Continuous Road Partition Detection Using U-Disparity Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting continuous road partitions with height, such as those involving radar and vision data fusion, face challenges with stability and computational efficiency, especially in complex environments.
Innovation Solution
A method and apparatus that utilize disparity maps and U-disparity maps to track and detect continuous road partitions by leveraging intermediate detection results from historical frames, reducing computation and improving accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If radar and vision data fusion is used to detect continuous road partitions, then material information can be obtained that cannot be obtained from gray scale images, but the computation amount becomes very large and stable results are difficult to obtain
Solution Approach 1:
The patent extracts and utilizes only the necessary material information (disparity values) from the radar data, rather than processing complete radar images. By focusing on specific disparity map characteristics and using U-disparity maps derived from stereo vision, the system obtains material information about road partitions without the computational burden of full radar-vision fusion.
Solution Approach 2:
The detection process is segmented into distinct stages: obtaining disparity maps, generating U-disparity maps, detecting road partitions in initial frames, and tracking in subsequent frames. This segmentation allows the system to process information in manageable steps, reducing overall computational complexity while maintaining detection accuracy.
2Loss of information
If radar and vision data fusion is used to detect continuous road partitions, then material information can be obtained that cannot be obtained from gray scale images, but the detection stability deteriorates due to environmental conditions and visual geometrical distortion
Solution Approach 1:
The system uses tracking of detected road partitions across multiple frames to provide feedback that stabilizes detection. By maintaining detection results from previous frames and comparing them with current frame detections, the system compensates for environmental variations and geometric distortions, improving overall detection stability while retaining material information from disparity maps.
Solution Approach 2:
The patent performs preliminary detection of continuous road partitions in initial frames to establish baseline detection results. These preliminary detections serve as a foundation for subsequent tracking, allowing the system to maintain stable detections even when environmental conditions cause variations in individual frame quality.
3Measurement precision
If detection is performed on each frame independently without using historical frame results, then detection accuracy can be maintained, but the computation amount is large and detection efficiency is low
Solution Approach 1:
The system dynamically adjusts its detection strategy based on frame sequence position. In initial frames, full detection is performed to establish accurate baseline results. In subsequent frames, the system transitions to tracking mode, dynamically reducing computational effort while maintaining accuracy through the use of intermediate detection results from previous frames.
Solution Approach 2:
Detection is performed comprehensively in initial frames to establish accurate intermediate results beforehand. These preliminary detection results are then reused in subsequent frames through tracking, allowing the system to maintain high detection accuracy while significantly improving efficiency by avoiding redundant full-detection operations in every frame.
4Measurement precision
If full detection is performed on every frame, then detection accuracy can be maintained, but the computation amount becomes very large
Solution Approach 1:
Instead of performing full detection on every frame, the system performs partial detection only when necessary (in initial frames or when tracking fails). The majority of frames use lighter tracking operations that maintain accuracy while reducing computational load, applying the principle of doing just enough detection work to maintain performance without excessive computation.
Solution Approach 2:
Comprehensive detection is performed in initial frames to establish accurate intermediate results that can be reused. This preliminary full-detection action provides a foundation for subsequent efficient tracking, allowing the system to maintain high accuracy while avoiding the need to perform equally intensive detection on every subsequent frame.
Data Source
AI summary
A method and an apparatus for detecting a continuous road partition with a height that includes obtaining disparity maps having the continuous road partition, and U-disparity maps corresponding to the disparity maps; obtaining an intermediate detection result of the continuous road partition detected from the U-disparity maps of first N frames; and detecting the continuous road partition from the U-disparity map of a current frame, based on the obtained intermediate detection result.