Lane Marking Recognition Using Leading Vehicle Reference Position
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lane boundary line detection systems often fail to accurately recognize left and right lane markings, especially in situations like traffic jams where the leading vehicle is close, leading to reduced recognition precision due to shortening white line lengths and interference from old white lines, road repair markings, or braking marks.
Innovation Solution
An image processing device that includes a lane marking recognition section and a leading vehicle recognition section, which determines true lane marking pairs by calculating width direction discrepancies between lane markings and a leading vehicle's reference positions within a threshold value, ensuring high precision recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If separate determination of left and right lane boundary lines is performed using distance to leading vehicle end sections, then processing is simplified, but recognition precision deteriorates due to incorrect pair combinations
Solution Approach 1:
The patent merges the separate determination processes for left and right lane boundary lines into a unified pair determination process. By calculating the distance between left and right lane boundary line candidates simultaneously and comparing it with the expected lane width, the system correctly identifies lane marking pairs without the errors that occur when determining boundaries separately.
2Ease of operation
If separate determination of left and right lane boundary lines is performed, then individual boundary detection is simplified, but false positives increase from old white lines and road markings
Solution Approach 1:
The patent combines the detection of left and right lane boundary lines into a pair-based detection system. By verifying that the distance between detected boundaries matches the expected lane width, the system filters out false positives from old white lines, road repair markings, and braking marks that would not form correct geometric pairs.
3Measurement precision
If the leading vehicle is close during traffic jam, then vehicle detection is easier, but the recognizable white line length shortens reducing precision
Solution Approach 1:
The patent changes the detection parameters from relying on long white line segments to using the geometric relationship between left and right boundaries. By measuring the distance between paired boundaries and comparing with expected lane width, the system maintains high precision even when white line length is short due to close proximity to the leading vehicle.
4Measurement precision
If threshold based on lane width is applied, then recognition accuracy improves, but processing complexity increases
Solution Approach 1:
The patent applies a simple threshold comparison between the measured distance of paired lane boundaries and the expected lane width. This single parameter threshold check provides high recognition accuracy without introducing complex processing, maintaining system simplicity while improving precision.
Data Source
AI summary
The image processing device includes a lane marking recognition section that recognizes left and right lane markings from a surroundings image captured by an imaging section, and a leading vehicle recognition section that recognizes a leading vehicle from the surroundings image. The lane marking recognition section determines that the left and right lane markings are a true lane marking pair in cases in which a width direction discrepancy, or a value based on the discrepancy, between a reference position between the left and right lane markings and a reference position of the leading vehicle is within a threshold value.