Parking Line Recognition Using Luminance Checks Against Shadows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional parking assistance devices inaccurately detect boundary lines due to shadows on the road surface, causing color blurring that leads to erroneous recognition.
Innovation Solution
A parking assistance device that includes a candidate line detection unit, a boundary line recognition unit, and a movement control unit, which utilize luminance and width thresholds, as well as pixel resolution criteria, to differentiate between actual boundary lines and shadow edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing is used to detect boundary lines, then the detection process is simple, but shadow edges are erroneously recognized as boundary lines due to color blurring
Solution Approach 1:
The patent segments the boundary line detection process into two distinct stages: candidate line detection (simple edge detection) and boundary line recognition (luminance difference verification). This segmentation allows the system to first identify potential lines and then filter out false positives like shadow edges by checking luminance differences across the candidate line, thereby improving accuracy without overwhelming complexity
Solution Approach 2:
The patent introduces an intermediary verification step (luminance difference calculation) between candidate detection and final boundary line recognition. This intermediary process acts as a filter that checks whether the luminance difference across a candidate line is below a threshold, effectively distinguishing true boundary lines from shadow edges without requiring complex image processing algorithms
2Measurement precision
If luminance difference verification is performed for all candidate lines, then detection accuracy improves, but processing time increases
Solution Approach 1:
The patent performs preliminary candidate line detection using simple edge detection algorithms before applying the more computationally intensive luminance difference verification. By pre-identifying only those lines that meet basic edge criteria, the system reduces the number of candidates requiring detailed verification, thus minimizing processing time while maintaining accuracy
Solution Approach 2:
The patent applies the luminance difference verification selectively only to candidate lines that meet certain criteria (such as orientation relative to the vehicle or position in the image), rather than uniformly to all detected edges. This localized application of the verification process reduces overall computation time while maintaining detection accuracy for critical areas
Data Source
AI summary
A parking assistance device includes: a candidate line detection unit that detects a candidate line that is a candidate for a boundary line in image data obtained by imaging a periphery of a vehicle; and a boundary line recognition unit that executes a determination process to determine whether or not a difference in luminance between areas on an outside of the candidate line in a width direction in the image data, the areas being on both sides of the candidate line, is equal to or larger than a luminance threshold, does not recognize the candidate line as a boundary line when determining that the difference in luminance is equal to or larger than the luminance threshold, and recognizes the candidate line as a boundary line when determining that the difference in luminance is less than the luminance threshold.