Lane Marking Recognition Using Virtual Lines for Automated Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Instability in sensor recognition leads to erroneous identification of lines similar to lane markings as actual lane markings, causing potential misrecognition during automated driving.
Innovation Solution
A mobile object control device that includes a lane marking candidate recognizer, a virtual line setter, and a lane marking searcher, which sets virtual lines outside recognized lane markings and searches for the current lane marking by determining the closest candidate to these virtual lines, excluding those within a predetermined distance as boundary lines rather than lane markings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If sensor recognition is used to identify lane markings, then automated driving control can be implemented, but erroneous recognition occurs when lines similar to lane markings are mistaken for actual lane markings
Solution Approach 1:
The system performs preliminary recognition of lane markings in past images before processing current images. By establishing reference lane marking positions from historical data, the system creates a baseline for comparison that helps filter out false detections in real-time processing, thereby improving recognition accuracy while maintaining automation.
Solution Approach 2:
The system uses feedback by comparing lane marking candidates from current images against previously recognized lane markings. The relative position relationships and consistency checks create a feedback mechanism that validates detections and eliminates erroneous recognitions, resolving the contradiction between automation and precision.
2Measurement precision
If virtual lines are set outside recognized lane markings to improve accuracy, then lane marking recognition precision improves, but device complexity increases due to additional processing components
Solution Approach 1:
Virtual lines serve as intermediary elements that mediate between recognized lane markings and lane marking candidates. These virtual lines act as reference boundaries that simplify the comparison process, improving accuracy without requiring complex algorithms. The intermediary virtual lines reduce computational complexity by providing clear reference points for validation.
3Measurement precision
If lane marking candidates are filtered by distance from virtual lines, then recognition accuracy improves, but processing time increases due to additional distance calculations
Solution Approach 1:
The system applies local quality by performing detailed distance calculations only for lane marking candidates in critical regions near virtual lines, rather than uniformly processing all candidates. This selective approach maintains high accuracy for borderline cases while reducing overall processing time by skipping unnecessary calculations for clearly valid or invalid candidates.
Data Source
AI summary
A mobile object control device of an embodiment includes a lane marking candidate recognizer configured to recognize candidates for a lane marking that demarcates a traveling lane in which a mobile object travels from an image including surroundings of the mobile object captured by an imaging unit, a virtual line setter configured to set at least one or more virtual lines outside of the lane marking recognized in the past as a lane marking that demarcates the traveling lane, when viewed from the mobile object, and a lane marking searcher configured to search for a lane marking that demarcates a current traveling lane of the mobile object on the basis of the lane marking candidates recognized by the lane marking candidate recognizer and the one or more virtual lines.