Sea Creatures Inspire Artificial Eyes with Improved Night Vision

Photo: Stephanie Precourt

Amateur photographers attempting to capture the moon with their cell-phone cameras are familiar with the limitations of low-light imaging. The long exposure time required for nighttime shots causes minor shakes to add up into majorly blurry images. Yet, fuzzy photos aren’t merely an annoyance—bomb-diffusing robots, laparoscopic surgeons, and planet-seeking telescopes all need to resolve fine details through almost utter darkness.

“These days, we rely more and more on visual information. Any technology that can improve or enhance image taking has great potential,” says Hongrui Jiang, the Lynn H. Matthias Professor in Engineering & Vilas Distinguished Achievement Professor of Electrical and Computer Engineering at UW-Madison and the corresponding author on the study.

Most attempts to improve night-vision tweak the “retinas” of artificial eyes so that they respond more strongly to incoming packets of light. Usually technicians try to change the materials or electronics on the imager—such as a camera’s charge coupled device (CCD).

However, rather than interfering with efforts to boost sensitivity at the back-end, Jiang’s group set out to increase intensity of incoming light through the front-end optics. They found inspiration for the strategy from two aquatic animals that evolved different strategies to survive and see under murky waters.

Arrays of minuscule crystal cups arranged on the artificial eye’s surface collect and concentrate incoming light.

Photo: Hongrui Jiang

Elephant-nosed fishes resemble river-dwelling Cyrano de Bergerac impersonators. Looking between their prominent proboscises reveals two strikingly unusual eyes, with retinas composed of thousands of tiny crystal cups instead of the smooth surfaces common to most animals. These miniature vessels collect and intensify red light, which helps the fish discern its predators.

“We were thinking: Why don’t we apply this idea? Can we enhance the intensity to concentrate the light?” says Jiang.

The group emulated the fish’s crystal cups by engineering thousands of miniscule parabolic mirrors, each as tall as a grain of blue spruce pollen. Jiang’s team then shaped arrays of the light-collecting structures across the surface of uniform hemispherical dome. The arrangement, inspired by the superposition compound eyes of lobsters, concentrates incoming light to individual spots on the imager, further increasing intensity.

“We showed four-fold improvement in sensitivity—that makes the difference between a totally dark image you can’t see and an actually meaningful image.” says Jiang.

In this case, the devices discerned a picture of Bucky Badger, UW-Madison’s mascot, through pitch-black darkness. However, the device could easily be incorporated into existing systems to visualize a variety of vistas under low light.

“It’s independent of the imaging technology.” says Jiang. “We’re not trying to compromise among different factors. Any type of imager can use this.”

Although superposition compound eyes are exquisitely sensitive, they typically suffer from less sharp vision. Increased intensity costs clarity when broad amounts of light compress down to single pixels. To recover lost resolution, Jiang’s group captured numerous raw images then applied a super-resolution-processing algorithm to produce crisp, clear pictures.

Currently the group is working to refine the manufacturing process to further increase the sensitivity of the devices. With perfect precision, Jiang predicts that the artificial eyes could improve by at least an order of magnitude.

“It has always been very hard to make artificial superposition compound eyes because the curvature and alignment need to be absolutely perfect.” says Jiang. “Even the slightest misalignment can throw off the entire system.”

Despite the technical obstacles, Heiwei Liu, a postdoctoral scholar in Jiang’s research group and first author on the paper, rose to the challenge of fabricating the devices. Yinggang Huang processed super-resolution images. The National Institutes of Health, National Science Foundation and the University of Wisconsin-Madison provided support for the research. A link to the manuscript may be found at PNAS.