Detecting Degraded Ground Paint for Accurate Virtual Control Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining ground control points from images of road markings face challenges due to degradation of painted road markings, which affects the accuracy of digital map data required for location-based services like autonomous driving.
Innovation Solution
A computer-implemented method and apparatus for detecting degraded ground paint in images using semantic segmentation, binary image generation, and hole-filling techniques to identify and correct faded or damaged road markings, enabling the generation of accurate virtual ground control points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If painted road markings are used as ground control points, then the cost and time for map data collection is reduced, but the accuracy of ground control points deteriorates due to paint degradation
Solution Approach 1:
The system performs preliminary detection of paint degradation using image processing and semantic segmentation before the paint markings become completely unusable. By identifying degraded areas in advance, the system can either correct them through image processing or flag them for replacement, ensuring that ground control points maintain their accuracy throughout the mapping process
Solution Approach 2:
The patent introduces an intermediate detection and classification layer between the painted road markings and the ground control point identification process. This intermediary system uses semantic segmentation and image analysis to assess the quality of paint markings, filtering out degraded markings before they compromise the accuracy of ground control points while allowing healthy markings to be used efficiently
2Measurement precision
If traditional ground control points are used for precise alignment, then the accuracy of digital map data is improved, but the cost and complexity of data collection increases
Solution Approach 1:
The system creates a digital copy or representation of ground control points by using painted road markings captured in images. Instead of requiring physical ground control points, the system processes images to identify and validate virtual ground control points based on paint markings, thereby simplifying the data collection process while maintaining accuracy through image-based verification
Solution Approach 2:
The patent replaces the mechanical/physical system of traditional ground control points with an image processing-based system. By using semantic segmentation and computer vision algorithms to identify and validate ground control points from painted road markings in images, the system eliminates the need for physical measurement equipment and manual ground control point establishment, reducing complexity while preserving accuracy
3Speed
If degraded paint markings are not detected, then the processing speed is maintained, but the reliability of ground control points deteriorates
Solution Approach 1:
The system applies partial action by selectively processing only those image regions containing painted road markings, rather than analyzing entire images. The semantic segmentation focuses computational resources on relevant areas, and degradation detection is performed only on identified paint regions, maintaining processing speed while ensuring reliability through targeted analysis of critical areas
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An approach is provided for detecting degraded ground paint in an image. The approach, for example, involves performing semantic segmentation on the image to determine one or more pixels of the image that are classified in a ground paint category. The approach also involves generating a binary image that contains the one or more pixels of the image that are classified in the ground paint category. The approach further involves generating a hole-filled binary image by filling in the binary image to generate one or more curvilinear structures from the one or more pixels. The approach further involves determining a difference between the image and the hole-filled binary image to identify one or more degraded ground paint pixels of the image and providing the one or more degraded ground paint pixels as an output.