New image sensor captures both colour and near-infrared at same time
Wed 8 Jun 2016
Researchers have developed a new single imaging sensor which is able to capture and display high-quality colour (RGB) and near-infrared (NIR) images simultaneously.
The team of scientists from Tokyo Institute of Technology and Japanese optics specialist Olympus claims that the device is the first of its kind that is capable of acquiring both forms of information at the same time.
Today’s single-chip sensors, used in digital cameras, smartphones and scanners, capture colour images through a standard Bayer colour filter array (CFA) – an array of R, G and B (RGB) filters laid over the image sensor (see diagram below). Image processing algorithms are then performed to acquire a full-colour image.
Now the research group, led by Professor Masatoshi Okutomi, has advanced the technology, pairing both RGB with NIR by modifying the traditional sensor layout. The prototype presents a new CFA format which contains both RGB and NIR pixels. The team integrates this filter into a new image processing system which can execute algorithms, such as de-mosaicing [PDF] and colour correction, in real-time. The system can capture both sets of spectral information at 60 frames per second (fps).
The combined sensor could be a welcome addition across a range of applications where both visual and NIR information is needed, including food inspection and fruit and vegetable sorting. Tokyo Tech further suggested that the technology could be of particular interest in fields such as remote sensing, surveillance, robotics, agriculture, and medical imaging, where NIR data is also useful resource.
It is interesting also to consider the opportunities for biometrics, with the research community considering tracing patterns of human blood vessels below the skin’s surface with infrared cameras. These patterns are unique to each individual and could provide a digital fingerprint, and in this case would offer an additional layer of verification after a colour facial scan.