Photonic Chip with Sensing and Computing Speeds Machine Vision

A new intelligent photonic-sensing computing chip processes, transmits, and reconstructs images of a scene within nanoseconds. This will enable extremely high-speed image processing edge intelligence for machine vision apps, including autonomous driving, industrial inspection, and robotic vision.

Researchers at Tsinghua University in China claim the new chip performs all the processes in nanoseconds by keeping them all in the optical domain. The research is published in the Optica journal, where the team describes the new chip as an optical parallel computational array (OPCA) chip. With a potential processing bandwidth of up to one hundred billion pixels and a response time of six nanoseconds, the chip is six orders of magnitude faster than current methods. They also created an optical neural network with the chip that integrates image perception, computation, and reconstruction.

The challenge has been in performing both image acquisition and analysis on the same chip in the optical domain is to find a way to convert the free-space spatial light used for imaging into an on-chip guided light wave. The new chip consists of a sensing-computing array of dedicated designed ring resonators that convert a free-space optical intensity image into a coherent light signal that can be guided on the chip. A micro-lens array enhances the process by focusing the scene onto the OPCA chip.

The researchers showed the chip could be used to classify a handwritten image and to perform image convolution, a process that applies a filter to an image to extract features. The chip architecture can effectively complete information compression and scene reconstruction, indicating its potential for widespread applications.