Most of the breakthroughs scientists are producing on the micro scale are consistently both fascinating and freaky. Airborne microbots, for example, will “float on the wind” in order to study climate changes and the spread of disease. Now, in another big micro-scale breakthrough, a team of researchers from Princeton University and the University of Washington have developed a camera the size of a coarse grain of salt. And all we can think about is how this thing would be perfect as an eye for bug bots.

New Atlas reported on the salt grain-sized camera, which iterates on previous, micro-scale cameras. The researchers note in their paper, published in the journal Nature Communications, that “ultra-small imagers” already in existence have the small-size factor down, but not the image quality.

This Fully Functional Camera Is the Size of a Grain of Salt_1
Princeton/Ethan Tseng, et al.

To improve on available micro cameras the researchers developed a unique type of metasurface; that is, an artificial sheet material with sub-wavelength thickness. The researchers’ metasurface consisted of 1.6 million photosensitive “posts” on its .5-millimeter-wide film surface. With each post standing as tall as an HIV virus is long.

To record an image the posts absorbed incoming photons, with each one acting as an “optical antenna” of sorts. The metasurface then translated these incoming photons—which each stand as the smallest discrete amount of electromagnetic radiation—into signals that a computer could read.

Using neural networks—that is, machine learning algorithms—the researchers were then able to decode the signals; turning otherwise excessively blurry images into relatively crisp and clear images. Below is a sample of what the researchers’ new camera can record (right) versus what previous micro-sized cameras could. The researchers say the new camera is ten times better than previous ones at filtering out errors.

Princeton/Ethan Tseng, et al.

“Although the approach to optical design is not new, this is the first system that uses a surface optical technology in the front end and neural-based processing in the back,” Joseph Mait, a former senior researcher and chief scientist at the U.S. Army Research Laboratory said in a Princeton press release. Mait, who was not involved with the research, added that the significance of the this micro-scale camera comes from its completing the “Herculean task” of designing the metasurface’s post layout and pairing it properly with the neural networks.

Felix Heide, the study’s senior author, says he and his colleagues are now working on more computational abilities for the camera. Beyond optimizing image quality, they also want to add object detection. As well as sensors relevant to medicine and robotics. Plus, engineers could also apparently use this tech to turn almost any surface into a camera. Which doesn’t sound terrifying at all!

Feature image: Princeton/Ethan Tseng, et al.