Multi-Rank Target Tracking for Crowded Multi-Camera Scenes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing target tracking methods in computer vision are prone to errors due to viewing angles, light conditions, and crowded targets, leading to a high risk of losing track of targets, particularly in multi-camera scenarios.
Innovation Solution
A target tracking method that includes acquiring a video stream, performing target detection on each frame, ranking candidates based on confidence and intersection over union, and employing a multi-rank matching strategy using trackers to determine tracking trajectories, improving accuracy and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional target tracking methods are used, then the system is simple to implement, but the tracking accuracy and stability deteriorate due to viewing angles, light conditions, and crowded targets
Solution Approach 1:
The patent segments the target tracking process into multiple independent modules: detection module, ranking module, matching module, and trajectory generation module. Each module handles a specific aspect of the tracking task, allowing the system to process complex multi-camera scenarios through structured decomposition while maintaining manageable system complexity
Solution Approach 2:
The patent introduces a multi-rank matching mechanism that operates across different confidence levels and camera views. By adding the dimension of multiple ranking levels and multi-camera perspectives, the system achieves more robust tracking without proportionally increasing complexity
2Measurement precision
If simple detection methods are used, then the processing speed is fast, but the detection accuracy deteriorates leading to high risk of losing track
Solution Approach 1:
The patent performs preliminary ranking of candidate targets based on confidence scores and intersection-over-union metrics before the matching stage. This preliminary organization of detection results allows the matching module to efficiently process only the most promising candidates, maintaining high detection accuracy while avoiding exhaustive processing of all possibilities
Solution Approach 2:
The patent dynamically adjusts matching parameters and confidence thresholds based on the specific scene conditions, target characteristics, and camera configurations. By adapting parameters such as IoU thresholds and confidence levels to different scenarios, the system optimizes the balance between detection precision and processing efficiency
3Measurement precision
If single-camera tracking is used, then the system complexity is low, but the tracking accuracy deteriorates in multi-camera scenarios
Solution Approach 1:
The patent designs a universal matching framework that can handle both single-camera and multi-camera scenarios through the same core algorithms. The multi-rank matching mechanism and trajectory association logic work consistently across different camera configurations, allowing the system to achieve high tracking accuracy in multi-camera environments without requiring entirely separate processing pipelines for each scenario
Data Source
AI summary
A target tracking method, including: acquiring a video stream of a preset acquisition region; performing target detection on tth video frame of the video stream, and determining a first candidate box of the candidate target in the tth video frame; ranking the candidate targets, according to a confidence of the candidate target and a first intersection over union between the first candidate boxes of respective candidate targets, and determining a candidate target in at least one matched rank; performing at least one matching on the candidate target in the at least one matched rank, respectively, according to a set of trackers of the video stream in (t−1)th state and a preset matching strategy, and determining a target tracking result of the tth video frame; and determining a tracking trajectory of the target in the video stream according to the target tracking results of the T video frames.