Automatic Parking System Using Top View Image Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional automatic parking systems face difficulties in recognizing parking spaces and adapting to various environments due to their reliance on ultrasonic sensors, which struggle to detect spaces without obstacles and are inflexible in different situations.
Innovation Solution
An automatic parking method and system that captures top view images of the surroundings, recognizes parking spaces, detects vehicle entry points using obstacle information, and calculates a parking path, even when one entry point is undetectable by generating a new entry point based on obstacle positions and parking lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrasonic sensors are used to recognize parking spaces, then obstacles can be detected, but parking spaces cannot be recognized when no obstacles are present
Solution Approach 1:
The patent combines multiple sensing methods (ultrasonic sensors for obstacle detection and camera-based vision systems for parking space recognition) into a unified automatic parking system. This merging allows the system to leverage the strengths of each method: ultrasonic sensors detect obstacles accurately, while the vision system identifies parking spaces even without obstacles, thereby resolving the contradiction between obstacle detection accuracy and parking space recognition capability.
Solution Approach 2:
The patent implements a multi-functional sensing system where the camera-based vision system serves multiple purposes: identifying parking spaces, detecting obstacles, and providing environmental context. This universal approach allows the system to recognize parking spaces in various conditions (with or without obstacles), overcoming the limitation of ultrasonic-only systems that fail when no obstacles are present.
2Device complexity
If conventional sensors are used, then simple obstacle detection is achieved, but flexibility in coping with various situations and environments is limited
Solution Approach 1:
The patent replaces traditional mechanical/contact-based sensing approaches with optical sensing (camera-based vision systems). This substitution enables the system to capture visual information about parking spaces and obstacles, providing greater environmental adaptability while maintaining relatively simple device architecture. The vision system can identify parking spaces, obstacles, and environmental features without physical contact, enhancing flexibility across various situations.
3Device complexity
If only one entry point is detected, then parking path calculation is simplified, but parking cannot be performed when one entry point is undetectable
Solution Approach 1:
The patent performs preliminary detection of multiple vehicle entry points (first and second entry points) before calculating the parking path. By identifying multiple potential entry points in advance, the system ensures that if one entry point becomes undetectable during parking execution, alternative entry points are already known, allowing the parking operation to continue reliably. This preliminary multi-point detection resolves the contradiction between path calculation simplicity and parking execution reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to an automatic parking method and system of a vehicle. An automatic parking method according to the present disclosure comprises the steps of: taking an image of the surroundings of the vehicle and converting the captured image into a top view image; recognizing a parking space on the basis of the converted top view image; sensing obstacles in the parking space surroundings; detecting a first vehicle entry point and a second vehicle entry point in the parking space for parking the vehicle using information on the detected obstacles and information on a parking line in the parking space; and calculating a parking path of the vehicle using the first vehicle entry point and the second vehicle entry point which have been detected. The present disclosure can be used to allow a vehicle to be automatically parked by recognizing a parking space on the basis of a top view image of the surroundings of the vehicle and detecting two vehicle entry points using obstacles in the surroundings of the vehicle.