Hyundai Mobis Co. has announced that it plans to apply the deep learning-based camera image detection technology to its vehicles to support autonomous driving for the first time from 2022 for mass production.
South Korea’s biggest auto parts maker said Thursday, ahead of the opening of the Seoul Motor Show 2019, that it will complete the development of the deep learning-based camera image detection technology which uses an artificial intelligence (AI) system by the end of this year. Image recognition technology is considered necessary to develop self-driving cars, as their camera sensors should be able to detect and distinguish one object from another, including vehicles and pedestrians, instead of a human driver
By training the system with a large amount of camera image, Hyundai Mobis plans to develop a camera sensor capable of conducting a trillion operations per second (teraflop). The company will also apply the advanced technology to its Surround View Monitor (SVM), which helps drivers park their vehicles by offering a 360-degree view. This will allow future cars to automatically control movements to prevent a collision from the front, side, and back.
“We are aiming to secure deep learning-based information and communication technology (ICT) to differentiate our autonomous driving technology in the fields of future vehicles,” said Jang Jae-ho, head of the Electrical and Electronics R&D Center of Hyundai Mobis.