5
\$\begingroup\$

A common way to make an infrared camera is to remove a commodity camera's IR-blocking filter (and add another filter to block the visible spectrum).

Is there a similarly cheap way to convert an RGB camera to sense three different bands of infrared (three colors), i.e., multispectral? Searching online for 3-channel infrared hits the red herring of remote-control toys.

\$\endgroup\$
3
  • \$\begingroup\$ The R, G, and B sensors (or built in filters) are going to have a response centered around these colors... This means that removing the IR blocking filter would probably only impact the R (which probably doesn't go much into IR to start with)... You didn't say what kind of camera you're looking to convert but I assume it's a color sensor camera? \$\endgroup\$
    – Guy Sirton
    Commented Dec 11, 2013 at 6:44
  • 2
    \$\begingroup\$ "When you wish to request a question moved to another site, all you need to do is flag for moderation attention and ask for the question to be migrated." -- Yannis 2013 \$\endgroup\$
    – David Cary
    Commented Dec 12, 2013 at 15:50
  • \$\begingroup\$ There is some fascinating info on the Basics of Infrared Photography page, so it looks like what you want to do might be possible, but I don't have the time to turn this into an answer at the moment. \$\endgroup\$
    – Mark Booth
    Commented Dec 13, 2013 at 9:25

2 Answers 2

4
\$\begingroup\$

Yes. In fact, every thus-modified camera senses three different bands of near-infrared. (This answer is an expansion of Mark Booth's comment.)

In Basics of Infrared Photography, the first diagram's shaded areas show the sensitivity to wavelength of the R,G,B sensors of an unmodified camera. The third diagram (copied in this answer) shows the same for a camera whose IR blocker has been removed. Those three shaded areas are identical past 900 nm, but from 700 to 900 they differ enough to be fairly considered as three different "colors." The difference can be augmented by boosting saturation in an image editor.

(Note that the blue channel now represents the longest wavelengths, red the shortest.)

third figure of external webpage

\$\endgroup\$
1
  • 1
    \$\begingroup\$ A common technique to emulate Kodak Ektachrome Infrared is to shoot a camera modified for full-spectrum use through a #12 blue-blocking filter. That gives you infrared in the blue channel, IR plus red in the red channel, and IR plus green in the green channel. Then, subtract the blue from each of the others, and move green to blue, red to green, and blue (IR) to red channels. flickr.com/groups/55027594@N00/discuss/72157601045395608 \$\endgroup\$ Commented Dec 12, 2022 at 16:44
2
\$\begingroup\$

Some photographers put a few different infrared bandpass filters (a), (b), (c), also called infrared transmitting filters, in front of the camera, one at a time, to photograph a scene in a few different bands of infrared light.

Some astronomers attach a prism to a telescope (a) (b), smearing out each pinpoint of light from a star into a spectrum with many recognizable bands of infrared colors. A diffraction grating might work better than a prism. (multispectral imaging, hyperspectral imaging, spectral signature, etc.).

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Prisms and diffraction gratings are only useful in cases where you want to separate a point source of light into its component parts -- the grating converts a 0-dimensional point into a 1-dimensional line, while the prism converts a 1-dimensional line into a 2-dimensional rainbow pattern. In either case, you won't get something that resembles the original image. \$\endgroup\$
    – Ian
    Commented Dec 12, 2013 at 15:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.