The AI That Allows Self-Driving Cars to ‘See’ Around Corners

Self-driving technology has been expanding rapidly over the last few years.  New (REVEAL) technologies involving lasers eliminate “noise” and allow images from around corners to be produced.  This is important for accident avoidance and safer self-driving.  Read the article below for more information. Sterling Performance is proud to be a part of the expanding AI industry!

Researchers at US universities have created an imaging system powered by artificial intelligence that could help self-driving cars “see” around corners in minute detail to identify hazards.

The imaging system uses a conventional camera sensor and a laser beam that can be “bounced” off walls and onto objects to create a pattern – visually similar to the static on an old untuned television to the naked eye.

The image is then reconstructed by an AI algorithm, eliminating the ‘noise’ and able to produce images of even 1cm tall letters. Deep-learning is a form of artificial intelligence that mimics the working of the human brain to process data and create patterns.

It is a particularly powerful subset of machine learning AI, able to learn unsupervised with unstructured data.

“Compared to other approaches, our non-line-of-sight imaging system provides uniquely high resolutions and imaging speeds,” said research team leader Christopher Metzler from Stanford University and Rice University.

“These attributes enable applications that wouldn’t otherwise be possible, such as reading the license plate of a hidden car as it is driving or reading a badge worn by someone walking on the other side of a corner”.

The system is designed to image small objects at high resolutions, even able to produce sub-millimeter detail in unseen areas from over a meter away. Combined with other imaging systems, the new algorithm will be able to produce much larger reconstructions.

Previous algorithms used for real-time imaging of this sort have been tested, but the researchers say this system is far more “robust” to the noise created by the short exposure images needed for real-time imaging.

As well as the potential to make autonomous cars safer, allowing the system to identify hazards across busy intersections or around parked vehicles, the researchers say that this new form of real-time imaging could have a wide range of uses.

They cite installing the systemon satellites and spacecraft, allowing imaging of caves within asteroids and other planets, as well as more Earthly practices.

“Non-line-of-sight imaging has important applications in medical imaging, navigation, robotics and defense,” said co-author Felix Heide from Princeton University. “Our work takes a step toward enabling its use in a variety of such applications.”

The research is part of DARPA’s Revolutionary Enhancement of Visibility by Exploiting Active Light-fields (REVEAL) program, which is testing various technology to “see” around corners. The team’s next task is to expand its systems field of view.

The research was published in the science journal Optica.

Original article published on

Comments are closed.