Road Division Line Recognition Using Adaptive Local Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately recognizing road division lines, particularly when luminance values of division lines are indistinguishable from those of the road surface, especially under conditions like shadows or strong reflections.
Innovation Solution
A division line recognition apparatus that sets a processing area, calculates statistics, and uses multiple threshold values to classify pixels and extract feature points, distinguishing between road surfaces and division lines, and decides the division line based on these features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single binarization threshold value is used to detect lane marks, then the detection process is simple, but the accuracy decreases when luminance values of division lines are indistinguishable from road surface under shadows or strong reflections
Solution Approach 1:
The patent divides the captured image into multiple blocks and calculates luminance statistics for each block separately. This segmentation allows the system to adapt threshold values to local luminance conditions, improving division line recognition accuracy in areas with shadows or reflections while maintaining manageable processing complexity through localized rather than global processing.
Solution Approach 2:
The patent applies different binarization threshold values to different blocks of the image based on their local luminance characteristics. By calculating the average luminance of each block and setting threshold values relative to these local averages, the system achieves adaptive detection that accounts for local variations in lighting conditions, thereby improving recognition accuracy without requiring a single complex global threshold.
2Reliability
If luminance value alone is used to identify division lines, then the processing is fast, but the reliability decreases when division line luminance is similar to road surface luminance
Solution Approach 1:
The patent changes the parameter used for threshold determination from a fixed value to a dynamic value based on the average luminance of each image block. By setting the binarization threshold as a function of local average luminance (e.g., threshold = average luminance + offset), the system adapts to varying lighting conditions and maintains reliable division line detection even when absolute luminance values are similar between road surface and division lines.
3Measurement precision
If the entire image is processed with a uniform threshold, then the processing is simple, but the measurement precision decreases in areas with varying luminance conditions such as shadows and reflections
Solution Approach 1:
The patent segments the image into multiple blocks and processes each block independently with its own threshold value. This approach improves feature point detection precision in areas with varying luminance by adapting to local conditions, while the block-based processing keeps the computational load manageable compared to pixel-by-pixel analysis of the entire image.
Solution Approach 2:
The patent applies processing to multiple blocks rather than the entire image at once, which allows for more precise local detection without the full computational cost of processing every pixel with adaptive thresholds. The block-based approach provides a balance between precision and computational efficiency.
Data Source
AI summary
An apparatus for recognizing a division line on a road from an image captured by a camera includes: a processing area setting unit to set a processing area to the image; a statistics calculation unit to calculate statistics of the image in the processing area; a threshold value setting unit to set a plurality of threshold values on the basis of the statistics; a division line feature point extraction unit to classify a plurality of pixels contained in the image on the basis of the white line threshold value and the road surface threshold value and extracts feature points of the division line on the basis of classification results of the plurality of pixels; and a division line decision unit configured to decide the division line on the basis of the feature points extracted by the division line feature point extraction unit.