Automotive computer vision aims at safety and autonomous driving
The automotive industry is rapidly expanding the use of cameras for a variety of safety-enhancing and autonomous driving functions. Reversing cameras are enhanced with computer vision functions to automatically brake and prevent accidents when resetting, surround view systems significantly increase the visibility of the environment and include automated parking functions, front cameras are used to maintain distance or, if necessary, brake, and side cameras replace mirrors. Autonomous vehicles use a variety of cameras with computer vision techniques to capture and understand their entire environment.
Videantis has been working for some time with Adasens, a company of the Spanish Ficosa Group. According to Videantis, intelligent cameras with the company’s vision processors will go into series production in 2019.
The background of the cooperation lies in the fact that today’s tasks of automotive image processing often require several powerful CPUs and GPUs, which consume very high electrical power (typically several hundred watts) to process images into meaningful information for the control of a vehicle. The Videantis processor architecture is said to perform these complex computer vision and image processing tasks much faster and with significantly lower power consumption, so that this technology can be embedded in smaller ECUs and even directly into tiny cameras.
At the CES 2018 in Las Vegas from 9 to 12 January, Videantis intends to show a vehicle camera from Ficosa in which the Adasens algorithms run on the Videantis Embedded Vision processor.