Pedestrians can be nearly invisible on dark roads, but Toshiba has developed a new image processor to help cars spot them.
The TMPV7608XBG image-recognition processor works with cameras mounted on a car to increase driver awareness and traffic safety.
It can alert the driver to nearby pedestrians even along dark roads where there are only a few lux of ambient light, according to the electronics maker, which claims it works as well as similar systems for daytime use.
Newly developed image-processing techniques allow the unit to detect stationary objects such as pylons, fallen trees and debris from landslides, as well as their distance, height and width.
The processor contains eight new image-processing engines so it can recognize pedestrians and vehicles simultaneously in 50 milliseconds, which is half the time of Toshiba’s previous processor.
“Although there are several image-recognition processors for automotive applications in the market including ours, we believe our new processor is superior in the industry with the following features,” a Toshiba spokesman wrote in an email, “night-time pedestrian detection; detection of the stationary obstacles on the road not detected by conventional pattern-recognition algorithms; and simultaneous execution of up to eight applications.”
The processor is meant to support smart car functions such as autonomous emergency braking, traffic sign recognition and forward collision warning systems that are slated to become part of the European New Car Assessment Programme, a testing regime for vehicles, from 2018.
Toshiba will ship samples of the processor starting in January, with mass production slated to begin in December 2016.
Carmakers have been adopting awareness technology at a quickening pace. Honda recently enhanced its 2015 Legend with a millimeter-wave radar in the grille and a monocular camera inside the windshield to detect nearby cars and people. Ford has said its 2015 Mondeo in Europe would feature a similar system, which will intervene when vehicle or pedestrian crashes seem imminent.