Camera-Based Parking Slot Detection with 2D Bounding Shapes

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing parking assistance systems face challenges in detecting parking slots due to variability in slot types, environmental conditions, and the need for real-time processing, with non-visual methods limited by coverage and cost, and visual methods constrained by computational complexity and limited range.

Innovation Solution

A parking assistance system using an image acquisition device and a parking slot detection module that extracts 2D bounding shapes from image data, allowing detection of parking slots with varied shapes and orientations, and optionally generating 3D projections, without requiring complex pre-processing, suitable for real-time operation on standard automotive platforms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If non-visual detection methods (ultrasonic sensors, radar, LIDAR) are used to detect parking slots, then robustness against illumination changes and straightforward data processing are achieved, but coverage is limited due to directionality and small detection area

Engineering Contradiction:
Improverobustness against illumination changesVSAvoidcoverage area
Core Design Contradiction:
ReliabilityVSArea of stationary object

Solution Approach 1:

The system divides the detection task into multiple segments by using several image acquisition devices positioned at different locations on the vehicle (front, rear, sides) to capture different portions of the surrounding environment, thereby achieving comprehensive coverage without requiring each individual sensor to have wide coverage

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system transitions from non-visual detection (ultrasonic/radar) to visual detection using cameras, moving from detecting physical signals to capturing optical images that provide richer spatial and contextual information about parking slots and surrounding environment

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If non-visual detection methods (LIDAR) are used to achieve high-precision parking slot detection, then stable and high-precision results are obtained, but vehicle cost increases significantly

Engineering Contradiction:
Improveparking slot detection precisionVSAvoidvehicle manufacturing cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The system replaces expensive LIDAR sensors with relatively inexpensive camera modules that can be mass-produced and integrated into standard automotive platforms, achieving cost-effectiveness while maintaining adequate detection precision through software-based image processing algorithms

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Solution Approach 2:

The system substitutes physical measurement systems (LIDAR, ultrasonic sensors) with optical imaging systems (cameras) combined with computer vision algorithms, replacing hardware-intensive approaches with software-intensive processing to reduce overall system cost

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Area of stationary object

If visual methods with complex pre-processing (stitching to generate top-view) are used to detect parking slots, then detection coverage is improved, but computational effort increases and real-time processing is limited

Engineering Contradiction:
Improvedetection coverageVSAvoidcomputational complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The system extracts only the essential information needed for parking slot detection directly from individual camera images without performing complex pre-processing steps like stitching multiple images to create a top-view, thereby reducing computational complexity while maintaining detection effectiveness

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system performs preliminary calibration and parameter configuration during system setup, storing detection parameters and geometric relationships in advance, so that during real-time operation only simple image processing and pattern matching are required, enabling real-time processing on standard hardware

Inventive Principle:
Principle #10Preliminary action

4Ease of manufacture

If traditional computer vision algorithms are used to detect parking slots by identifying low-level features (lines, corners), then parking slot detection is achieved, but performance degrades when markers are not clearly visible

Engineering Contradiction:
Improvealgorithm implementation simplicityVSAvoiddetection reliability under poor visibility
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The system employs deep learning models that learn optimal detection parameters and features from training data, automatically adapting to various lighting conditions, marker qualities, and parking slot configurations, thereby maintaining high detection reliability across diverse environmental conditions without requiring manual parameter adjustment

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4607472A1Parking assistance system and method for operating a parking assistance system
Publication Date: 2025.08.27 FICOSA AUTOMOTIVE S L U
  • EP4607472A1 patent drawingFigure 1~3B
  • EP4607472A1 patent drawingFigure 4A~6
  • EP4607472A1 patent drawingFigure 7~9

AI summary

The present disclosure relates to a parking assistance system (5) for a vehicle (10). The parking assistance system (5) comprises an image acquisition device (20) configured to acquire image data (21) and a parking slot detection module (31) configured to process the acquired image data (21) to extract one or more 2D bounding shapes (25) to demarcate corresponding parking slots (11) in the image data (21). The 2D bounding shapes (25) are formed by a finite number of segments intersecting at corresponding vertices to enclose a close region. At least one vertex angle is greater than 90°. The present disclosure further relates to methods for operating a parking assistance system (5).