Unstructured Road Boundary Detection via HSI Color Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional edge-based approaches for road boundary detection in autonomous vehicles fail on unstructured roads due to the lack of clearly defined features in images.

Innovation Solution

A method for detecting unstructured road boundaries using a color image processing technique that involves noise suppression through an improved vector median filtering algorithm, conversion to the HSI color space, and model-based road segmentation with a Kalman filter for boundary estimation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional edge-based approaches are used for road boundary detection, then the detection can work on structured roads with clearly defined features, but it fails on unstructured roads lacking clearly defined features

Engineering Contradiction:
Improveroad boundary detection reliabilityVSAvoidadaptability to unstructured roads
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent transforms the detection approach from edge-based spatial features to color-based spectral features by converting images to HSI color space. This parameter change allows the system to detect road boundaries through color characteristics (hue, saturation, intensity) rather than relying on edge definitions, enabling reliable detection on unstructured roads without clearly defined geometric features

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical edge-detection algorithm with a color-space-based segmentation approach using Kalman filter for boundary estimation. This substitution moves from detecting geometric discontinuities to tracking color-based road regions through temporal filtering, achieving robustness on unstructured surfaces

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If noise suppression filtering is applied to reduce computational complexity, then real-time performance improves, but detection precision may be affected

Engineering Contradiction:
Improvereal-time detection performanceVSAvoidroad boundary detection precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent implements a Kalman filter that uses feedback from previous frame predictions to refine current boundary estimates. The filter continuously updates road boundary positions by combining predicted values with current color-space measurements, reducing noise while maintaining precision through iterative refinement rather than aggressive filtering

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary road region segmentation in HSI color space before boundary estimation, pre-identifying candidate road pixels based on color characteristics. This preliminary action reduces the search space and computational load for subsequent boundary detection while preserving detection accuracy through color-based region masking

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3036714B1Unstructured road boundary detection
Publication Date: 2019.12.04 HARMAN INT IND INC
  • EP3036714B1 patent drawingFigure 1
  • EP3036714B1 patent drawingFigure 2
  • EP3036714B1 patent drawingFigure 3~4

AI summary

A method for detecting unstructured road boundary is provided. The method may include: obtaining a color image; selecting a candidate road region within the color image according to a road model; identifying a seed pixel from the candidate road region; obtaining a brightness threshold and a color threshold, where the brightness threshold and the color threshold are determined according to brightness distances and color distances from pixels in the candidate road region to the seed pixel; and performing road segmentation by determining whether the pixels in the candidate road region belong to a road region based on the brightness threshold and the color threshold. The amount of computation can be reduced greatly by using the improved unstructured road boundary detection method.