Images in this spectral range cannot be obtained with ordinary silicon-based CCD and CMOS detectors. Quants of wavelengths above 1 μm are unable to induce electrons in silicon-based detectors, since quantum efficiency rapidly drops to zero in SWIR range.
InGaAs-based detectors tend to be used for registering radiation in SWIR-range. Well, a couple of years ago we got our hands on such a commercial detector. Its resolution is modest: 320x256. Its spectral response curve is presented on a figure below.
It seemed like no unexpected challenges would be met, and the process of developing this camera would be no different from developing a regular visible one, but we were mistaken. The main obstacles came from dark current being pretty high and characteristics of separate elements varying quite a lot. Look at the histogram below:
During the time period of 16 ms, potential well of some elements of this detector quickly fills up to 3-5%, and when the frame rate is 25 fps (40 ms) it’s 8-12%. Full well capacity of 6 million electrons for an element of such a detector translates to 600 000 e- of dark current per element, and a noise level of about 800 e- per pixel. Is it too much or too little? It’s fine for registering well lit objects, but when it comes to a highly sensitive camera that is capable of registering self-radiation of objects with temperatures below 100°C (like in the first video) — 800 e- noise level is extremely high.
The graph below demonstrates blackbody radiation. It’s obvious that for objects with 300-400 K, temperature radiation in the range of 1-2 μm is very feeble.
The second peculiarity is wide spread of every element’s characteristics. The development process took several years, and emphasis had been placed on developing low-noise analog circuitry and on approximation of elements’ characteristics as a function of temperature. Once again, the detector was a commercial one, we weren’t able to cool it down to lower dark current levels directly. However, we managed to partially stabilize its characteristics by thermostating the detector.
We have mentioned this camera in some of our previous articles, and compared its performance with that of visible range detectors and the image intensifier «ЭОП 3»:
«How different cameras and detectors see at night»
and demonstrated this camera’s ability to observe stars during daytime:
«Observing stars during the day or daylight astronomy»
In this piece we want to add on to what was published earlier and demonstrate other unique features of SWIR-cameras.
The most frequent question is — «how do cameras see through the fog?». It’s not easy to come across exemplary fog, so sorry for not the most demonstrative video ever. Footage on the left was filmed with a visible range Panasonic GM1 camera to show how the scene looked to normal human eyes.
Here’s the video filmed with the SWIR camera itself:
Available links to original videos:
«VS320 original footage»
«Panasonic GM1 original footage»
Just in case, let us warn you that fogs can differ from each other. In some cases it is impossible to see through one regardless of spectral range. The result is heavily influenced by dispersion of water droplets.
The sensitivity of the camera is in fact illustrated by the first video in this article. It captures an ordinary mug of freshly brewed coffee. At the start of the video, we observe thermal radiation of the object, and after the light was turned on — reflected radiation. As of today, VS320 camera is the only one capable of capturing thermal radiation of objects with the temperature below 100°C. We presented this video on exhibitions several times and were always met with skepticism.
For reference: thermal radiation of red-hot metal 500°C and higher is visible to a naked eye and an ordinary color camera, soldering tip heated up to 400°C can be perceived by a black-and-white CCD sensor. VS320 SWIR camera can «see» objects starting with temperature as low as 50-60°C.
More objective measurements based on the blackbody model are presented below. The noise level of detector’s elements aligns with blackbody model’s signal at a temperature of about 50°C.
A link to the original video (be warned, it’s not compressed, so its size is quite large)
«VS320 blackbody video»
One peculiar moment we have run into was a special anti-counterfeiting measure — likely luminescent ink put onto banknotes in a specific way. Photos of the bills taken at normal lighting are no different from ones posted on the Central Bank of Russia’s website. For example, this is a 500-ruble bill under normal lighting:
but when the source of lighting is a luminescent lamp, certain markers become visible. Banknotes of different denominations have those markers at different spots, so that could be used for automated sorting.
Central Bank of Russia’s website does not mention that fact.
Apparently, these varied markers have been dropped for the new release, so now the marker is uniform for all types of banknotes: round shape with «₽» in the middle.
Here are all these banknotes together:
It’s worth noting that the night sky looks very bright when shot in SWIR range. It helps SWIR cameras successfully compete with other night-vision equipment and opens up such applications for them as object detection in that «bright» night sky.
«VS320. Night sky in SWIR range»
It’s the other way around during the day — compared to the brightness of the sky as seen in visible range, it seems way darker. Here’s a shot taken on a very sunny day.
This property is useful when it comes to observing celestial objects during the day, one particular occasion of which has been described in the article by the name of «Observing stars during the day or daylight astronomy».
The most important feature of a SWIR camera (along with enhanced visibility in fog) is enhanced visibility in haze. Here’s a comparison of shots taken in different spectral ranges.
And here’s a video of a cable-stayed bridge at a distance of 9-10 km shot with a SWIR camera.
Smolny filmed at a distance of 9 km (in the middle of the video local contrast enhancement function (analogous to HDR/DDE) is turned on).
In summary, SWIR cameras can be used for the following purposes:
- Visibility enhancement in fog, haze, or smog
- Alternative for night-vision equipment
- Detection of objects in the sky during daytime
- Development of multispectral cameras (when you need to find a warm object, concealed for the visible range devices)
- Specialized industrial applications
- Detection of camouflaged objects (for cases when some parts of them seem lighter and some seem darker in SWIR range)
