Multi-Camera Video Tracking System Using Geometry Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional security mechanisms face difficulties in manually tracking moving targets across multiple video feeds, especially in crowded areas with obstacles, leading to loss of targets and increased operational costs due to the need for extensive personnel training and resources.
Innovation Solution
A system utilizing a geometry engine to calculate the nearest and best video devices for tracking, combining video feeds into a single user interface, and allowing for seamless transition between video devices as the target moves, with features for recording and replaying video to track targets backward or forward in time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual tracking of moving targets across multiple video feeds is performed, then operator experience and knowledge are required to track targets accurately, but this leads to increased training time and operational errors
Solution Approach 1:
The system automatically performs target tracking across multiple video feeds without requiring operator intervention. The automated target recognition and tracking system independently identifies, tracks, and monitors targets across different camera feeds, eliminating the need for operators to manually follow targets while maintaining high accuracy.
Solution Approach 2:
The patent replaces the mechanical manual tracking process with an automated electronic system. Instead of operators visually tracking targets across multiple video feeds, the system uses computer vision algorithms, image processing, and automated recognition to perform the same function, thereby eliminating training requirements and reducing operational errors.
2Area of stationary object
If multiple security cameras are deployed to cover all areas, then comprehensive surveillance coverage is achieved, but the complexity of monitoring all cameras simultaneously increases
Solution Approach 1:
The system segments the surveillance task by automatically identifying and isolating specific areas of interest (such as targets or objects of concern) from the overall video feeds. Instead of requiring operators to monitor all cameras simultaneously, the system divides the monitoring function into automatic target detection, individual target tracking, and relevant area highlighting, thereby reducing operational complexity while maintaining comprehensive coverage.
Solution Approach 2:
The patent introduces an automated target recognition and tracking system as an intermediary between the multiple cameras and the operator. This intermediary automatically processes video feeds from all cameras, identifies targets, and presents only relevant information to the operator, thereby reducing the complexity of monitoring multiple cameras while maintaining complete surveillance coverage.
3Reliability
If automated target recognition and tracking systems are implemented, then manual tracking errors are reduced, but the cost of deploying and maintaining the system increases
Solution Approach 1:
The automated target recognition and tracking system is designed to perform multiple functions: target detection, target tracking across video feeds, real-time monitoring, and historical data analysis. By consolidating these functions into a single integrated system, the patent reduces overall system complexity and cost compared to having separate systems for each function, while maintaining high tracking reliability.
Data Source
AI summary
Tracking a target across a region is disclosed. A graphical user interface is provided that displays, in a first region, video from a field of view of a main video device, and, in a plurality of second regions, video from a field of view of each of a plurality of perimeter video devices (PVDs). The field of view of each PVD is proximate to the main video device's field of view. A selection of one of the plurality of PVDs is received. In response, video from a field of view of the selected PVD is displayed in the first region, and a plurality of candidate PVDs is identified. Each candidate PVD has a field of view proximate to the field of view of the selected PVD. The plurality of second regions is then repopulated with video from a field of view of each of the plurality of identified candidate PVDs.