Camera Attitude Correction for Reliable Vehicle Object Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object detection systems in vehicles struggle with accurate object detection and distance calculation without external range sensors, leading to increased system complexity and cost, and unreliable detection due to environmental changes during vehicle operation.

Innovation Solution

An external environment recognition device that calculates camera pose parameters and adjusts geometric image transformation parameters to maintain accurate object detection and distance calculation, using onboard stereo cameras and sensors to adapt to changing environmental conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If geometric image transformation is performed using initially set camera pose parameters without considering environmental changes, then the system configuration remains simple, but object detection accuracy deteriorates and erroneous detections increase

Engineering Contradiction:
Improvesystem configurationVSAvoidobject detection accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system dynamically updates camera pose parameters based on detected landmarks and environmental changes, rather than using static initial parameters. This allows the system to adapt to changing conditions (vehicle movement, road geometry changes) while maintaining accurate object detection, resolving the contradiction between system simplicity and detection accuracy

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements a feedback mechanism where detected landmarks are used to recalculate and update camera pose parameters, which are then used to correct subsequent object positions. This closed-loop approach maintains detection accuracy without requiring complex external sensors by continuously adapting to environmental changes

Inventive Principle:
Principle #23Feedback

2Measurement precision

If external range sensors are added to accurately identify object positions, then object detection accuracy improves, but system complexity and cost increase

Engineering Contradiction:
Improveobject position identification accuracyVSAvoidsystem configuration
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system uses naturally occurring environmental landmarks (buildings, poles, road markings) as intermediaries to indirectly determine camera pose and object positions. Instead of directly measuring distances with external range sensors, the system infers position information through image-based landmark recognition and geometric relationships, achieving accurate detection without additional expensive sensors

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system uses the camera's own image data to calculate camera pose parameters and correct object positions, rather than relying on external range sensors. By leveraging the existing camera system and environmental landmarks, the method achieves accurate measurement while avoiding the need for additional sensing hardware

Inventive Principle:
Principle #25Self-service

3Productivity

If camera pose parameters are not updated during vehicle operation, then processing speed remains high, but detection reliability deteriorates due to environmental changes

Engineering Contradiction:
Improveprocessing speedVSAvoidobject detection reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system performs periodic updates of camera pose parameters by detecting landmarks in sequential images and recalculating parameters at regular intervals. This periodic recalibration maintains detection reliability under changing environmental conditions while avoiding continuous processing that would reduce overall system productivity

Inventive Principle:
Principle #19Periodic action

Data Source

PatentEP3889946B1External environment recognition device
Publication Date: 2025.12.10 ASTEMO LTD
  • EP3889946B1 patent drawingFigure 1
  • EP3889946B1 patent drawingFigure 2
  • EP3889946B1 patent drawingFigure 3(a)~3(d)

AI summary

Provided is an external environment recognition device, in which it is possible to: calculate a camera attitude parameter and adjust a geometric image conversion parameter to adjust the relationship of one or more device cameras relative to the current environment; reduce the false positive rate and at the same time maintain the accuracy of object detection, and thereby improve the object detection reliability; and improve the accuracy of calculating the distance to a detected object without adding an external distance sensor. This invention is provided with: an image acquisition unit for acquiring an image from a camera sensor for capturing an image of the external environment; a sensor information acquisition unit for acquiring sensor information collected from a sensor for detecting information relating to the current attitude of the camera sensor; a camera attitude parameter calculation unit for calculating, on the basis of the sensor information acquired by the sensor information acquisition unit, at least one camera attitude parameter defined by the camera pitch angle, the camera roll angle, and the camera yaw angle, for correcting the current attitude of the camera sensor relative to a predefined reference; and an image conversion unit for performing, on the basis of the camera attitude parameter calculated by the camera attitude parameter calculation unit, a geometric image conversion on the image acquired by the image acquisition unit.