Ouster, Inc.
DRIVER VISUALIZATION AND SEMANTIC MONITORING OF A VEHICLE USING LIDAR DATA

Abstract:

Methods are provided for using a light ranging system of a vehicle. A computing system receives, from light ranging devices, ranging data including distance vectors to environmental surfaces. A distance vector can correspond to a pixel of a three-dimensional image stream. The system can identify a pose of a virtual camera relative to the light ranging devices. The light ranging devices are separated from the pose by first vectors that are used to translate some of the distance vectors using the first vectors. The system may determine colors associated with the translated distance vectors and display pixels of the three-dimensional image stream using the colors at pixel positions specified by the translated distance vectors. The system may use one or more vehicle models with the ranging data to provide semantic labels that describe a region that has been, or is likely to be, in a collision.