Virtual Sensors for Object Tracking in Video Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring systems for areas of interest, such as intersections and shopping malls, require complex and expensive infrastructures of physical sensors and cameras, which are limited in object classification, tracking, and trajectory mapping, and are not adaptable for different use cases, leading to incomplete data and frequent hardware replacements.
Innovation Solution
A computer-implemented method and system using AI-based video processing that receives a video feed, separates frames, detects and tracks objects, and superimposes virtual sensors to gather information on object trajectories, enabling accurate detection, tracking, and classification without the need for physical hardware, and is adaptable for various use cases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If physical sensors and cameras are used to monitor the area of interest, then object detection and tracking can be achieved, but the device complexity and cost increase significantly
Solution Approach 1:
The patent uses virtual sensors that are software-based copies of physical sensors, implemented as computational models rather than physical devices. These virtual sensors process video feed data to detect and track objects, eliminating the need for complex physical sensor installations while maintaining detection capabilities
Solution Approach 2:
The patent replaces mechanical/physical sensor systems with a software-based computer vision system. Instead of using physical cameras and sensors mounted in the area, the system uses virtual sensors that process video feed through algorithms to detect objects, trajectories, and behaviors
2Area of stationary object
If more physical sensors are added to cover a wider area, then the coverage area increases, but the device complexity and maintenance requirements increase
Solution Approach 1:
The virtual sensor system is designed to be universal and adaptable to different areas and use cases. The same software-based virtual sensors can be deployed to monitor intersections, shopping malls, or any other area by simply changing the video feed source and detection parameters, without installing new physical infrastructure
Solution Approach 2:
Instead of physically expanding the sensor network to cover wider areas, the system creates virtual copies of sensors that can be positioned anywhere in the monitored space through software, allowing unlimited coverage expansion without additional physical hardware
3Loss of information
If current camera-based systems are used, then object classification information can be obtained, but the data interpretation becomes difficult and hardware replacement is frequently required
Solution Approach 1:
The virtual sensor system processes and interprets data automatically through integrated algorithms that detect objects, classify them, track trajectories, and generate meaningful insights. The system serves itself by converting raw video data into structured information without requiring external hardware replacements or manual intervention
Solution Approach 2:
The virtual sensors act as an intermediary layer between the video feed and the analysis system. They transform raw video data into structured object information, trajectories, and behavioral data, making the information easily interpretable and usable for various applications
4Measurement precision
If radar-based IoT sensors are used, then object detection is achieved, but object classification capability is limited and the area coverage remains small
Solution Approach 1:
The virtual sensor system is dynamic and can be configured for different use cases by adjusting detection parameters, object classes to monitor, and analysis algorithms. The same infrastructure can adapt to monitor traffic flow, pedestrian behavior, vehicle classification, or other specific applications without hardware changes
Solution Approach 2:
The system provides universal monitoring capabilities that can serve multiple use cases simultaneously. Virtual sensors can detect and classify various object types (pedestrians, vehicles, animals) and perform different analyses (trajectory tracking, behavior detection, counting) using the same infrastructure
Data Source
AI summary
A method, system, and computer program for gathering information of an object moving in an area of interest by using AI-based video processing and analytics, object detection, and tracking moving objects from a video frame to frame. The system, method, and computer may be implemented as a platform including an analytics dashboard and a backend dashboard. The analytics dashboard enables via the UI to display to the user the camera feeds being monitored and analytics results from those feeds over a period of time. The backend dashboard allows the user to set up a camera feed themselves and post-process pre-captured video files using algorithms made available via the system that the user can pick and choose between.