Vehicle Surroundings Imaging With Cartesian 3D Projection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional three-dimensional shell-shaped environmental models used in camera-based driver assistance systems suffer from increasing sampling point resolution with distance from the vehicle, leading to distortion artifacts and uneven image quality due to their polar coordinate-based projection algorithms.

Innovation Solution

A three-dimensional environmental model with equidistantly arranged projection points in Cartesian coordinates, forming a bowl-shaped projection surface, where the distance between points is defined using Cartesian coordinates and adjusted based on distance from the vehicle and other parameters to maintain constant image resolution and reduce distortions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a polar coordinate-based projection algorithm is used with a three-dimensional shell-shaped environment model, then the projection of real image data onto the virtual model can be achieved, but the sampling point resolution increases with distance from the vehicle, causing distortion artifacts and uneven image quality

Engineering Contradiction:
Improvesampling point resolutionVSAvoidimage quality uniformity
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent changes the coordinate system parameter from polar to Cartesian coordinates, and modifies the projection surface from a shell-shaped model to a planar model. This parameter change ensures that sampling points are equidistantly distributed across the entire projection area, maintaining constant resolution regardless of distance from the vehicle, thereby eliminating distortion artifacts and achieving uniform image quality throughout the displayed view

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If the distance between projection points increases with radial distance from the vehicle, then the projection algorithm can be implemented using polar equations, but the virtual image resolution becomes non-uniform with best resolution achieved in areas where no real image data is available

Engineering Contradiction:
Improveprojection algorithm implementationVSAvoidvirtual image resolution
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

Instead of allowing the distance between projection points to increase with radial distance as in polar coordinates, the patent inverts this approach by using Cartesian coordinates where the distance between projection points remains constant throughout the projection area. This inversion ensures that the highest resolution is achieved uniformly across all areas, including regions where real image data is available, rather than concentrating best resolution in areas without real data

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentEP3308361B1Method for generating a virtual image of vehicle surroundings
Publication Date: 2024.04.10 CONTINENTAL AUTONOMOUS MOBILITY GERMANY GMBH
  • EP3308361B1 patent drawingFigure 1
  • EP3308361B1 patent drawingFigure 2
  • EP3308361B1 patent drawingFigure 3

AI summary

The invention relates to a method for generating a virtual image of vehicle surroundings, in which real images of the vehicle surroundings are recorded by means of one or more cameras positioned on the vehicle and in order to generate the virtual image, individual pixels from the real images captured by means of the cameras are projected onto projection points (Pi) in a three-dimensional surroundings model of the vehicle surroundings.