Highly accurate positioning and object-finding are critical for many automotive and mobility related use-cases – from driver-facing AR through accurate package drop-off and pick-up to autonomous navigation in urban canyons or parking lots. Existing systems like AGPS are not accurate enough in many situations and rely on a grid system that everything is fixed to. LIDAR and radar are useful but see only partial images (outlines) thus losing much of the relevant information.
White Raven’s visual positioning system uses deep neural networks (DNNs) to learn an imagery model of a location, and compares it in real time with camera input to identify buildings, landmarks and points of interest accurately. These can be accurately marked in the visual output (screen or heads-up-display) for the benefit of the driver and passengers. Furthermore this highly accurate system can be used to triangulate the location of the camera (i.e. the car) with high accuracy to enable autonomous operations.