Computer vision – a mobile approach to modern automotive

The automotive industry has been being revolutionized by three important trends:

(1) the shift to Electric Vehicles,

(2) Internet of Vehicle for interconnections between vehicles and connections between vehicles and infrastructure, and

(3) Autonomous Driving.

The former is driven not only by the shortage of fossil energy but also by the full intervention in steering control of Electric Vehicles, facilitating Autonomous Driving function. The advances of camera technology and computer vision techniques have been emerging as indispensable keys to revolutionize the automotive industry. It enables various Advanced Driving Assistance System (ADAS) products and the recent rise of Autonomous Driving.

Although there exist a variety of commercial computer vision-based ADAS products and Autonomous Driving systems, our quest for future generation of intelligent vehicles is still in the middle of nowhere. These products are implemented on embedded platforms which are still the most dominant in the automotive industry. However, the computation capability of personal mobile devices like iPad and Android tablets is increasingly powerful, making it very suitable to automotive applications. This white paper therefore addresses a mobile computer vision-based approach to modernize the automotive industry in years to come.