Parking Lot Map Generation with Vehicle Icon Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for guiding vehicle drivers to available parking spaces in parking lots are inefficient due to distorted bird's-eye view images, making it difficult to identify parked vehicles and available spaces.
Innovation Solution
An information processing apparatus that acquires feature information about vehicles and objects in a parking lot using sensors, generates a parking lot map with icons representing vehicles and objects, and displays this information to the driver for easy identification of available parking spaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a bird's-eye view image is generated by converting camera images to show parking areas, then the overall parking lot layout becomes visible, but vehicle distortion occurs making it difficult to identify parked vehicles and available spaces
Solution Approach 1:
The system segments the parking lot into individual parking space units, each represented by an icon in the map view. This segmentation allows the overall layout to be visible while providing clear, undistorted representation of each parking space status (occupied/vacant), resolving the contradiction between showing the big picture and maintaining vehicle identification accuracy.
Solution Approach 2:
Instead of displaying the distorted bird's-eye view image directly, the system creates a simplified symbolic copy (icon-based map) that represents the parking lot layout, vehicles, and parking spaces. This copy eliminates the distortion problem while preserving the essential information needed for navigation and parking space identification.
2Measurement precision
If drivers directly check surroundings by driving through the parking lot to find vacant spaces, then accurate identification of available parking is achieved, but time is lost and collision risk increases
Solution Approach 1:
The system performs preliminary identification of vacant parking spaces and displays this information on the map before the driver arrives. The map shows the locations of available parking spaces in advance, allowing the driver to plan the route and arrive at the destination without unnecessary driving through the entire parking lot, thus saving time while maintaining accurate identification.
Solution Approach 2:
The system provides continuous feedback to the driver through the map display, showing real-time information about parking space availability and guiding the driver to vacant areas. This feedback mechanism eliminates the need for trial-and-error driving and direct surrounding checks, reducing both time loss and collision risk.
3Area of stationary object
If a bird's-eye view image is used to display parking areas, then a comprehensive view is provided, but the distorted image makes it difficult to determine vehicle types and correspondence to actual vehicles
Solution Approach 1:
The system applies different representation qualities to different elements: parking spaces are shown as simple geometric areas, while vehicles are represented by icons that preserve key identifying features (type, color, size). This local differentiation maintains comprehensive coverage of the parking lot while preventing information loss about vehicle characteristics.
Solution Approach 2:
The system uses color coding in the icon representations to convey vehicle feature information. Different colors indicate different vehicle types or attributes, allowing the map to display comprehensive parking lot coverage while preserving essential vehicle identification information that would be lost in a distorted bird's-eye view.
Data Source
AI summary
An image processing apparatus for this determination includes processing circuitry configured to acquire feature information about a vehicle or a person in a parking lot, e.g., information on at least one feature among a color, an orientation, a location, a size, a vehicle brand, and a vehicle type, using information detected by a sensor; and generate a parking lot map including an icon corresponding to the vehicle or the person. The icon includes information on at least one feature among a color, an orientation, a location, a size, a vehicle brand, and a vehicle type of an object such as a vehicle and a person.