Automated Video Tracking with Inertial Sensors for Overlapping Athletes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image analysis tools struggle to reliably track multiple overlapping athletes in video footage due to obscuration, collisions, and the inability to identify athletes by distinguishing features, leading to inconsistent tracking and inaccurate timekeeping.
Innovation Solution
A system that combines machine learning models with motion data from wearable monitors to identify and assign persistent identifiers to athletes, correlating video events with motion events to enhance tracking accuracy and correct time information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If existing image analysis tools are used to track athletes in video footage, then tracking can be performed without additional equipment, but tracking reliability deteriorates when athletes are obscured, collide, or move behind other objects
Solution Approach 1:
The patent combines video footage analysis with inertial sensor data from wearable monitors to create a hybrid tracking system. The video processing component identifies athlete positions and events visually, while the inertial sensors provide continuous motion data that persists even when athletes are obscured from view. By merging these two data streams, the system achieves reliable tracking without requiring complex additional infrastructure.
Solution Approach 2:
The patent introduces inertial sensors as intermediary devices worn by athletes that bridge the gap between video coverage and actual athlete motion. These sensors act as mediators that provide direct motion measurements independent of visual obscurity, allowing the system to maintain tracking reliability when athletes are obscured, collide, or move behind other objects.
2Device complexity
If existing image analysis tools are used to identify athletes by distinguishing features, then no additional identification equipment is needed, but identification accuracy deteriorates when multiple athletes are tracked simultaneously
Solution Approach 1:
The patent uses inertial sensors as intermediary devices that provide unique identification data for each athlete. These sensors capture individual motion patterns and serve as mediators that enable precise athlete identification without requiring complex image analysis of distinguishing features like jersey numbers or facial appearance.
Solution Approach 2:
The system employs feedback mechanisms where inertial sensor data continuously informs and refines athlete identification. The motion data from sensors provides ongoing verification and correction of athlete identities, allowing the system to maintain high identification accuracy throughout the sporting event even when visual features become indistinct.
3Quantity of substance
If video footage alone is used for athlete tracking, then no additional data collection equipment is required, but tracking persistence deteriorates throughout the duration of sporting events
Solution Approach 1:
The patent merges video footage with inertial sensor data to create a persistent tracking system. The inertial sensors continuously record motion data throughout the sporting event, providing an unbroken tracking record that extends beyond the limitations of video coverage. This combination allows tracking to persist throughout the entire duration of the event without requiring continuous clear visual contact.
Solution Approach 2:
The system uses inertial sensors to provide continuous motion data that exceeds what is strictly necessary for basic tracking. This excessive data collection ensures that even when video footage temporarily loses track of an athlete, the inertial sensors maintain the tracking record, thereby extending tracking persistence throughout the entire sporting event.
Data Source
AI summary
A method of tracking video objects within video footage includes receiving, at a computing device, video footage of a space and event data from a plurality of monitors. Each monitor includes a motion sensor configured to measure movements of a respective monitored video object within the space. The event data includes motion events transmitted by the monitors. The method also includes using a machine learning model to identify and track the monitored video objects within the video footage and to identify, from the video footage, video events performed by the monitored video objects. The method also includes using the computing device to assign a respective persistent identifier to each of the monitored video objects within the video footage based at least in part on commonality between video events performed by the monitored video object and motion events transmitted by one of the monitors.