Virtual Field Mapping from Single Video Stream

Solution Overview

Problem

Existing technologies require multiple simultaneous video streams to perform advanced sports analytics, which is not feasible for most observers who only have access to a single broadcast stream.

Innovation Solution

A system that translates a single two-dimensional video stream of a sporting event into a virtual three-dimensional field-of-play rendering, using a reference database and processors to identify camera angle settings and translate objects-of-interest onto the virtual rendering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple simultaneous video streams are used to perform advanced sports analytics, then measurement precision of object locations is improved, but device complexity and cost increase

Engineering Contradiction:

Improvelocation accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a virtual copy of the physical field-of-play with identical geometric features and dimensions. This virtual field serves as a reference model that can be mapped against the single video stream, enabling precise location determination without requiring multiple physical cameras. The virtual field is populated with coordinate information for all relevant locations, allowing the system to calculate object positions by matching visual features from the single video stream against the known geometry of the virtual field.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent transitions from analyzing multiple two-dimensional video streams to creating a three-dimensional virtual representation of the field from a single two-dimensional video stream. By building a virtual field with depth and spatial coordinates, the system achieves three-dimensional location accuracy using only a single two-dimensional video input, effectively adding dimensional information through computational modeling rather than through multiple camera feeds.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If multiple simultaneous video streams are used to perform advanced sports analytics, then measurement precision of object locations is improved, but accessibility and cost increase

Engineering Contradiction:

Improvelocation accuracyVSAvoidaccessibility

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent creates a virtual copy of the physical field-of-play with identical geometric features and dimensions. This virtual field serves as a reference model that can be mapped against the single video stream, enabling precise location determination without requiring multiple physical cameras. The virtual field is populated with coordinate information for all relevant locations, allowing the system to calculate object positions by matching visual features from the single video stream against the known geometry of the virtual field.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent creates a universal virtual field model that can be applied to any sporting event on a standardized field. This single virtual field representation can serve multiple purposes: it can analyze player positions, track object movement, determine distances, and support various analytics applications all from a single video stream. The system is designed to work with different sports and field types by loading appropriate virtual field configurations, making the analytics capability universally applicable rather than requiring specialized multiple-camera setups for each event.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If multiple simultaneous video streams are used to perform advanced sports analytics, then measurement precision of object locations is improved, but loss of time for data collection increases

Engineering Contradiction:

Improvelocation accuracyVSAvoiddata processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent pre-populates the virtual field with coordinate information for all relevant locations, boundaries, and features before the sporting event begins. This preliminary setup includes defining the geometry, dimensions, and coordinate systems of the entire field. During the event, the system only needs to match visual features from the video stream against this pre-established virtual model, rather than building and calibrating the field model in real-time. This preliminary preparation significantly reduces the computational burden and time required during actual data collection and analysis.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12217359B2Three-dimensional rendering of sporting event from two-dimensional video stream

Publication Date: 2025.02.04 GRATZ JOSEPH

AI summary

Methods and apparatus are disclosed for three-dimensional mappings of sporting events from two-dimensional video streams. An example system comprises processor(s) that generate a virtual three-dimensional rendering of a physical field-of-play. The processor(s) generate camera angle entries of a reference database each of which corresponds with a virtual image of the virtual rendering and includes a virtual camera angle setting at which the virtual image is captured and two-dimensional coordinates of static feature(s) of the virtual rendering. For each frame within a video stream, the processor(s) identify two-dimensional coordinates of static feature(s) of the physical field-of-play, determines a matching virtual camera angle setting by comparing the two-dimensional coordinates, and translate objects-of-interest within the frame onto the virtual rendering based on a virtual camera position setting and the matching virtual camera angle setting.