Ball Tracking Device Using Trajectory Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing moving body tracking systems face challenges in accurately tracking objects like balls in sports videos due to similarities in color or shape with other moving bodies, leading to errors and occlusions, especially when the ball is near a player or concealed.

Innovation Solution

A moving body tracking device that includes a video input unit, ball candidate detector, ball tracking unit, image generator for detecting ball trajectory, player region detector, and ball position output unit, which calculates ball likelihood and trajectory based on feature amounts and optical flow, and corrects positions using both initial and trajectory-based outputs to ensure accurate tracking.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a moving body tracking device uses prediction position based on past frame information and extracts candidate objects with predetermined features, then the tracking process can be simplified and executed in real-time, but tracking accuracy deteriorates when the ball resembles other objects (like player shoes) or is occluded

Engineering Contradiction:
Improvereal-time tracking speedVSAvoidball position accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the tracking process into two independent modules: a fast prediction module that uses past frame information to predict ball position, and an accurate trajectory detection module that uses multi-frame analysis to detect ball trajectory. The final position is determined by selecting from these modules based on confidence levels, thus resolving the contradiction between speed and accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by using only necessary frames for trajectory detection when the ball is occluded or ambiguous, rather than always using full multi-frame analysis. This maintains real-time performance while improving accuracy only when needed, based on confidence level assessment.

Inventive Principle:
Principle #16Partial or excessive action

2Device complexity

If the tracking system relies on color and shape features to identify the ball, then the detection process remains simple and fast, but tracking reliability deteriorates when the ball color or shape is similar to other moving bodies like player shoes

Engineering Contradiction:
Improvedetection algorithm complexityVSAvoidball identification reliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent dynamically adjusts the detection approach based on confidence levels. When confidence is high (clear color/shape distinction), the simple feature-based method is used. When confidence is low (similar features to other objects), the system switches to trajectory-based detection using optical flow, thus adapting to varying reliability conditions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent introduces confidence level as an intermediary that mediates between simple feature-based detection and complex trajectory-based detection. This intermediary assesses the reliability of color and shape features and directs the system to appropriate detection methods, resolving the contradiction between simplicity and reliability.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Loss of time

If the system tracks the ball using only single-frame candidate extraction, then processing time is reduced, but tracking accuracy worsens when the ball is occluded by players or other objects

Engineering Contradiction:
Improveprocessing timeVSAvoidball position precision
Core Design Contradiction:
Loss of timeVSMeasurement precision

Solution Approach 1:

The patent implements periodic action by alternating between single-frame candidate extraction (for normal conditions) and multi-frame trajectory detection (for occlusion conditions). The system periodically assesses confidence levels and switches methods accordingly, reducing processing time while maintaining precision during occlusions.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The patent performs preliminary action by detecting ball candidates in advance using single-frame extraction, then uses multi-frame trajectory analysis only when necessary to confirm the ball position. This preliminary detection reduces processing time while ensuring accuracy when occlusions occur.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3179444B1Moving body tracking method and moving body tracking device
Publication Date: 2021.02.17 PANASONIC HOLDINGS CORP
  • EP3179444B1 patent drawingFigure 1
  • EP3179444B1 patent drawingFigure 2
  • EP3179444B1 patent drawingFigure 3

AI summary

A moving body tracking method and a moving body tracking device detect a position of a moving body in each of a plurality of frame images which configure a video, detect a trajectory of the moving body based on the image which is obtained by using the plurality of frame images, determine a final position of the moving body in each of the plurality of frame images based on the detected position and the position of the moving body which is obtained from the detected trajectory, and output the determined position.