Researchers at Princeton and Washington University have developed an ultra-compact camera the size of a grain of salt. The device is capable of producing clear and colourful images that are not significantly different in quality from images taken with a standard camera equipped with a lens 500,000 times larger.
Conventional cameras use a series of curved glass or plastic lenses to focus light rays, but the new system is based on a different technology called metasurface. This 0.5 x 0.5 mm structure is filled with 1.6 million cylindrical columns. Each one has a precise design, a unique geometry and acts as an optical antenna.
When light hits these structures, it hits the integrated optical surface and is processed by special algorithms. This whole setup is known as Neural Nano Optics and is a big step forward from previous mini-cameras.
Camera with it’s software can send the data through wireless and fibre broadband connections, so internet streaming is a possibility.
A miniature camera with excellent performance
The main innovation of the camera is the integrated design of the optical surface and the signal processing algorithm that generates the image. This has improved the device’s performance in natural light, unlike previous meta-field cameras that needed pure laser light or other ideal conditions to produce high-quality images.
The researchers compared the images produced by their system with images taken with earlier meta-field cameras and conventional cameras. Apart from slight blurring at the edges of the frame, images from the new device are comparable to those taken with conventional cameras with lenses 500,000 times larger.
Other ultra-compact, metastable lenses have severe image distortion and a lower field of view. They also have a limited ability to capture the full spectrum of visible light corresponding to the perceptual characteristics of the human eye, known as RGB imaging.
“Designing and tuning these tiny microstructures to do what we wanted them to do was a challenge,” says co-author Ethan Tseng of Princeton University. “For this specific task of capturing wide-field RGB images, it was not clear until now how to design millions of nanostructures together using post-processing algorithms,” he adds.
Metasurfacing is a standardised and low-cost method for device fabrication.
It is made of silicon nitride, a glass-like material that follows the standard semiconductor manufacturing method used in computer chips. This means that the metasurface camera design can be mass-produced at a lower cost than conventional camera lenses.
If you put thousands of these cameras on any surface, the entire surface becomes a camera. “We can turn individual surfaces into very high-resolution cameras, so instead of three cameras on the back of a mobile phone, the entire back of the phone becomes one big camera,” says Felix Heide of Princeton University.
The team is now working to expand the device’s computational capabilities. The researchers not only want to further improve image quality, but also add object detection capabilities and other functions that could make it a revolutionary tool in medicine and robotics.
Such miniaturised devices enable minimally invasive endoscopy. They can be used to diagnose and treat diseases. There is even a possibility that with a very little latency an expert doctor could perform a procedure remotely by using a wireless or fibre broadband connection to the software. It can also be used as a robotic vision system where limited size and light weight are important. The intelligence services of many countries are also happy about this.