Object Tracking via Camera-Tag Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for tracking player movements in sports activities face challenges such as loss of tracking during occlusions, inability to distinguish player identities, high costs, and inaccuracies, particularly in indoor settings and for non-professional teams.
Innovation Solution
A system that uses cameras and tags with inertial sensors to capture video and sensory data, synchronizing the data to generate performance profiles for objects, allowing for accurate tracking and identification of players and objects without the need for GPS or human operators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video tracking solutions are used to track players, then player positions can be captured, but tracking is lost when players occlude one another and human operators are required to maintain accuracy
Solution Approach 1:
The patent combines video tracking data with inertial sensor data from tags worn by players. When video tracking is lost due to occlusion, the inertial sensors continue to track player position through dead reckoning, eliminating the need for human operators to intervene and maintain tracking accuracy.
Solution Approach 2:
The inertial sensors act as an intermediary system that bridges gaps in video tracking. During occlusion events, the inertial measurement units (IMUs) provide continuous position estimates that supplement the visual tracking data, allowing the system to maintain accuracy without human intervention.
2Measurement precision
If GPS receivers are used for tracking, then player positions can be determined, but the receivers are heavy and bulky making them unwieldy for young players
Solution Approach 1:
The patent extracts the inertial sensing function from the heavy GPS receiver system and implements it using lightweight IMUs integrated into tags. This separation allows the system to achieve position determination accuracy without requiring heavy GPS hardware, making the tags suitable for young players.
Solution Approach 2:
The patent uses lightweight, low-cost inertial sensors in disposable or reusable tags rather than expensive, heavy GPS receivers. These lightweight tags can be easily worn by young players and discarded or replaced if needed, eliminating the weight burden of traditional GPS systems.
3Ease of operation
If inertial sensors are used without GPS, then tracking is lighter and more portable, but accumulated errors cause position data to become unreliable quickly
Solution Approach 1:
The patent uses video tracking data as feedback to correct accumulated errors in the inertial sensor measurements. The visual system provides periodic position references that reset the drift in the inertial integration, maintaining long-term reliability while keeping the system portable and lightweight.
Solution Approach 2:
The patent merges inertial navigation with visual odometry to create a complementary system. The inertial sensors provide continuous, lightweight tracking while the video system periodically corrects accumulated errors, achieving both portability and reliability through data fusion.
4Area of stationary object
If multiple high-resolution cameras are used for tracking, then tracking coverage is improved, but the system becomes expensive and complex
Solution Approach 1:
The patent enables each player to serve as their own tracker through wearable inertial tags. This eliminates the need for multiple expensive cameras to cover the entire field, as each player independently measures their own position, reducing system complexity and cost while maintaining full field coverage.
Solution Approach 2:
The patent segments the tracking function from the camera system and assigns it to individual player tags. Instead of using multiple cameras to track all players, each player carries their own sensing unit, dividing the tracking task into independent, simple units that reduce overall system complexity.
Data Source
AI summary
A method and system for tracking movements of objects in a sports activity are provided. The method includes matching video captured by at least one camera with sensory data captured by each of a plurality of tags, wherein each of the at least one camera is deployed in proximity to a monitored area, wherein each of the plurality of tags is disposed on an object of a plurality of monitored objects moving within the monitored area; and determining, based on the video and sensory data, at least one performance profile for each of the monitored objects, wherein each performance profile is determined based on positions of the respective monitored object moving within the monitored area.