Lane Marking Geometry Extraction for Automated HD Map Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional map building procedures for high definition maps are resource intensive and costly, requiring significant human input for manual pinpointing of control points, which is neither cost efficient nor turnaround optimized.
Innovation Solution
The use of environment sensors such as camera sensor arrays and LiDAR to detect stripe-shaped objects like lane markings, which are then used for automatic localization geometry generation and map updates, reducing the need for human intervention and optimizing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual pinpointing of control points is used to accurately localize map objects, then manufacturing precision is improved, but productivity deteriorates
Solution Approach 1:
The patent replaces manual mechanical pinpointing operations with automated image processing and pattern recognition algorithms. The system automatically detects stripe-shaped objects (lane markings) in images, extracts their geometric features, and localizes map objects without human intervention, thereby maintaining accuracy while dramatically improving productivity
Solution Approach 2:
The system enables self-service by allowing the map building process to automatically identify and localize control points through algorithmic analysis of stripe-shaped patterns in images. The automated geometry generator independently performs localization tasks that previously required human operators, achieving both high precision and efficient throughput
2Manufacturing precision
If higher definition maps with more pixels are created, then manufacturing precision is improved, but loss of time increases
Solution Approach 1:
The patent extracts only the essential stripe-shaped objects (lane markings) from the full image data to serve as control points for localization. By focusing on these key geometric features rather than processing all pixels manually, the system achieves high-definition map quality while significantly reducing the time required for map building
Solution Approach 2:
The system performs preliminary automated detection and extraction of stripe-shaped object geometries before the actual map localization process. This pre-processing step prepares high-precision control points in advance, enabling subsequent map generation to proceed quickly with already-processed geometric data
3Device complexity
If LiDAR-only systems are used for detection, then device complexity is reduced, but measurement precision deteriorates due to calibration errors and occlusions
Solution Approach 1:
The patent creates a two-dimensional image representation (copy) of the three-dimensional scene captured by LiDAR. This image copy can be processed using robust image processing techniques to detect stripe-shaped objects, providing a complementary view that is less susceptible to LiDAR calibration errors and occlusion issues while maintaining system simplicity
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
This approach conserves both human and computer resources by automating the localization and map update processes, improving efficiency and accuracy in high definition map building, while also overcoming limitations of LiDAR-only systems such as calibration errors and traffic occlusions.
Implementation Method 1
The stripe-shaped object may include lane markings or road adjacent structures such as guardrails or barricades
Implementation Method 2
The use of environment sensors such as camera sensor arrays and LiDAR to detect stripe-shaped objects
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
Apparatus and methods are described for generating geometries for stripe- shaped objects. An image is identified that includes a roadway having one or more stripe-shaped objects. The stripe-shaped objects may include lane lines for road edges or lanes of the roadway. The stripe-shaped objects may include a barrier. At least one targeted region within the image is determined. The at least one targeted region is shaped to intersect the one or more stripe-shaped objects and includes a plurality of pixels. An image analysis is performed on the image to determine when the at least one target region includes a pixel in common with the one or more stripe-shaped objects. A geometry is constructed using the pixel in common. The geometry may be used to update a map or subsequently perform localization.