Drivable Area Detection Using Corrected Obstacle Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic driving systems face challenges in accurately and smoothly detecting drivable areas due to inconsistencies in static boundary detection and dynamic trajectory prediction, leading to instability in drivable area boundaries across consecutive frames.
Innovation Solution
A method and apparatus for detecting drivable areas that corrects historical trajectories and position information using map information, key point comparisons, and type-specific correction parameters, improving the smoothness and accuracy of static and dynamic obstacle detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional drivable area detection is used without correction, then the detection process is simple and fast, but the accuracy and smoothness of boundary detection deteriorate
Solution Approach 1:
The patent applies preliminary action by correcting historical trajectories and position information before using them for drivable area detection. The system pre-processes detection data from multiple frames, adjusting trajectories of dynamic obstacles and positions of static obstacles based on map information and key point comparisons. This preliminary correction ensures that the data used for boundary determination is already optimized, improving detection accuracy without adding complexity during the main detection process.
Solution Approach 2:
The patent implements feedback by using map information and key point comparisons to verify and correct detected positions and trajectories. The system continuously compares detected key points against map data and historical information, making iterative corrections to trajectory and position data. This feedback mechanism ensures that detection results converge toward accurate values, resolving the contradiction between simplicity and precision.
2Stability of the object's composition
If trajectory correction based on map information is applied, then the smoothness of static boundaries improves, but the computational time and processing complexity increase
Solution Approach 1:
The patent applies segmentation by dividing the correction process into discrete key points along trajectories and boundaries. Instead of correcting entire continuous paths at once, the system identifies and corrects specific key points (such as intersection points, turning points, and boundary markers) independently. This segmentation allows parallel processing and reduces computational complexity while maintaining boundary smoothness, as each key point can be corrected using local map information without affecting the entire trajectory.
Solution Approach 2:
The patent implements local quality by applying different correction strategies to different parts of the detection data based on their specific characteristics. Static obstacles near key points receive correction based on map information, while dynamic obstacles have their trajectories corrected using a combination of historical data and current frame information. The system adjusts the degree of correction locally, applying stronger constraints in areas requiring high precision (such as boundary regions) and more flexible correction in open areas, thereby optimizing processing efficiency.
3Measurement precision
If multiple video frames are analyzed for trajectory determination, then the accuracy of dynamic obstacle trajectory improves, but the detection complexity and data processing load increase
Solution Approach 1:
The patent applies preliminary action by pre-processing multiple video frames to extract and store trajectory information before the main detection process. The system analyzes historical frames in advance, establishing baseline trajectories and key point positions that can be referenced during real-time detection. This pre-extraction of trajectory data from multiple frames reduces the computational burden during actual drivable area calculation, as the system only needs to refine rather than compute trajectories from scratch.
Solution Approach 2:
The patent implements partial action by selectively processing only the necessary portions of multiple video frames. Instead of analyzing every pixel and object in all historical frames, the system focuses on extracting trajectories of relevant dynamic obstacles and positions of static obstacles near key points. The system processes excessive data (multiple frames) but only performs detailed analysis on the partial subset that directly impacts drivable area boundaries, thereby maintaining trajectory accuracy while controlling processing complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and an apparatus for detecting a drivable area, and a storage medium are provided. The method includes: collecting N consecutive video frames of a road when a vehicle is driving, where N is a positive integer greater than 1; determining a historical trajectory of a dynamic obstacle and position information of a static obstacle included in the N consecutive video frames by analyzing the N consecutive video frames; correcting the historical trajectory and the position information based on a preset rule; determining a predicted trajectory of the dynamic obstacle based on the corrected historical trajectory; and determining the drivable area of the vehicle based on the predicted trajectory and the corrected position information. The invention may improve the smoothness of static boundaries and the accuracy of dynamic trajectories, improve the effect of detecting the drivable area and improve the safety performance of automatic driving.