Lane Recognition Device Using Cumulative Voting for Stud Marks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lane recognition devices struggle to accurately recognize stud-shaped lane marks due to limited edge features, noise sensitivity, and increased calculation and memory requirements, especially at high speeds and when vehicle orientation is not parallel to the lane.
Innovation Solution
A lane recognition device that extracts candidate positions from input images and uses cumulative voting to generate a feature image, weighting vote values by elapsed time to determine lane boundary positions, reducing noise sensitivity and computational demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If template matching method is used to detect stud-shaped lane marks, then detection capability is improved, but noise sensitivity increases and false detections occur
Solution Approach 1:
The patent combines multiple detection results from different time points through cumulative voting. Detection results from sequential images are integrated by accumulating vote values at corresponding parameter space locations, which merges multiple observations to improve reliability and reduce noise-induced false detections.
Solution Approach 2:
The patent implements feedback through iterative cumulative voting where detection results from previous time steps influence current detection. The system continuously refines lane mark detection by feeding back accumulated evidence from multiple frames, allowing the detection accuracy to improve over time while filtering out transient noise.
2Reliability
If multiple time-series images are composed to detect stud-shaped lane marks, then detection reliability is improved, but calculation amount and memory consumption increase
Solution Approach 1:
The patent extracts only the essential parameter space coordinates from multiple images rather than processing and storing complete image data. By extracting and voting only the relevant parameter values (position, orientation) at each detection point, the system achieves reliable detection while minimizing calculation and memory requirements.
Solution Approach 2:
The patent transforms the detection problem from image space to parameter space through Hough transformation. This dimensionality change allows cumulative voting to be performed in the parameter space domain rather than requiring composition and processing of multiple full-resolution images, significantly reducing computational complexity while maintaining detection reliability.
3Measurement precision
If Hough conversion is used to extract straight lines from lane mark images, then lane boundary detection is improved, but stud-shaped lane marks cannot be recognized well due to limited edge features
Solution Approach 1:
The patent performs preliminary detection of candidate lane mark positions using template matching or feature detection before applying Hough conversion. This preliminary action identifies potential stud-shaped lane mark locations, which then serve as seed points for cumulative voting, enabling Hough conversion to effectively detect lane boundaries even when edge features are limited.
Solution Approach 2:
The patent introduces cumulative voting as an intermediary mechanism between initial feature detection and final lane boundary determination. This intermediary process accumulates evidence from multiple detection attempts and time points, bridging the gap between limited edge features in individual frames and reliable lane boundary detection through aggregated voting results.
Data Source
AI summary
The recognition of a lane boundary line expressed by a stud-shaped lane mark is realized with high accuracy against noise and with a small calculation amount and a small memory consumption amount.A lane recognition device extracts a candidate position of a lane mark from a received input image. A vote value is weighted according to the elapsed time with regard to the extracted candidate position of the lane mark to cumulatively perform voting to a parameter space of an approximate curved line or an approximate straight line, and, thus, to generate a time-series cumulative voting feature image. A candidate of a lane boundary line position is extracted based on the generated time-series cumulative voting feature image to determine the lane boundary line position. The lane recognition device may be provided with an image output device which images an image and outputs the imaged image.