Road Marking Semantic Vectors for Robust Vehicle Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle positioning methods for intelligent driving face challenges such as poor accuracy due to illumination variations affecting feature points, and high computational complexity from storing excessive invalid information, which hinders efficient processing.
Innovation Solution
A visual semantic vector-based method that classifies pixel points in road images, projects them to a ground coordinate system, and determines semantic vectors for road surface marking, using a pre-trained neural network to partition pixel sets and perform PCA for robust guidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If feature point method is used for vehicle positioning, then positioning function can be implemented, but accuracy deteriorates due to illumination variations affecting feature points
Solution Approach 1:
The patent extracts only the essential semantic information (semantic vectors) from road images rather than using all feature points. By selecting and extracting only the most relevant semantic features for positioning, the system avoids the illumination sensitivity of traditional feature point methods while maintaining positioning functionality.
Solution Approach 2:
The patent changes the parameter representation from traditional feature point coordinates to semantic vectors that encode both position and category information. This parameter transformation makes the positioning system more robust to illumination variations by using semantic categories that are less sensitive to lighting changes.
2Loss of information
If dense semantic point clouds are generated based on semantic segmentation, then comprehensive semantic information is obtained, but storage resources are excessively consumed
Solution Approach 1:
The patent extracts only the necessary semantic information (semantic vectors representing road surface markings and their positions) rather than generating complete dense semantic point clouds. This selective extraction maintains the essential semantic content needed for positioning while dramatically reducing storage requirements by filtering out redundant information.
Solution Approach 2:
The patent segments the semantic information into discrete semantic vectors corresponding to specific road surface markings rather than storing continuous dense point clouds. This segmentation approach preserves the critical semantic categories and positions while reducing overall data volume by focusing only on relevant road features.
3Loss of information
If dense semantic point clouds are stored, then semantic information is preserved, but processing efficiency deteriorates due to excessive data volume
Solution Approach 1:
The patent extracts and stores only the essential semantic vectors needed for positioning rather than preserving all semantic information in dense point clouds. This extraction strategy maintains the critical semantic content for road surface marking identification while significantly improving backend processing efficiency by reducing data volume.
Solution Approach 2:
The patent segments semantic information into discrete, processed semantic vectors that are ready for immediate use in positioning calculations. This pre-segmentation and vectorization approach eliminates the need for complex processing of dense point clouds, thereby enhancing backend processing speed and efficiency.
Data Source
AI summary
A visual semantic vector-based vehicle guidance method and system, a device, and a medium are provided. The method includes: acquiring a road image, and classifying pixel points in the road image to obtain pixel point categories; performing point set partitioning according to pixel point positions and categories to obtain a plurality of pixel point sets, each pixel point set consisting of the pixel points with continuous positions and the same category; projecting the pixel points in each pixel point set to a ground coordinate system to obtain three-dimensional coordinate values of the pixel points in each pixel point set; determining a semantic coordinate and a direction of a corresponding pixel point set as a semantic vector of the pixel point set according to the three-dimensional coordinate value of each pixel point; and road surface marking positioning is performed according to the semantic vector to guide a vehicle to travel. Therefore, the robustness of the semantic vector can be enhanced and reliable data support can be provided for subsequent vehicle positioning.