HyperCam would let you see the unseen

Because regular cameras just process visible light, the images that they produce look like what we see with our own eyes. By contrast, hyperspectral cameras process additional wavelengths, showing us things that we wouldn't otherwise be able to see. Unfortunately, they also tend to be big, expensive, and thus limited to scientific or industrial applications. That could be about to change, however, as scientists from the University of Washington and Microsoft Research are creating a compact, inexpensive consumer hyperspectral camera. It may even find its way into your smartphone.

Known as the HyperCam, the device both emits and then images 17 different wavelengths of light within the visible and near-infrared part of the electromagnetic spectrum. By contrast, a conventional camera just works with the red, green and blue (RGB) bands of visible light.

The HyperCam then compares each of those images to one regular RGB photo of the same subject, to see which ones differ most from what the human eye would see. Those select images are combined into one composite photo, which is presented to the user.

So, what sort of things might it show them? Examples include veins beneath the skin, or the ripeness of fruits beneath their skins. In lab tests, the HyperCam was able to accurately identify 25 different people based on their vein patterns, with 99 percent accuracy. It was also able to gauge the ripeness of 10 types of fruit with 94 percent accuracy – by contrast, a conventional RGB camera had a success rate of only 62 percent.

In its current form the technology would cost about US$800 as a standalone camera, although it might only cost about $50 to add to a smartphone in the manufacturing process. Before that can happen, however, the camera has to be further miniaturized and adapted to work better in bright light.

Researchers at Tel Aviv University are working on a similar project, although one that's perhaps even more ambitious. They're developing a smartphone hyperspectral camera that could identify the unique electromagnetic "fingerprints" of various substances, meaning that you could see what something was made of just by pointing your phone's camera at it.

Source: University of Washington