Parking Space Identification Using Corner Pairing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current parking space identification systems for driverless vehicles rely on obstacles or adjacent parking spaces, limiting their application and accuracy in identifying isolated or single parking spaces.
Innovation Solution
A marked parking space identification method and system that uses an image capturing device and a parking space identification module to detect corners, pair adjacent corners into groups, and determine the presence of a parking space based on these groups, capable of identifying single and multiple parking spaces, including parallel, perpendicular, and angled parking arrangements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If distance sensors are used to detect parking spaces, then detection capability is improved, but obstacles of certain heights must be present on both sides of the parking space
Solution Approach 1:
The patent replaces the mechanical/distance-based sensing system with an optical image recognition system. Instead of using ultrasonic or radar distance sensors that require physical obstacles, the system uses image capturing devices to detect line markings on the ground, enabling identification of parking spaces without requiring obstacles of certain heights on both sides.
Solution Approach 2:
The patent introduces line markings on the ground as an intermediary element for parking space identification. These markings serve as visual cues that the image recognition system can detect, replacing the need for physical obstacles while still providing reliable parking space boundaries.
2Measurement precision
If current image recognition method is used to identify line markings, then adjacent concatenated parking spaces can be identified, but single isolated parking spaces cannot be identified
Solution Approach 1:
The patent segments the parking space identification process into distinct steps: first identifying line markings, then detecting intersections of vertical lines, and finally determining parking spaces based on these intersections. This segmentation allows the system to handle both single isolated parking spaces and adjacent concatenated parking spaces by treating each detected intersection pattern as a potential parking space indicator.
3Reliability
If obstacles are required for distance sensor detection, then detection reliability is improved, but system complexity and limited application scenarios increase
Solution Approach 1:
The patent replaces the complex distance sensor system that requires obstacle detection and height verification with a simpler image recognition system that detects line markings on the ground. This substitution reduces system complexity while maintaining or improving detection reliability across diverse parking scenarios.
Data Source
AI summary
A marked parking space identification system and a marked parking space identification method are disclosed. The marked parking space identification method includes: receiving a first image of a first parking space with markings; identifying a plurality of corners in the first image of the first parking space with markings; pairing up two adjacent corners to become a parking corner group; and using the parking corner group to identify a first marked parking space.