Miniaturizing augmented reality displays into glasses to expand their uses

Augmented reality (AR) takes digital images and overlays them on real-world scenes. But AR isn’t just a way to play video games; it could transform surgery and self-driving cars. To make the technology easier to integrate into common personal devices, researchers report in ACS Photonics how they combine two optical technologies into a single high-resolution AR display. In their glasses prototype, the researchers used computer algorithms to remove distortions to improve the quality of the images.

AR systems, like those found in bulky goggles or car head-up displays, require portable optics, but when typical four-lens AR systems are scaled down to the size of glasses or smaller, they typically result in poorer computer-generated image quality and a smaller field of view.

Ma Youguang and his colleagues may have found a solution to condense the technology: They combined two optical technologies, metasurfaces and refractive lenses, with a microLED screen (containing an array of tiny green LEDs to project the image), to create a compact, single-lens hybrid AR design.

The display’s metasurface is an extremely thin, lightweight silicon nitride film etched with a pattern that shapes and focuses light from green microLEDs. Black and green images are then formed onto a synthetic polymer refractive lens, refining the image by sharpening and reducing optical aberrations.

The final image is projected from the system and superimposed onto an object or screen. To further increase the resolution of the projected image, Ma and his team used computer algorithms to identify small imperfections in the optical system and correct them before the light is emitted from the microLEDs.

The researchers integrated the hybrid AR display into glasses and tested the performance of their prototype with computational image correction. Images projected from the single-lens hybrid system had less than 2% distortion across a 30-degree field of view, which is comparable in image quality to current commercial AR platforms with four lenses. The researchers found that their computational pre-processing algorithm improved the reprojected AR image of the red panda.

The reprojected red panda was 74.3% structurally similar to the original image, a 4% improvement from the uncorrected projection of the image. The researchers say that with further development, the platform could be extended from green to full color, enabling a new generation of mainstream AR glasses.