Cross-Camera Obstacle Tracking With Global ID Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current unmanned 360-degree visual perception systems have low accuracy in tracking obstacles in overlapping camera areas, necessitating a high-performance circular multi-target tracking system.
Innovation Solution
A cross-camera obstacle tracking method that determines the main camera, integrates obstacles with global identifiers in overlapping areas, and performs correlation consistency detection to improve tracking accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a 360-degree visual perception system tracks obstacles in each camera independently, then the system can cover all surrounding areas, but the tracking accuracy in overlapping camera areas deteriorates
Solution Approach 1:
The patent merges tracking results from multiple cameras by establishing mapping relations between local identifiers in different cameras and global identifiers. Obstacles detected in overlapping areas are integrated into associated sets, and correlation consistency detection is performed to fuse tracking data across cameras, thereby improving tracking accuracy while maintaining comprehensive coverage.
Solution Approach 2:
The patent introduces global identifiers as an intermediary to bridge local identifiers from different cameras. The mapping relation between local and global identifiers serves as a mediator to correlate tracking results across multiple cameras, enabling accurate identification and fusion of obstacles in overlapping detection areas.
2Reliability
If multiple cameras are used to track obstacles, then the coverage and detection capability are improved, but the system complexity increases
Solution Approach 1:
The patent segments the complex multi-camera tracking problem into manageable components: (1) obtaining tracking results from each camera independently, (2) establishing mapping relations between local and global identifiers, (3) integrating obstacles in overlapping areas, and (4) performing correlation consistency detection. This segmentation reduces system complexity by breaking down the overall task into distinct processing stages.
Solution Approach 2:
The system performs self-service through automated processing of multi-camera data. The mapping relation establishment and correlation consistency detection are automatically executed without manual intervention, allowing the system to handle multiple cameras efficiently while maintaining reliability.
3Measurement precision
If obstacle tracking results from multiple cameras are integrated, then the tracking accuracy is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-establishing mapping relations between local identifiers and global identifiers before full integration. Obstacles in overlapping areas are pre-integrated into associated sets, and preliminary correlation consistency checks are performed, which reduces the processing time required for complete multi-camera data fusion.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Embodiments of the present provide a cross-camera obstacle tracking method, apparatus, device, system and medium. The method includes: obtaining (S110, S210) obstacle tracking results of images captured by at least two cameras, wherein each obstacle tracking result comprises results after identifying and tracking at least one obstacle in the image captured by each camera, and each obstacle is labeled with a local identifier for each camera; in (S120, S220) response to a triggering condition being satisfied, establishing a mapping relation between local identifiers of the obstacles in the cameras and global identifiers according to the obstacle tracking result of each of the cameras; determining (S130, S250) similarities of obstacles according to the obstacle tracking result of each of the cameras, and performing fusion of the global identifiers on the same obstacle according to the similarities; and obtaining (S140, S260) obstacle tracking results labeled with the global identifiers. Therefore, the obstacle tracking accuracy is improved.