Processing IR Photographs
Basic PP
Just by setting the right wite balance and making some corrections to the contrast by adjusting the tone curve or levels settings, it is possible to get good results in IR photography. With the right settings, photos can be viewed and used unprocessed from the camera. Using a RAW format does ofcourse give the photographer more freedom in post processing. Unprocessed photographs have dependent on the type of IR filter that is used (665nm, 720nm or 850nm) a limited color separation and may have different tones depending on the white balance stting. Sepia, red or orange tone are most common, but you can use any color you like.
Black & White
The most used and very easy final IR photo is converted to black & white. B&W photos are possible direcly from the camera with most models, but often it's neccessary to adjust the contrast using levels or curve settings. The use of an increased local contrast (clarity) setting may enhance and strengthen the image. If a IR filter with high cutoff is used (> 800nm) then the photos will be in monotone without any editing.
Blue Sky effect
The sky in IR photos is usually dark and is brightest at lower wavelengths, so usually in the red band of pictures. In most unedited photos the sky looks pinkish. By swapping the red and blue channels of an image, the sky get a blue color and the rest of the picture will turn pinkish or yellow. To bring out the blue sky, it is posssible to shift the hue a bit and selectively desaturate any colour but blue. This blue sky effect can be very pleasing and is just one example of how infrared pictures can be edited to get great results.
Bluesky workflow Photoshop »
Bluesky workflow Photoshop 2nd example »
Color Infrared
Color Infrared photos are known from old IR-film and also from aerial photography. Infrared and visible color information is combined in these pictures to create a false color composite. Spectral information is commonly mixed IR as Red, Red as Green and Green as Blue (RGB=IR-R-G). This combination is referred to as ColorInfraRed (CIR). With a converted digital camera it normally not possible to record CIR images, because only IR information is recorded. It is possible to create CIR pictures, but the process is rather laborious. You need two *exact* the same pictures, one taken using an IR camera and one with a normal camera. If you use a full spectrum camera, two frames using different filters (hot-mirror and IR-pass) can be used, this would be ideal. I own both a converted and a nomal 30D, and use both with the same lens. In PP the color information from both images are combined.

In short, the process to create a CIR image is separating the channels of both images to B&W layers. These layers are then mixed back to the color channels using mixture settings. This way of working allows to play freely with the different color channels and to create lots more different color composites. One big problem with combining two frames is that it is virtually impossible to create pixel perfect matching frames. This causes the composite picture the look unsharp, or cause heavy color fringing. Partly this can be solved by applying a gaussian blur to the IR layer before combining and adding a sharpning layer (highpass filter, overlay (or soft\hard light) blending) based on the color image. You can use the photoshop auto align function (edit > Auto-Align layers) to align both pictures as good as possible. If you look closely to the image, there is still quite a bit of fringing visible in the tree at the foreground, because it was quite windy at the day the pictures were taken and the tree was moving a lot.
This Album contains examples of what is possible just using different layer blending: