Multi-Camera Object Tracking via Feature Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing location tracking systems are intrusive as they require objects to be equipped with positioning sensors, which is undesirable in scenarios like business areas, senior centers, and parking lots, where it's impractical or ineffective to expect individuals or vehicles to carry tracking devices.
Innovation Solution
A non-intrusive location tracking method using a plurality of cameras that capture images of objects, analyze features, and match them to stored information to track movements within an area, allowing for the identification and monitoring of objects without the need for sensor-equipped devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS sensors or other positioning sensors are equipped on objects to be tracked, then location tracking accuracy is improved, but the intrusiveness and complexity of the system increases
Solution Approach 1:
The patent creates optical copies (images) of objects using cameras positioned at multiple locations. Instead of tracking objects with onboard sensors, the system captures visual copies from multiple angles and uses image processing to determine object locations and movements, thereby eliminating the need for sensor-equipped devices while maintaining tracking capability
Solution Approach 2:
The patent introduces cameras and image processing algorithms as intermediary elements between the objects to be tracked and the tracking system. The cameras capture images and the processing system extracts location information from these images, serving as a non-intrusive mediator that enables tracking without direct sensor attachment to objects
2Reliability
If sensors are required to be carried by individuals or vehicles for tracking, then tracking reliability is improved, but ease of operation and user acceptance deteriorates
Solution Approach 1:
The system enables objects to be tracked passively without any action required from the objects themselves. The cameras and processing system automatically capture and analyze images to track object movements, allowing guests, patrons, and residents to move freely without carrying or wearing any tracking devices
Solution Approach 2:
By creating and analyzing visual copies of objects through camera images, the system achieves reliable tracking without requiring objects to carry sensors. The optical copies contain sufficient information for location determination, eliminating the burden on users while maintaining tracking effectiveness
3Area of stationary object
If multiple cameras are deployed to cover an area, then tracking coverage is improved, but device complexity and cost increases
Solution Approach 1:
The patent divides the tracking area into multiple zones covered by different cameras. Each camera is responsible for capturing images within its specific field of view, and the system processes images from multiple cameras to achieve comprehensive area coverage. This segmentation allows scalable deployment where cameras are added only where needed
Solution Approach 2:
The patent employs a universal image processing system that handles images from multiple camera sources with different positions and orientations. The same processing algorithms extract location information regardless of which camera captured the image, allowing the system to scale to multiple cameras without proportionally increasing processing complexity
Data Source
AI summary
A tracking system obtains first recognized object and associated first information being detected in at least one image captured by camera(s). First information includes first recognized object's first feature(s), first additional feature(s) derived from the first feature(s), first location(s) of first feature(s), and first real-world dimensions. A list of tracking objects is obtained, each including second recognized object and associated second information. Second information including second recognized object's second feature(s), second additional feature(s) derived from second feature(s), second location(s) of second feature(s), and second real-world dimensions. The system compares first additional feature(s) with second additional feature(s) stored in given tracking object, first location(s) with second location(s) stored in given tracking object, and first real-world dimensions with second real-world dimensions stored in given tracking object. When they match, the first information comprising first additional feature(s) are stored in given tracking object.