0
\$\begingroup\$

I was developing a digital system to process raw images stemming from an image sensor but I would like to deal with "universal" (as possible) spectrum image sensors with no filter. I know that for long wavelengths (microwaves), the mechanism is different with respect to the detection of infrared-visible-UV-X-ray light since the last ones should use the photoelectric effect to capture photons, but between infrared-visible-UV-X ray that use photoelectric effect, the material of the sensor changes because the photon has different energy.

  1. Does a new technology or a new kind of sensor exist that is able to detect "at least" radiation from Far Infrared to X-Ray wavelengths?

  2. CMOS and CCD sensors have both roughly the same spectrum sensitivity (from NIR to part of Near UV)? Or exists CMOS and CCD sensors, made with different materials, able to detect also at FIR and X-ray wavelength?

  3. In the market there exist CMOS or CCD image sensors with no filters? In this case, the resulting image is monochromatic?

  4. A Bayer filter that gives RGB color to a detected image, since color is related to the visible spectrum only, this filter has a role on the building of a IR or UV (or X-ray) digital image? If I have a row image got by a sensor with no filter, thus an image with spectrum from Infrared to X-Ray (or UV), is it possible to make a digital filter? For example, starting from this row image with its entire spectrum with no filters, building digitally the IR only image, or UV only image or visible only image?

\$\endgroup\$
  • \$\begingroup\$ The sensors don't really care what frequency the photon had, so no, you can't use digital filters to get different images. If that would be possible it would have been implemented a long time ago because you could reduce the amount of pixels on a RGB sensor by a factor of 3. \$\endgroup\$ – Arsenal Mar 21 '17 at 16:27
  • \$\begingroup\$ Good 'ol photographic (silver halide) panchromatic film had trouble at the infrared end, but was fairly monochromatic. \$\endgroup\$ – glen_geek Mar 21 '17 at 19:51
0
\$\begingroup\$

The closest thing to full-spectrum sensor comes from the field of photoacoustic spectroscopy. With a modulation applied (a chopping wheel, basically), almost any kind of EM radiation that interacts with matter will make a small thermal expansion signature, and the resulting vibration can be detected with a microphone (or seismometer, or piezoelectric transducer... depending on the chopping frequency). For an image to be created thus, a lot of processing might be required, but it should be possible to implement moving-blockage scanning, even with X-rays.

There is an older technology that also works: the bolometer. A thermocouple can be made VERY small, and energy turning to heat can be detected rapidly. Thermistors, too, can be printed on a grid for imaging. Thermal sensors were the earliest IR sensors, after all IR discovered.

\$\endgroup\$
  • \$\begingroup\$ Hi @Whit3rd. Thank you for the answer. Is there a well and full explained reference about the first solution that you cited based on photoacoustic spectroscopy? \$\endgroup\$ – D3m Apr 5 '17 at 9:17
  • \$\begingroup\$ Sorry for double comment. I read about photoacoustic imaging but I think it works if the sensor is near the source of the resulting vibration. If I am observing a far object emitting IR, I don't think I can detect it by photoacoustic imaging due to the distance, or is it possible? I read also that according to the intensity of the radiation, the pressure wave changes. How can I distinguish different radiations? For example what does it change of ultrasound wave if the material is hit by an X-rays radiation or microwave radiation? Does ultrasound frequency change? \$\endgroup\$ – D3m Apr 5 '17 at 11:19
  • \$\begingroup\$ To distinguish different radiations AND image a visual field, is possible with time-varying filters, or multiple images with different gratings applied. All detectors have signal-to-noise issues, but modulation (like, light chopping) accomplishes some noise cancellation. Light intensity from 'a far object emitting IR' has to be compared to noise sources to determine detectability. \$\endgroup\$ – Whit3rd Apr 6 '17 at 5:30
2
\$\begingroup\$

A partial answer:

  1. Yes, there are one-chip CCD cameras that yield a monochromatic image where all wavelengths are present. Color separation is achieved by (acousto-otpical) band pass filters; often used in microscopy. Foveon sensors attempt(ed) to detect multiple wavelengths at different depths of one sensor.

  2. Bayer mask cameras are sensitive to IR light, that is why they have IR blocking filters. You can make a night vision camera by knocking the IR filter from a web cam (and get a blurry image during the day). But no, you cannot discriminate color signals from a monochromatic camera.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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