Automatic Parking Obstacle Classification for Accurate Maneuvering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic parking systems face challenges in accurately recognizing obstacles in parking spaces, leading to inaccurate parking and potential vehicle damage due to insufficient obstacle analysis.
Innovation Solution
An automatic parking method that utilizes image processing and semantic segmentation to determine the size and category of obstacles, adjusting parking strategies based on these factors to improve recognition accuracy and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional obstacle detection methods are used in automatic parking, then the system is simple to implement, but the obstacle recognition accuracy is insufficient leading to inaccurate parking
Solution Approach 1:
The patent segments the obstacle recognition process into multiple stages: initial obstacle detection, classification into different types (soft obstacles like grass/flowers vs. hard obstacles like branches/leaves), and size-based categorization. This segmentation allows the system to apply different processing strategies to different obstacle types, improving overall recognition accuracy while managing system complexity through modular processing.
2Manufacturing precision
If all obstacles are treated uniformly in the parking process, then the control logic is simple, but the parking accuracy is reduced due to inability to differentiate between obstacle types
Solution Approach 1:
The patent applies local quality by treating different obstacle types with different control strategies. Soft obstacles (grass, flowers) are ignored in the parking process, while hard obstacles (branches, leaves) require maintaining specific distances. This differentiated approach improves parking accuracy by matching the control strategy to the specific characteristics of each obstacle type rather than applying a uniform approach.
Solution Approach 2:
The system changes the control parameter (distance to maintain) based on the obstacle type and size. For small hard obstacles, a smaller distance is maintained, while for larger hard obstacles, a greater distance is maintained. This parameter adjustment based on obstacle characteristics enables precise parking control tailored to each situation.
3Measurement precision
If detailed obstacle analysis is performed for all objects, then parking accuracy is improved, but computing overheads increase
Solution Approach 1:
The patent extracts and processes only the critical features of obstacles that are necessary for safe parking decisions. Instead of performing comprehensive analysis on all detected objects, the system extracts key attributes such as obstacle type (soft vs. hard) and size category, then uses this extracted information to make parking decisions. This selective feature extraction maintains high recognition accuracy while reducing unnecessary computational overhead.
Solution Approach 2:
The system performs partial analysis by focusing only on the most relevant obstacle characteristics needed for safe parking. Rather than analyzing every aspect of each obstacle in detail, the system performs sufficient analysis to categorize obstacles into the necessary groups (soft/hard, small/large) and then applies appropriate control strategies, avoiding excessive processing of irrelevant details.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
This application provides an automatic parking method and apparatus, and a vehicle, and relates to the field of intelligent driving. The method includes: obtaining a first image; recognizing the first image, to determine a size of an obstacle and a target category to which the obstacle belongs; and determining a parking solution based on the size of the obstacle and the target category to which the obstacle belongs. In technical solutions of this application, the parking solution is determined by comprehensively considering the size of the obstacle and the target category to which the obstacle belongs. This improves accuracy of obstacle recognition in an automatic parking process, and also improves accuracy of automatic parking. In this way, differentiated parking is implemented for different obstacles to further implement intelligent parking.