Observation and analysis of large-scale human motion

Area of Science:

• Sports Science
• Biomechanics
• Human Movement Analysis

Background:

• Traditional biomechanical systems excel at detailed motion analysis but are limited for large-scale, game-time athlete tracking.
• Observing player movement across entire playing fields requires alternative data acquisition methods.

Purpose of the Study:

• To develop and validate a video-based, computer-aided system for acquiring large-scale human movement data in team sports.
• To enable the study of athlete motion over large areas and long durations, complementing existing high-accuracy systems.

Main Methods:

• A video-based system was developed, sacrificing some spatial accuracy and temporal resolution for broad coverage.
• The system was tested using handball match recordings, analyzing player positions on a 40 x 20 m court.
• Measurement error was assessed through field tests, including comparisons with established biomechanical measurement systems.

Main Results:

• The system successfully acquired position data for 14 handball players over 1 hour.
• Root-mean-square (RMS) error for player positions was below 0.6 meters.
• The system demonstrated feasibility for capturing large-scale motion data relevant to game analysis.

Conclusions:

• The developed video-based system offers a viable solution for large-scale motion acquisition in team sports.
• This approach provides valuable insights into player movement patterns that are not captured by traditional, high-resolution biomechanical tools.
• The system's accuracy and coverage support its application in analyzing game dynamics and player strategies.