Vehicle Positioning via Camera-Aerial Image Alignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Satellite-based positioning methods for determining a vehicle's position on a multi-lane road lack sufficient accuracy, making it difficult to reliably identify the current lane and detect danger spots such as traffic jams, intersections, and potholes.

Innovation Solution

A method utilizing a camera and a sensor unit on the vehicle to capture images and generate a transformed image with a virtual downward perspective, combined with satellite-based positioning, to accurately determine the vehicle's position by recognizing features and segments in both the camera and aerial images, and using distance sensors for depth mapping to enhance precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If satellite-based positioning is used to determine vehicle position, then the positioning system is simple to implement, but the measurement precision is insufficient (1-10m accuracy cannot reliably identify current lane)

Engineering Contradiction:
Improvevehicle position accuracyVSAvoidpositioning system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines satellite-based positioning with image processing technology to achieve high-precision vehicle positioning. The control unit integrates GPS coordinates with camera images and digital terrain models, merging multiple positioning approaches to overcome the limitations of satellite positioning alone and achieve lane-level accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces a digital terrain model as an intermediary between the satellite positioning system and the final position determination. The terrain model provides detailed road geometry information that enhances the coarse satellite position data, enabling precise lane identification through image matching and feature recognition.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If a full camera image is processed to generate transformed image, then the position detection accuracy is improved, but the computational power and processing time increase

Engineering Contradiction:
Improveposition detection accuracyVSAvoidprocessing speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the camera image processing into specific regions of interest, particularly focusing on road surface areas and hazardous location markings. By processing only relevant image sections rather than the entire image, the system maintains high position detection accuracy while reducing computational load and processing time.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and processes only the essential image features needed for position determination, such as road markings, hazardous location signs, and terrain characteristics. This selective extraction of relevant information from the full image reduces computational requirements while preserving the accuracy needed for lane and position identification.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3721371B1Method for position determination for a vehicle, controller and vehicle
Publication Date: 2024.07.24 ROBERT BOSCH GMBH
  • EP3721371B1 patent drawingFigure 1
  • EP3721371B1 patent drawingFigure 2
  • EP3721371B1 patent drawingFigure 3

AI summary

A method for position determination for a vehicle (100), wherein the vehicle (100) has at least one camera (110) and a sensor unit (140) for a global satellite navigation system arranged on it, having the following method steps: capturing (310) at least one camera image of the surroundings (200) of the vehicle (100) by means of the camera (110); producing (320) a transformed image on the basis of the captured camera image, wherein the transformed image has a vertically downwardly directed virtual perspective (121); determining (330) a satellite position of the vehicle (100) by means of a satellite-assisted position determination; and providing (340) an aerial image of the surroundings (200) of the vehicle (100) on the basis of the determined satellite position; wherein subsequently a detection (394) of an orientation of the transformed image in the provided aerial image and an ascertainment (395) of a vehicle position on the basis of the detected orientation of the transformed image in the provided aerial image are performed.