Radar and Vision Fusion for Markerless Golf Club 3D Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing golf launch monitoring systems face challenges in accurately measuring club parameters without requiring additional markers, which can alter the appearance of the club and introduce hardware complexity, and lack the ability to provide comprehensive 3D reconstruction for improved performance analysis.
Innovation Solution
A combined radar and vision-based system measures club head movement and orientation from the back or side of the player without markers, using 3D scanning and stereo vision to reconstruct the launching scene in high fidelity, allowing for interactive 3D replay and accurate measurement of club parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vision-based solution is used to measure club parameters, then measurement accuracy for certain parameters (lie angle, total spin, spin axis) is improved, but additional markers/stickers on club face are required which affect appearance and increase hardware complexity
Solution Approach 1:
The patent combines radar-based measurement system with vision-based measurement system into a unified golf launch monitoring system. The radar system provides club head speed, attack angle, and club path measurements, while the vision system captures club head position and orientation. By merging these two systems, the patent achieves comprehensive measurement capabilities without requiring markers on the club face, as the radar system can independently measure club parameters through radar signals reflected from the club head and ball.
Solution Approach 2:
The patent introduces radar signals as an intermediary medium to measure club parameters. Instead of requiring direct visual contact with markers on the club face, the radar system uses electromagnetic waves to detect club head motion, speed, and orientation. The radar signals act as a mediator that enables measurement without physical contact or visual markers, thereby eliminating the need for additional hardware on the club while maintaining measurement accuracy.
2Reliability
If markers are placed on club face to improve vision-based measurement robustness, then measurement reliability is improved, but club appearance is altered and user acceptance decreases
Solution Approach 1:
The patent replaces the mechanical marker system with a radar-based measurement system. Instead of using physical markers on the club face that alter appearance, the radar system uses electromagnetic wave reflection to detect club head parameters. This substitution eliminates the need for visual markers while maintaining measurement reliability, as the radar system can accurately track club head position, speed, and orientation through signal processing without requiring physical modifications to the club.
3Loss of information
If high speed cameras are used to capture club and ball movement, then performance analysis quality is improved, but hardware cost increases significantly
Solution Approach 1:
The patent merges radar-based measurement with vision-based measurement to achieve comprehensive performance analysis. The radar system provides precise measurements of club head speed, attack angle, and club path, while the vision system captures club head position and orientation. This combination allows for detailed performance analysis without requiring expensive high-speed cameras, as the radar system can independently provide critical motion data through electromagnetic signal processing.
Solution Approach 2:
The patent changes the measurement parameters from high-frame-rate video capture to radar signal processing. Instead of using high-speed cameras to capture every frame of club and ball motion, the radar system processes electromagnetic signals to extract key parameters such as club head speed, direction, and trajectory. This parameter transformation enables detailed performance analysis with lower hardware requirements, as radar signal processing can achieve the necessary measurement precision without expensive high-speed imaging equipment.
4Measurement precision
If camera-based system is used to measure full 3D postures of club head, then measurement capability is improved, but system complexity increases due to lack of consistent distinguishing features
Solution Approach 1:
The patent introduces radar signals as an intermediary to measure club head 3D posture. The radar system emits electromagnetic waves that reflect off the club head, and by analyzing the phase and frequency changes of the reflected signals, the system can determine club head position, orientation, and motion in 3D space. This radar-based intermediary approach provides consistent measurement capability without requiring visual markers or complex camera systems, as the radar signals can detect club head parameters regardless of lighting conditions or background variations.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate 3D reconstruction of club and ball movements, providing detailed performance analysis from any perspective, reducing hardware complexity and cost, and supporting both left- and right-handed players without additional markers.
Implementation Method 1
a radar device configured to transmit a radar signal and receive a reflected signal from the club
Implementation Method 2
a camera configured to capture a first image of the club and a second image of the club
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
An example method of modeling a portion of a golf club and a golf swing includes scanning the golf club to obtain scanning information, training a convolutional neural network using the scanning information, using at least one camera to obtain a series of images, converting the series of images into parameterized motion representations, using at least one radar to obtain a radar signal, converting the radar signal into time-frequency images, inputting the parameterized motion representations and the time-frequency images into the convolutional neural network, receiving golf club parameters and golf swing parameters as an output of the convolutional neural network, and generating a visual model of the golf club and the golf swing in a virtual space using the golf club parameters and the golf swing parameters.