|Digital Reflected UV and NIR Photography|
A while back, I became aware that most digital cameras could effectively image into the UV (below 400 nm), and into the NIR (above 700 nm) range of the electromagnetic spectrum (EM). Needless to say, this intrigued me to no end. I got most of my information from Bjørn Rørslett who has an incredible wealth of information on the subject.
As you may know, I do this sort of thing for a living. I am a "Remote Sensing Specialist" for a living, which basically means that I spend lots of time looking at digital satellite imagery well that extends well beyond the range of "normal" human vision (400 nm to 700 nm, or blue to red as you normally see in a prism). While most satellites that I deal with see well above the visible range (up to 2500 nm and sometimes beyond), most people in my business (and there are not that many, I think I know most of them personally) effectively treat the UV as "noise". This has to with the way the atmosphere works with scattering, and how "soft" the blue region is to begin with. There are many things to see beyond normal human vision.
I own a few digital SLR cameras (a couple of Fuji Finepix S2 Pro's, and a couple of Fuji point-and-shoot cameras, I like Fuji because of their CCD produces almost no noise in low light levels.) I decided that I "needed" to try this out (much the same way that I decided I needed to levitate matter). Well, about $100 later, I owned a 77 mm B+W 403 UV filter, which interestingly enough, seems to have transmission windows in the EM spectrum at about 350 nm (UV) and 750 nm (NIR), while completely blocking out any light transmission between about 400 nm and 700 nm. This effectively filters out almost all of the visible range of light (for normal humans), while allowing it to be used for both UV (well, UVA at least, which goes from about 400nm to 320 nm. For reference, UVB goes from 320 nm to 290 nm, and UVC goes from 290 nm to 200 nm luckily for us, is completely blocked by the earth's ozone layer) and NIR photography. Most cameras these days have an NIR filter over the CCD anyway because by default, the CCD is very sensitive to NIR radiation. And most lenses have a coating that diminishes UV interference. Most prosumers go beyond that and add at least an external UV filter to further improve filtering performance (and to protect the front element of their expensive lens). You may think that these things working together will render your excursion into UV/NIR for naught, well you may find that you are wrong. If you truly want good images (and have money to spare), there are both special lenses designed for UV photography, and any number of websites that give tutorials about how to remove the UV reduction coatings off of your expensive lenses (this will obviously degrade the lenses performance for "visible" photography, and shouldn't be performed on any lens that you intend on ever using for "visible" photography again). However, just because a signal is reduced, that doesn't mean it isn't strong enough to be recorded. Checking with the spectral transmission windows (which is the sort of thing I do for a living), I see that my Sigma 70-200mm f2.8 lens does pass some UV range information to the CCD, although not very much. Its still possible to capture some of it, but we are talking shutter speeds in the range of 20 seconds or longer. Believe it or not, its almost impossible to get a plant in natural surroundings to stay stationary for that long. You can still get some images however.
While I was getting images, I wanted better images. I can't seem to locate a true UV Nikon lens (made with quartz lenses, Nikon stopped making them, and the ones that are out there are very scarce, and expensive). Of course, It wouldn't be in the spirit of this website if I didn't purchase a lens and remove its multicoating. Anything worth doing, is worth doing right. The next best thing seems to be a very simple lens, with the least number of elements and groups you can find. I can tell you that the method of using a polishing compound seems to be the only way to go. I have attempted to use Ferric Chloride to remove the coating, with no visible success. It could be that I didn't wait long enough, or that I didn't have the solution heated, but I was wary about letting the solution on the lens long enough for capillary action to draw it into the interior of the lens. That would most definitely be a "bad thing". I acquired a Nikon Series E 35mm lens for testing purposes. And after the failed Ferric Chloride attempt (which I still believe should remove the multicoating, which I believe is a Magnesium Flouride compound, but I didn't want to risk it). I purchased some Cerium Oxide powder, mixed it to a paint-like consistency using normal water, and polished my lens using a dremel tool and a soft buffing pad (I trimmed the normal dremel attachment to let me apply the pad face first to the lens in order to get a orbital polishing pattern). You can do it by hand if you have the patience. If you intend on doing it with some mechanical help, I wouldn't go above about 500 rpm's (the slower the better). And be sure to check the temperature fairly often, and make sure it doesn't go dry. You definitely don't want to get the glass too hot.
