Extended wavelength response

Traditional vision applications are limited to the recording and capture of images in the visible and near infrared (NIR) spectrum. This limitation is caused by silicon, the material used in the majority of sensors.

In silicon, photons in the visible wavelength have enough energy to release the electrons and penetrate material such as silicium. Shortwave light on the other hand is richer in energy, but sensors with quartz glass or removable cover glass are necessary for the shortwave light to reach the photosensitive silicon layer.

UV imaging

One method of extending a sensor's response towards shorter wavelengths is to use a fluorescent coating. One type is known as a 'lumogen coating', which fluoresces when struck by a UV photon. This fluorescence is detected by the pixels in the adjacent light sensitive layer, as the wavelengths emitted by the coating are close to the peak responsivity of the sensor.

The lumogen coating process needs to be highly accurate as the thickness of the coating is critical if good results are to be obtained. If the coating is too thick, the fluorescent emission is scattered and absorbed within the coating itself and if the coating is too thin, photons will not be detected. In order to ensure excellent UV image quality, this type of sensor must always be used with UV transmissive lenses fitted with visible cut and UV pass filters.

In physics infrared radiation describes electromagnetic waves in the spectral area between visible light and longwave terahertz radiation. Infared imaging is a method to convert radiant energy in the infrared wavelengths into a detectable or measurable form. The spectral range of 780 nm to 1 mm is called infrared.

The spectral range is divided into near infrared (NIR, 750 nm - 1400 nm), short wavelength infrared (SWIR, 1.4 µm-3 µm), mid wavelength infrared (MWIR, 3 µm - 8 µm), long wavelength infrared (LWIR, 8 µm-15 µm) and far infrared (FIR, 15 µm - 1000 µm). The human eye reacts to visible light in the wavelength of about 390 to 750 nm. The NIR-spectrum is adjacent to this visible spectrum. Silicon based CCD and CMOS sensors are used where possible for NIR applications due to their lower cost compared with other sensor materials.

As silicon based sensors only show a limited efficiency in converting photons into electrons in the wavelength >1000 nm, manufacturers try to use special circuits or sensor coatings to get the remaining sensitivity beyond 1000 nm.

SWIR or short wavelength infrared is a very interesting spectral range. Images created by SWIR cameras are almost similar to those taken by a monochrome CCD camera, but the detector material is sensitive in a wavelength band where water vapour causes maximum absorption and silicon is transparent. Infrared imaging can be used to detect features that are not apparent in visible wavelengths for applications such as inspection of silicon wafers or solar cells, sorting of fruits or vegetables, control of plant growth, recognition of safety features, etc.

The highly sensitive NIR-SWIR cameras mainly use InGaAs (Indium Gallium Arsenide) detectors, offering a high quantum efficiency at 900 - 1700 nm and are normally combined with CMOS readout electronics. Most InGaAs cameras feature additional electronics to execute a background and defect pixel correction to achieve perfect image quality for the subsequent image analysis.

Shortwave infrared cameras can be cooled or not, depending on the application requirements. Uncooled sensors are typically used for imaging up to 1.7 µm, cooled sensors offer improved reduced noise on higher speed sensors, but are generally a more expensive solution due to the thermoelectric cooling.

Following the SWIR-range is the mid wavelength infrared (MWIR, 3 µm - 8 µm). Cameras for this range are often called thermal cameras and are most commonly based on Mercury Cadmium Telluride (MCT) or Indium Antimonide (InSb) sensors. Again using the photoelectric effect, MCT based sensors offer sensitivity at around 3 - 8 µm, where InGaAs is insensitive. As with shortwave infrared, midwave infrared imaging can be used for detecting reflected or emitted infrared. This sensor type requires cooling to ensure that the detected signal is not saturated by inherent dark current in the camera.

Longwave infrared (LWIR): Short and mediumwave infrared imaging sensors use the photoelectric effect to provide a signal and produce an image. Longwave infrared detectors (or micro bolometers) detect heat radiation by measuring changes in capacitance or resistance within the structure of the pixels. Commonly based on Amorphous Silicon (ASi) or Vanadium Oxide (VO), these sensors can detect wavelengths of 8 to 14 µm where the measured values relate to the temperature of an object.

This enables LWIR cameras to work where there is no infrared source and with objects at much lower temperatures compared to short and midwave infrared. As this technology is not dependent on an external IR source for imaging, it can be used for security and surveillance applications in all lighting conditions.

Existing longwave IR detectors are relatively low resolution with pixel resolutions typically of 320 x 240 or 640 x 480, but next generation sensors are now coming to market which offer megapixel resolutions.

Optics

SWIR and MWIR cameras use standard glass lenses as glass has good transmittance at these wavelengths. For LWIR, an alternative lens type is required with sapphire, crystalline silicon and germanium being the most common materials used. These are expensive to source and process and in most cases are of fixed focal lengths, although zoom lenses are becoming more common.

Thermography - temperature measurement

Short, medium and longwave infrared can all be used for thermography. LWIR cameras are most commonly used for this, but SWIR and MWIR sensors can be used for materials at higher temperatures. For example when heated to high temperatures, steel emits wavelengths in the visible spectrum - the shorter the wavelength, the higher the temperature. Thermography is a large subject with temperature measurements dependent not just on emissivity of an object, but also on reflectivity and transmission of IR by an object. The camera also needs to be calibrated using a black body (a perfect thermal energy emitter) so that sensor readings can be accurately related to specific temperatures, down to a single degree Centigrade.

The images show a scene under visible light (top), SWIR (left) and LWIR (right). The car on the right has been driven recently, as can be seen from the hot tyres and bonnet in the LWIR image. The cars centre and left have been stationary for some time, although the car on the left has a warm bonnet having been heated by sunlight. The SWIR image shows the glass as non-transparent, while the visible image shows none of this information.

Scene under visible light

Scene under SWIR

Scene under LWIR

The images that follow show surface mounted LEDs both with and without cooling. The images use pseudo-colour to indicate the change in temperature and clearly demonstrate the value of this technique. There are important security and medical uses for thermal imaging in the detection of potentially harmful fever among groups of people and these can be deployed at any transport hub. Other uses for thermal imaging include carpool lane checking (detecting the number of passengers in a vehicle) and various recycling applications.