Multi-Camera Object Tracking via Subnetwork Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current intelligent CCTV and smart remote security technologies are limited to tracking objects in single-camera environments, making it difficult to track specific objects in real-time across multiple cameras that do not overlap.
Innovation Solution
A method and system that selects a subnetwork of cameras, receives image frames, determines the camera with the object of interest, locally tracks the object, and re-identifies it across multiple cameras using local path tracking information and similarity data, allowing for real-time global tracking in a multi-camera environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If object tracking is performed in a single-camera environment, then tracking precision is maintained, but tracking coverage and ability to track objects across multiple camera zones is limited
Solution Approach 1:
The system divides the multi-camera environment into multiple subnetworks, with each subnetwork containing one or more cameras. Each subnetwork independently performs local tracking, and the server module coordinates between subnetworks to achieve global tracking across non-overlapping camera zones while maintaining identification accuracy through distributed processing.
Solution Approach 2:
The server module acts as an intermediary that receives tracking information from multiple client modules (each managing a subnetwork), performs global matching of tracking targets across subnetworks, and coordinates re-identification when objects move between camera zones. This mediator enables accurate cross-camera tracking without requiring direct camera-to-camera communication.
2Area of stationary object
If multiple cameras with non-overlapping zones are used to expand tracking coverage, then tracking coverage area increases, but system complexity increases due to coordination requirements
Solution Approach 1:
The system segments the multi-camera system into independent subnetworks, each managed by a client module. This segmentation allows each subnetwork to operate semi-independently with local tracking capabilities, reducing the coordination burden on the central server and simplifying the overall system architecture while still achieving global coverage.
Solution Approach 2:
The system performs preliminary local tracking within each subnetwork before global matching is required. Client modules pre-process image frames and generate local tracking information, so that when objects move between zones, the server module only needs to perform matching on pre-processed data rather than raw images, reducing computational complexity.
3Speed
If local tracking is performed in each camera independently, then processing speed is maintained, but ability to re-identify objects when they move between camera zones is lost
Solution Approach 1:
The server module receives local tracking information from multiple client modules and performs global matching to determine whether tracking targets are the same object across different subnetworks. This feedback mechanism enables re-identification of objects when they move between camera zones while maintaining high processing speed through distributed computation.
Solution Approach 2:
The system merges local tracking results from multiple independent subnetworks into global tracking information. By combining local path tracking information, similarity information, and multi-camera topology data, the server module achieves cross-camera re-identification while preserving the processing efficiency of independent local tracking.
Data Source
AI summary
The present disclosure relates to a method and system for tracking an object of interest in real-time in a multi-camera environment capable of performing data communication without the occurrence of overlapping, and according to the present invention, an object of interest can be tracked in real-time on the basis of mutually captured image data even though capturing zones of a plurality of cameras capable of performing data communication according to the connection therebetween are not overlapped with each other.