Dynamic Augmented Reality Image Adjustment for Vehicle Road Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Augmented reality images in vehicles often appear misplaced due to changes in the vehicle's relative orientation to the road surface, especially on non-flat roads or when the vehicle inclines, and cannot account for real-time objects in the field of view, with existing solutions requiring extensive terrain mapping or processor-intensive lidar systems.
Innovation Solution
A system using a front-facing camera to detect lane lines and determine the road's three-dimensional ground plane, adjusting the augmented reality image to align with the road surface and projecting it correctly in a head-up display region, while also considering the driver's eye position and real-time objects, utilizing existing camera technology to reduce processing intensity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional AR image projection is used without dynamic adjustment, then the system is simple and computationally lightweight, but the AR image appears misplaced on non-flat roads or when the vehicle inclines
Solution Approach 1:
The system dynamically adjusts the AR image projection based on real-time detection of lane lines and ground plane orientation. The processor continuously updates the projection parameters to match the current road geometry and vehicle orientation, transforming a static projection system into a dynamic one that adapts to changing conditions.
Solution Approach 2:
The system introduces an intermediary ground plane model that mediates between the camera input and the AR image projection. By detecting lane lines and interpolating to determine the ground plane orientation, the system creates a virtual reference surface that guides the AR image placement, ensuring accurate alignment without requiring direct complex terrain mapping.
2Measurement precision
If extensive terrain mapping or lidar systems are used, then AR image placement accuracy improves, but computational complexity and system cost increase significantly
Solution Approach 1:
The system extracts only the essential information needed for AR image placement from the camera input - specifically the lane lines and ground plane orientation. By focusing on extracting these key geometric features rather than processing complete terrain maps or using lidar point clouds, the system achieves accurate ground plane detection with significantly reduced computational complexity.
Solution Approach 2:
The system replaces expensive, complex lidar systems and extensive terrain mapping infrastructure with a simple camera and lightweight image processing algorithms. The camera captures transient visual information that is processed in real-time to determine ground plane orientation, eliminating the need for costly persistent terrain databases or complex sensing systems.
3Ease of operation
If the AR image is projected without accounting for driver's eye position, then the projection system is simpler, but the visibility and usability of the AR information decreases
Solution Approach 1:
The system applies local quality adjustment by tailoring the AR image projection specifically to the driver's eye position and orientation. By monitoring eye position and adjusting the projection parameters accordingly, the system ensures that the AR information is optimally positioned in the driver's field of view, enhancing visibility and usability without requiring complex multi-point or adaptive display systems.
Data Source
AI summary
A system and method to adjust an augmented reality (AR) image involve a front-facing camera in a vehicle. The camera obtains an image that includes a road surface ahead of the vehicle. The system also includes a processor to detect lane lines of the road surface in three dimensions, to perform interpolation using the lane lines to determine a ground plane of the road surface, and to adjust the AR image to obtain an adjusted AR image such that all points of the adjusted AR image correspond with points on the ground plane.