Multi-Camera Target Tracking via Candidate Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-camera surveillance systems often fail to provide complete coverage, making it difficult to track targets as they move between camera fields of view, especially in crowded environments, and require human intervention for effective target recognition and tracking.
Innovation Solution
A multi-camera surveillance system with a graphical user interface that allows users to interactively track targets by selecting primary and secondary cameras based on likelihood and proximity, generating target samples, and searching secondary cameras for matching video snippets to create a chronological playlist of the target's path.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple cameras are installed throughout a facility to increase coverage, then the area covered by cameras increases, but the complexity of monitoring and tracking targets across multiple cameras increases significantly
Solution Approach 1:
The system divides the facility into multiple camera zones and segments the tracking task by maintaining separate target samples for each camera. Each camera's video feed is processed independently to extract target samples, which are then correlated across camera boundaries to reconstruct continuous target paths through the facility.
Solution Approach 2:
The system introduces an intermediary processing layer that correlates target samples across multiple cameras. This intermediary layer matches targets appearing in one camera with targets in other cameras based on spatial relationships and temporal continuity, automatically reconstructing target paths without requiring manual monitoring of all cameras simultaneously.
2Productivity
If automated target recognition is implemented to reduce manual monitoring, then productivity increases, but the system becomes less effective in crowded environments and without human intervention
Solution Approach 1:
The system performs partial automated recognition by extracting target samples from video feeds and performing automated matching across cameras. However, it allows user interaction to review, confirm, or correct target identifications, combining automated efficiency with human judgment to maintain reliability in crowded or complex environments.
Solution Approach 2:
The system implements feedback mechanisms where user interactions with target samples (confirmations, corrections, or rejections) are used to refine and improve the automated recognition process. This feedback loop allows the system to learn from user decisions and improve its target tracking accuracy over time while maintaining high productivity.
3Ease of operation
If users manually monitor multiple cameras simultaneously to track a target, then target tracking is possible, but the time required to locate and track targets increases significantly
Solution Approach 1:
The system performs preliminary automated actions by continuously extracting target samples from all camera feeds and pre-processing video data. When a user initiates a tracking request, the system has already prepared target samples and can quickly present relevant candidates, eliminating the need for users to manually scan through multiple cameras to locate targets.
Data Source
AI summary
Tracking of targets in video captured by a multi-camera surveillance system is often difficult due to incomplete camera coverage and due to the complexities associated with automated recognition in a dense and highly variable environment. The present disclosure embraces a system and method for target tracking in a multi-camera surveillance system that reduces the time required for a user to track a target by offering computer generated search results that include suggestions of candidates found in the video that match (to some degree) the target. When one of the candidates is selected by a user, the search is shifted to other cameras and continued. In this way, the target may be tracked quickly. Further, a playlist of video snippets of the target is accumulated during the search and can be played to show all captured video of the target moving about the facility.