Road Sign Detection Using Dictionary and GLCM Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for detecting road signs struggle to accurately differentiate between signs and light sources, leading to false positives and potential safety hazards due to driver distraction.

Innovation Solution

A device and method that utilize dictionary information generated from multiple images, including Gray Level Co-Occurrence Matrix and decision tree classifiers, to determine the presence of road signs by analyzing pixel brightness distributions and spatial relationships, and employing shape classification algorithms to reduce confusion with light sources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional road sign detection methods are used, then the system is simple to implement, but it produces false positives by confusing road signs with light sources

Engineering Contradiction:
Improveroad sign detection accuracyVSAvoidimage processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the image processing into multiple distinct stages: color space conversion to HSV, filtering based on color ranges, connected component analysis, and classification using dictionary information and Gray Level Co-Occurrence Matrix. This segmentation allows each stage to focus on specific aspects of road sign detection, improving overall accuracy while managing complexity through modular processing.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary classification system that uses dictionary information and Gray Level Co-Occurrence Matrix features to distinguish between road signs and light sources. This intermediary layer processes extracted features and provides final classification, acting as a mediator between raw image data and detection decisions, thereby reducing false positives.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If driver distraction is considered, then human factors are acknowledged, but automatic detection systems cannot compensate for driver inattention

Engineering Contradiction:
Improvesafety warning reliabilityVSAvoidautomatic detection capability
Core Design Contradiction:
ReliabilityVSExtent of automation

Solution Approach 1:

The system performs self-service by automatically detecting road signs and generating warnings without requiring driver intervention or awareness. The automated detection process continuously monitors the environment and provides warnings independently of driver state, ensuring safety even when drivers are distracted.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system provides feedback to the driver through automated warnings when road signs are detected. This feedback mechanism ensures that drivers are informed of potential safety issues even when they may be distracted, creating a closed-loop safety system that compensates for human limitations.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If dictionary information from multiple images is used, then the discrimination between road signs and light sources improves, but the data processing requirements increase

Engineering Contradiction:
Improvediscrimination accuracyVSAvoidimage data quantity
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent performs preliminary action by pre-processing and organizing image data into a structured dictionary format before actual detection. The Gray Level Co-Occurrence Matrix is pre-computed and stored as reference data, allowing rapid comparison during real-time detection without reprocessing the entire image dataset, thus managing data quantity requirements.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies local quality by analyzing specific regions of interest in images rather than processing entire images uniformly. The connected component analysis and Gray Level Co-Occurrence Matrix computation focus on localized areas that contain potential road signs, reducing the overall data processing burden while maintaining high discrimination accuracy.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3336745B1Device for determining the presence of a road sign
Publication Date: 2022.08.24 CONTINENTAL AUTONOMOUS MOBILITY GERMANY GMBH
  • EP3336745B1 patent drawingFigure 1~2
  • EP3336745B1 patent drawingFigure 3a~3e
  • EP3336745B1 patent drawingFigure 4

AI summary

The present invention relates to a device for determining the presence of a road sign. It is described to provide (210) at least one image, the at least one image relating to a scene external to a vehicle. Dictionary information is provided (220), wherein the dictionary information is generated on the basis of a plurality of images, and wherein at least some of the image data of the plurality of images is image data associated with at least one road sign. It is determined (230) if a particular type of road sign is present in the at least one image on the basis of image processing of the at least one image and the dictionary information. The image processing comprises determining (232) a distance and spatial relationship for at least one pixel value in the at least one image. Driving warning information is output (240) if the particular type of road sign is determined to be present.