47.3: Invited Paper: World Fixed Augmented‐Reality HUD for Smart Notifications

Abstract

In this paper, we present a Head‐Up Display (HUD) system that displays perspective correct surrounding information for driver safety. In our implementation, we display contrast enhanced ego lane information on the HUD during driving conditions with poor visibility. The use of HUD for Augmented Reality (AR) needs an accurate model of the picture generation process for proper visualization of the 3D content that is perspective accurate from the user view point. The system uses multiple computer vision techniques for the accurate visualization.First, we use a camera calibration algorithm that uses view independent spatial geometry and view dependent perspective transformation that compensates for the distortion of the optics as well as the combiner. The calibration is done by capturing the test patterns from the front facing camera located close to the rear view mirror and a camera from the driver's point of view. Further registration of the HUD image is done through a head tracker that uses the camera mounted for the driver monitoring on the steering wheel. The second module detects the lane markings on the road and returns the location of the polygon in the camera co‐ordinate system. The image data bounded by this polygon is further enhanced to reduce the effect of haze, rain and insufficient illumination and blended with the camera input from the front facing camera. The HUD then projects the synthesized image that is perspective correct rendered in the vehicle co‐ordinate system.