UV is very "soft", and I usually try to shoot with at least an f4.8 aperture (and most often much higher). With my Sigma 70-200 mm f2.8 lens (with its UV coating intact), its UV external filter removed, and the B+W 403 filter installed, that usually translates to shutter speeds in the 10's of seconds as opposed to fractions of a second. Using my Nikon Series E (with its UV coating removed by mechanical means) the shutter speeds come up to under the 10's of seconds. Either way, bring a tripod if your going to try this yourself.
NIR is much easier (the lens isn't fighting against you, as it comes from the manufacturer), and digital camera's are inherently sensitive to IR. Aperture is still important for crisp images, but you can get some decent shutter speeds on sunny days, and (depending on your lens), may be able to hand-hold for a NIR shot.
On the Finepix S2 Pro, the UV seems to be output mostly in the blue channel, and the NIR seems to fall mostly in the red channel. The green channel also picks up some NIR, but it seems to be a bleed over from the red channel (and its much more noisy than the image in the red channel). This makes it very easy to get my images out. And thanks to Fuji firmly refusing to give me spectral response curves for the CCD, I don't really have an exact range for what I am imaging yet (you do detect the sarcasm in that "thanks" don't you?).
Note: The UV and NIR images shown below are taken with the B+W 403 filter in place, which, I can confirm, completely blocks out the visible (look through the viewfinder with your lens cap on to see exactly what you will see with the filter in place)
My sample normal image is here (No B+W 403 filter), ISO 100, f2.8, 1/60th of a second. This is a Common Dandelion (Taraxacum officinale, Weber) taken in the visible range (as a normal photograph), but all images below are resampled much smaller for web viewing. These image are taken in full sunlight, so there will be some IR influence in the UV image.
The UV image is here (showing the typical "bullseye" pattern the species displays in the UV, bright flower head showing against darker background for easy attraction of insects), ISO 100, f4.8, 15 second exposure time, taken about 2 seconds after the visible image above. This was the image recorded in the blue channel.
The NIR image is here (showing typically bright leafy vegetation, but there are a few details on the flower that aren't readily seen in the visible image), ISO 100, f4.8, 15 second exposure time, taken at the same time as the UV image above. This was the image recorded in the red channel.
I picked up a florescent black light recently because it is capable of putting out UV, without producing any IR radiation. These images are taken only under UV light, so there should be no visible/IR contamination of the UV image.
Here is a small white flowering plant I came across. It claims to be a common Periwinkle (Vinca minor Alba), but I am not entirely sure I believe it. It does appear to have a response in the UV, with a definite darkening in the center of the flowers that I can't see either under visible light, or black light.
Here is its visible image (taken with the lights on, and with a flash).
And here is the UV image taken in black light with my newly modified Nikon Series E 35mm lens with its UV coating removed. You can see in this photo, and the photo above, that normal lenses have trouble focusing into the UV. Some correction needs to be made after you put on the UV filter. This can be compensated for manually, or you can mark the lens. The below image was not compensated for to display this property.
Here is the same plant taken with my EL-Nikkor 63mm f2.8 lens. Since this is an enlarger lens, Nikon states that it is basically focus corrected down into the UV because B/W photographic papers are very sensitive in this range. Note: You cannot simply attach an EL-Nikkor lens to a Nikon F-Mount. This particular Nikkor lens has a 39mm screw mount on it. This is a very common mount worldwide, so you can find some adapters to go from Nikon to M39. You will also need to rig some sort of focusing assembly because the EL-Nikkors are designed to be bellows focused. This particular image is taken using a couple of Russian made converter rings, and a Russian bellows unit. Notice that the focus is dead on with the visible image. A little further modifications to my focusing range with this lens, and I believe it will become my UV workhorse.
Here is the visible image:
And the same image with the UV filter installed under only black light:
This site was last updated 11/16/05