Road Boundary Detection Using 3D Height Maps

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting road boundaries demarcated by three-dimensional objects face challenges such as reduced accuracy due to environmental changes, low-height objects, and occlusions, particularly in roads with slope changes or distant three-dimensional objects.

Innovation Solution

A road boundary detection device using two cameras mounted on a vehicle, which calculates three-dimensional distance data and projects it onto planes to detect road boundaries by identifying strips of data, applying a road model to estimate the road shape and using template matching to determine the presence and position of three-dimensional objects, even in occluded areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If stereo image processing is used to detect three-dimensional objects, then road boundary detection capability is improved, but detection accuracy deteriorates under environmental changes such as brightness reduction

Engineering Contradiction:
Improveroad boundary detection capabilityVSAvoiddetection accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent transitions from two-dimensional image processing to three-dimensional spatial reasoning by constructing a bird's-eye view image from multiple camera inputs. This dimensional transformation allows the system to detect road boundaries by analyzing height differences and spatial relationships in 3D space, making detection more robust to environmental changes like brightness variations that affect 2D image processing.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent introduces a bird's-eye view image as an intermediary representation that synthesizes information from multiple cameras. This intermediate 3D spatial model serves as a mediator between raw image data and road boundary detection, enabling the system to reason about road boundaries through height maps and spatial relationships rather than directly processing potentially degraded 2D images.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Area of stationary object

If camera-based detection is used for road boundaries, then detection range is improved, but detection reliability deteriorates in cases of occlusion or unclear images

Engineering Contradiction:
Improvedetection rangeVSAvoiddetection reliability
Core Design Contradiction:
Area of stationary objectVSReliability

Solution Approach 1:

The patent merges data from multiple cameras mounted at different positions on the vehicle to construct a comprehensive bird's-eye view image. By combining information from multiple viewing angles and elevations, the system achieves wider detection coverage while the redundant information from multiple sources provides robustness against occlusion and unclear images through cross-validation.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent adds the height dimension to traditional 2D image processing by constructing a 3D height map from multi-camera inputs. This dimensional enhancement allows the system to detect road boundaries based on vertical height differences of objects like curbs and barriers, providing reliable detection even when 2D image features are occluded or ambiguous.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Area of stationary object

If multiple cameras are used for three-dimensional object detection, then detection coverage is improved, but system complexity increases

Engineering Contradiction:
Improvedetection coverageVSAvoidsystem complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The patent makes the bird's-eye view image construction process serve multiple functions: it simultaneously provides a wide-area detection view, enables 3D spatial reasoning for road boundary detection, and creates a unified coordinate system for integrating data from multiple cameras. This multi-functionality reduces the need for separate processing pipelines for each camera, simplifying the overall system despite using multiple sensors.

Inventive Principle:
Principle #6Universality (Multi-functionality)

4Speed

If traditional image processing is used for low-height objects, then processing speed is maintained, but detection precision deteriorates

Engineering Contradiction:
Improveprocessing speedVSAvoiddetection precision
Core Design Contradiction:
SpeedVSMeasurement precision

Solution Approach 1:

The patent addresses low-height object detection by transitioning to 3D height map analysis in the bird's-eye view representation. Low-height objects that are difficult to detect in 2D images become detectable through their height differences from the road surface in the 3D height map, maintaining processing efficiency while significantly improving detection precision for such objects.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP2336999B1Device for detecting/judging road boundary
Publication Date: 2020.04.22 HITACHI AUTOMOTIVE SYST LTD
  • EP2336999B1 patent drawingFigure 1~2
  • EP2336999B1 patent drawingFigure 3
  • EP2336999B1 patent drawingFigure 4

AI summary

There is provided a road boundary detection/judgment device resistant to environmental change and capable of detecting even a road boundary demarcated by a three-dimensional object in the distance. The device is provided with: an image acquisition section having two or more cameras for image-capturing the road area; a distance data acquisition section acquiring three-dimensional distance information about an image-capture area on the basis of an image obtained by the image acquisition section; a road boundary detection section detecting the height of a three-dimensional object existing in the road area on the basis of the three-dimensional distance information obtained by the distance data acquisition section to detect a road boundary; and a same boundary judgment section transforming the image, for a first road area where the height of a three-dimensional object corresponding to a road boundary could be detected and a second road area where the height of a three-dimensional object corresponding to a road boundary could not be detected, and judging whether the three-dimensional object corresponding to the first road area and the three-dimensional object corresponding to the second road area are the same. If it is judged that the three-dimensional objects corresponding to the first and second road area boundaries are the same, the second road area is reset as the first road area.