Lane Detection Edge Filtering for Wet Road Reflections
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional lane departure warning systems struggle to accurately distinguish between road markings and reflections, particularly on wet roads, leading to incorrect lane detection and potential system failures.
Innovation Solution
A method that identifies and removes edges with almost vertical orientations, likely caused by reflections from headlights, by determining edge angles relative to a reference edge and filtering out those within a predetermined angular range, thereby improving lane detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional image processing algorithms detect all high-contrast edges in the image, then lane detection can be performed, but reflections from headlights on wet roads are incorrectly classified as road markings leading to incorrect lane detection
Solution Approach 1:
The patent applies local quality by analyzing the orientation characteristics of individual edges at different locations in the image. Edges with orientations within a specific angular range (indicative of reflections) are treated differently from edges with other orientations (likely road markings). This localized differentiation based on orientation allows the system to maintain high detection precision while improving reliability by filtering out reflection artifacts.
Solution Approach 2:
The patent changes the parameter of edge orientation angle to distinguish between reflections and road markings. By calculating the angle of each detected edge relative to the image coordinate system and comparing it against a predetermined angular range, the system dynamically filters edges based on their orientation parameter. This parameter-based filtering resolves the contradiction by allowing reliable lane detection while maintaining precision in identifying actual road markings.
2Productivity
If the image processing algorithm recognizes all high-contrast structures as lane markings, then lane detection can be performed quickly, but the system fails on wet roads with reflections
Solution Approach 1:
The patent applies preliminary action by filtering edges based on their orientation angle before proceeding with lane detection. The predetermined angular range is established in advance, and edges falling within this range are removed prior to lane marking identification. This preliminary filtering step prevents reflections from interfering with subsequent detection processes, maintaining both speed and reliability.
Solution Approach 2:
The system uses the orientation angle parameter to efficiently filter out reflection edges in a computationally simple manner. By comparing edge angles against a predetermined angular range, the system quickly eliminates false positives without requiring complex analysis, thus maintaining high processing speed while improving reliability on wet roads.
3Reliability
If edges with vertical orientation are removed from detection, then reflections are eliminated, but some legitimate road markings may be missed
Solution Approach 1:
The patent applies local quality by considering the spatial context and orientation of edges. The predetermined angular range is specifically chosen to target the characteristic vertical orientation of reflections while preserving road markings that have different orientations. This localized orientation-based differentiation ensures that only reflection artifacts are removed, maintaining high reliability in reflection rejection while preserving legitimate road marking detection precision.
Solution Approach 2:
The system carefully selects the predetermined angular range parameter to distinguish between reflections and road markings. By setting this parameter to target only edges with orientations characteristic of reflections (typically near-vertical), the system achieves high reliability in rejecting reflections while minimizing the risk of removing legitimate road markings, thus maintaining detection precision.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances lane detection reliability by distinguishing reflections from actual road markings, especially on wet roads, resulting in more accurate and precise lane recognition for improved driver assistance systems.
Implementation Method 1
a digitized image of a lane in front of a vehicle is evaluated with the method
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a lane detection method wherein a digitized image (10) of a lane (12) is evaluated, comprising the following steps: detecting edges (14, 16, 18, 20) in the image (S10), determining the angles of the detected edges with respect to a reference edge (22; S12), removing the edges (16, 20) from the plurality of detected edges (S14), the angles of which lie within a predetermined angle sector, and detecting the lane (12) based on the remaining edges (14, 18) of the plurality of detected edges (S16).