Lane Line Determination Using Pixel Screening and Curve Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lane line positioning in autonomous driving relies heavily on manual labeling, which is time-consuming, costly, and prone to errors, and lacks automation in accurately determining lane lines from vehicle-front images.
Innovation Solution
A method and apparatus for automatically determining lane lines in road images by preprocessing, edge detection, pixel screening, and fitting, using modules for line determination, pixel screening, and error evaluation to assess positioning accuracy, thereby reducing manual intervention and improving automation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual labeling is used to acquire true value of lane line, then positioning accuracy can be evaluated, but time consumption and cost increase significantly
Solution Approach 1:
The patent uses automated algorithms to generate virtual true values of lane lines by processing road images through line detection and pixel fitting methods. These computationally generated lane line positions serve as copies or substitutes for manually labeled true values, enabling accuracy evaluation without actual manual labeling. The system creates synthetic reference data that replicates the function of manual annotations.
2Measurement precision
If manual labeling is used to determine lane line position, then true value can be obtained, but labor cost and operation complexity increase
Solution Approach 1:
The system performs self-service by automatically determining lane line positions through a series of computational steps: edge detection, line segment extraction, pixel screening, and curve fitting. The algorithm processes road images autonomously to generate lane line true values without requiring human operators to manually annotate images. The entire pipeline from image input to true value output is automated, making the system self-sufficient.
3Productivity
If automated lane line determination is implemented, then productivity increases, but measurement precision of lane line position may deteriorate
Solution Approach 1:
The automated determination process is divided into distinct segmentation steps: (1) edge detection to identify potential lane line regions, (2) line segment extraction to isolate candidate segments, (3) pixel screening to select pixels forming the lane line, and (4) curve fitting to determine the final lane line position. Each segment focuses on a specific aspect of the problem, allowing the system to maintain high precision through specialized processing at each stage while achieving high overall productivity.
Solution Approach 2:
The system performs preliminary actions by pre-processing road images through edge detection and line segment extraction before final lane line determination. These preliminary steps prepare the data by identifying and isolating relevant features, reducing the complexity of subsequent processing. The pixel screening step also performs preliminary selection of candidate pixels, ensuring that only high-quality pixels are used in the final fitting process, thereby maintaining precision while improving efficiency.
Data Source
AI summary
A lane line determination method and apparatus, a lane line positioning accuracy evaluation method and apparatus, a device and a storage medium are provided, which are related to a field of image processing, and particularly to fields of autonomous driving, intelligent transportation, computer vision and the like. The specific implementation is: determining a line in a received road image; screening pixels forming the line and determining pixels forming a lane line; and fitting the pixels forming the lane line to obtain the lane line. According to the technology of the present disclosure, the disadvantages of manual labeling can be overcome, and the lane line in the image collected by image acquisition device can be automatically recognized using an image recognition method, thereby improving the automation degree of lane labeling.