Autonomous Parking Line Detection Using Spatial Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous parking systems face challenges in accurately identifying and navigating parking spaces, particularly in complex environments, due to limitations in object recognition and noise filtering in vehicle images.
Innovation Solution
A vehicle system utilizing a camera to acquire images, derive spatial recognition data through deep learning, and a controller to extract feature points, filter noise, and determine reliable candidate parking lines, enabling precise autonomous parking by clustering and classifying objects, and controlling the vehicle to park within identified areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrasonic signals are used for autonomous parking, then the system can detect parking spaces, but the measurement precision and reliability are insufficient in complex environments
Solution Approach 1:
The patent combines multiple sensing modalities (ultrasonic signals, camera images, and spatial recognition data) into a unified parking detection system. The controller integrates data from these different sources to derive feature points and determine parking spaces, thereby improving both measurement precision and reliability compared to using ultrasonic signals alone.
Solution Approach 2:
The patent introduces spatial recognition data as an intermediary that bridges the gap between raw camera images and parking space identification. This intermediate processing layer extracts meaningful features and enhances the reliability of parking detection by providing additional contextual information beyond what ultrasonic signals or camera images alone can provide.
2Adaptability or versatility
If camera images are used for autonomous parking, then visual recognition capability is improved, but noise and interference in images reduce measurement precision
Solution Approach 1:
The patent extracts relevant feature points from camera images by applying spatial recognition algorithms and filters. This extraction process separates meaningful parking space information from visual noise and interference, thereby maintaining the adaptability benefits of camera-based recognition while improving measurement precision through selective feature extraction.
Solution Approach 2:
The patent transforms raw camera image data into spatial recognition data through parameter changes in the processing domain. By converting visual information into structured spatial features and applying filtering operations, the system enhances measurement precision while preserving the versatile visual recognition capabilities of the camera system.
3Measurement precision
If multiple feature points are derived from both image and spatial recognition data, then parking line identification accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the feature point derivation process into distinct stages: extracting feature points from camera images, extracting feature points from spatial recognition data, and then clustering these feature points to identify parking lines. This segmentation manages system complexity by breaking down the complex processing task into manageable sub-tasks while maintaining high identification accuracy through the comprehensive use of multiple feature sources.
Data Source
AI summary
A system for performing autonomous parking of a vehicle includes: a camera configured to acquire a surrounding image of the vehicle including a parking line; and a controller configured to derive spatial recognition data based on the surrounding vehicle image as an input, derive a feature point corresponding to the parking line based on the surrounding image and the spatial recognition data, determine a candidate parking line based on clustering of the feature point, and control the vehicle to perform parking in a parking area having the candidate parking line.