f. 'd' in micrometres).
For the LED used above l=950nm. Visible light can be taken as l=550nm (i.e. green approaching yellow in the mid-spectrum)
NA = numerical aperture of objective
The resolution R of a microscope objective with numerical aperture NA is R = l / (2 x NA)
For the 3.5x (NA 0.10) objective used, the resolving power would be reduced from ca. 2.8 micrometres (for visible light) to ca. 4.8 micrometres (for the IR emitting LED).
The microscope used was a Russian Biolam with planachromatic objectives 3.5x (NA 0.1) and 9x (NA 0.20).
The black and white CCD surveillance camera was Japanese with 380 line horizontal resolution. No camera controls apart from AGC.
Images were captured using a 'Snappy 2.0' in 'High Quality' still capture mode.
First published in June 1998 Micscape Magazine.
Please report any
Web problems or offer general comments to the Micscape Editor,
via the contact on current Micscape Index.
Micscape is the
on-line monthly magazine of the Microscopy UK web
site at Microscopy-UK