Rail Detection Region Shifting to Avoid Shadow False Positives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional rail detection systems erroneously detect rails due to obstacles such as shadows because they rely on brightness gradients, leading to inaccurate rail detection.
Innovation Solution
A rail detection method that divides the image into overlapping regions, sets starting points based on camera parameters, and iteratively refines rail segments by shifting regions and redesignating starting points to exclude obstacles, ensuring accurate rail detection even in the presence of shadows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If brightness gradient is used to detect rails, then detection speed is improved, but detection accuracy deteriorates due to false detection of shadows and obstacles
Solution Approach 1:
The patent divides the image into multiple search regions and further segments each region into candidate segments based on brightness gradients. By segmenting the detection process into multiple stages (initial search region division → candidate segment identification → rail segment selection), the system maintains detection speed while improving accuracy through progressive filtering of false positives like shadows
Solution Approach 2:
The patent applies different detection criteria to different regions. Brightness gradient-based candidate segmentation is applied locally within each search region, while additional verification criteria (such as rail width consistency, continuity checks, and comparison with adjacent regions) are applied specifically to distinguish true rail segments from shadow artifacts. This localized application of quality filters improves detection accuracy without sacrificing overall detection speed
2Device complexity
If search region is narrowed by brightness gradient, then detection complexity is reduced, but false detection of rails increases due to obstacles
Solution Approach 1:
The patent performs preliminary division of the image into search regions and preliminary identification of candidate segments based on brightness gradients before final rail detection. This preliminary action reduces the search space and complexity for subsequent processing while maintaining reliability by ensuring that all potential rail segments are captured in the initial candidate set, preventing false negatives
Solution Approach 2:
The patent incorporates feedback mechanisms where detected rail segments from one region inform the search and detection process in adjacent regions. Detection results are verified against multiple criteria including continuity, width consistency, and brightness gradient patterns. This feedback loop allows the system to correct false positives identified in earlier stages while maintaining reduced detection complexity through intelligent region-based processing
Data Source
AI summary
A rail detecting device includes: a starting-region rail-starting-point redesignating unit that shifts a starting region for detecting rails to set a modified starting region and designates modified starting points of two rails at the lower edge of the modified starting region when portions of the two rails are not detected based on rail segments in an image; a rail-candidate detecting unit that detects multiple lines in the modified starting region as multiple modified rail candidates, the lines being candidates of edges of the two rails; a starting-region-rail assigning unit that assigns modified rail candidates whose lowest ends are closest to the corresponding modified starting points out of the multiple modified rail candidates as tentative modified rail segments; and a starting-region-rail disappearance determining unit that establishes the tentative modified rail segments as modified rail segments when distances between the lowest ends of the tentative modified rail segments.