Parking Frame Edge Correction for Dashed Line Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing parking assist devices struggle to accurately detect parking frames due to short edges caused by light reflection or other factors being misinterpreted as dashed boundary lines, leading to incorrect detection of parking spaces.
Innovation Solution
An image processor that includes edge detection, grouping, determination, and correction components to identify and correct dashed line edges, using threshold values for brightness and color differences, and performing linear approximation to accurately set parking frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If edge detection is performed on all detected edges including short edges, then the detection coverage is improved, but false detection of parking frames increases due to misinterpretation of short edges as dashed boundary lines
Solution Approach 1:
The patent segments the edge detection process into multiple stages: initial edge detection, grouping of detected edges, pattern recognition to identify dashed line patterns, and selective correction. This segmentation allows the system to differentiate between genuine dashed boundary lines and false short edges by analyzing their spatial relationships and patterns, thereby improving detection accuracy while maintaining reliability
Solution Approach 2:
The patent performs preliminary grouping and pattern recognition on detected edges before final parking frame determination. By预先 (in advance) organizing edges into groups based on their spatial relationships and identifying dashed line patterns, the system can filter out false short edges before they affect the final parking frame detection, thus resolving the contradiction between comprehensive detection and accurate filtering
2Manufacturing precision
If linear approximation is applied to correct endpoint coordinates of dashed line edges, then the precision of parking frame setting is improved, but the complexity of image processing increases
Solution Approach 1:
The patent applies linear approximation correction only to specific endpoint coordinates of dashed line edges where correction is needed, rather than uniformly processing all edges. This localized approach improves parking frame setting precision for critical areas while minimizing unnecessary processing complexity, as the correction is applied selectively based on the identified dashed line patterns and their geometric properties
Data Source
AI summary
An image processor includes an edge detection portion for scanning an image and detecting, as edges, an arrangement of pixels in which brightness value difference or color parameter difference between the pixels is equal to or greater than a threshold; a grouping portion for grouping the detected edge based on edge length, a distance between endpoints of the edges, and an angle between the edges; a determination portion for determining the grouped edges as a dashed line edge group when a pattern, in which the brightness value difference or the color parameter difference between the pixels is detected, matches a predetermined pattern; a correction portion for performing a linear approximation process on the dashed line edge group to correct a coordinate value of an endpoint of the dashed line edge; and a parking frame setting portion for setting a parking frame using the corrected dashed line edge.