Amap launches ADAS warning navigation function: using visual AI technology to provide intelligent reminders of vehicle and pedestrian collisions

On November 18, the new version of Gaode Map recently launched the innovative ADAS warning navigation function. With the help of visual AI technology, it can intelligently identify vehicles and pedestrians in front, and provide collision warnings, lane departure and other safety reminders to prevent risks.

[[435738]]

According to reports, this function uses a mobile phone camera or a connected dashcam to capture the road ahead, and with the help of visual AI technology, it can intelligently identify traffic elements such as vehicles, pedestrians, and lane lines ahead. If there is a risk ahead, Amap will provide vehicle collision warnings, pedestrian collision warnings, lane departure warnings, and front vehicle start warnings through navigation screens and voice broadcasts to assist drivers in warning potential safety risks.

Amap's ADAS warning navigation function currently supports most mainstream iPhones and Android phones. It is suitable for regular driving navigation and the recently launched lane-level navigation, and can be used by all Amap users. During the pre-trip route planning and navigation process, users can enable this function by clicking "ADAS Warning Navigation" in the toolbar. It should be noted that since this function requires taking a picture of the road ahead, the phone needs to be fixed in a place with a good view ahead.

IT Home learned that mobile phone users can also connect to many mainstream brands of dashcams without having to turn on the phone camera. In the future, AutoNavi ADAS warning navigation function will be open to support more models.

According to AutoNavi, the ADAS warning and navigation function mainly relies on AutoNavi's leading visual AI technology. With the efficient visual tracking technology solution, it can lock the location of vehicles and pedestrians in the road image ahead within 10 milliseconds. At the same time, relying on millions of road traffic visual training models, its vehicle recognition accuracy rate is as high as over 97%.