Lane Marking Data Gap Completion via Unsupervised Feature Line Connectivity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems face challenges in accurately detecting lane markings due to incomplete or discontinuous data in map databases, which can lead to errors in autonomous vehicle navigation, potentially causing accidents.
Innovation Solution
The system employs an unsupervised process to identify and connect missing parts of feature lines in lane marking data using distance and orientation features, generating a complete feature line by forming a connecting line between border points of disconnected parts, thereby enhancing data accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional data collection methods are used for map data, then data collection is simple and cost-effective, but data completeness and accuracy deteriorate due to errors, occlusions, and false positives
Solution Approach 1:
The system uses unsupervised learning algorithms that automatically detect and fill missing lane marking data without requiring human annotation or manual correction. The algorithm processes the data itself to identify gaps and generate connecting lines, enabling the system to self-correct incomplete map data.
Solution Approach 2:
The patent replaces manual data collection and verification processes with automated computational algorithms. Instead of relying on physical sensors alone, the system uses image processing and machine learning models to detect, verify, and complete lane marking data, substituting mechanical data collection with intelligent processing.
2Measurement precision
If multiple data sources are used to improve detection accuracy, then detection reliability improves, but false positives and false negatives increase due to errors in sensor data and image recognition
Solution Approach 1:
The system implements a feedback mechanism where the unsupervised learning algorithm continuously refines lane marking detection by analyzing the consistency and logical validity of detected markings. The algorithm uses feedback from spatial relationships and geometric constraints to correct false detections and improve overall accuracy.
Solution Approach 2:
The patent changes the parameters used for lane marking detection from raw sensor data to processed features including spatial relationships, geometric consistency, and logical validity. By transforming detection criteria to include multiple validation dimensions, the system reduces false positives while maintaining detection coverage.
3Productivity
If existing image recognition algorithms are used for lane marking detection, then detection speed is acceptable, but completeness deteriorates due to occlusions by trees, vehicles, and faded markings
Solution Approach 1:
The system performs preliminary actions by pre-processing image data to enhance lane marking visibility and prepare the data for unsupervised learning analysis. This includes preprocessing to improve contrast, reduce noise, and highlight potential lane markings before the main detection algorithm processes them.
Solution Approach 2:
The patent introduces an intermediary processing layer between raw sensor data and final lane marking detection. The unsupervised learning algorithm acts as an intermediary that processes raw detections, identifies logical inconsistencies, and generates corrected lane marking data, thereby filling gaps caused by occlusions.
4Measurement precision
If manual verification of lane markings is performed, then data accuracy improves, but processing time and computational resources increase significantly
Solution Approach 1:
The system eliminates the need for manual verification by implementing self-service through unsupervised learning algorithms that automatically detect, validate, and correct lane marking data. The algorithm independently performs accuracy checks and corrections that would otherwise require human intervention.
Solution Approach 2:
The patent substitutes manual verification processes with automated computational algorithms that perform accuracy checking and data completion. The unsupervised learning model replaces human experts in validating lane marking detections, significantly reducing processing time while maintaining high accuracy.
Data Source
AI summary
A system is disclosed for generating a feature line, such as for lane marking data of a map database. The system, for example, determines connectivity data for the feature line based on at least one first point of the feature line and at least one second line, which is within a maximum distance from the at least one first point. Further, orientation data associated with the at least one first point and at least one second point associated with the at least one second line is generated, in response to determining the connectivity data for the feature line. Further, distance data associated with distance between the at least one first point and the at least one second point is determined to generate a connecting line for the feature line based on the at least one first point, the at least one second point, the orientation data and the distance data.