# US 2021/0343024 A1 · Capture-ball-based on-ball point distribution method and motion-posture recognition method, system, and apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0343024 A1
- Document ID: US20210343024A1
- Status: Published application
- Kind: A1
- Application number: 16/470,749
- Filed: 2018-06-12
- Published: 2021-11-04
- Earliest claimed: 2018-04-11
- Assignees: Shenzhen Realis Multimedia Technology Co., Ltd.
- Applicants: Shenzhen Realis Multimedia Technology Co., Ltd.
- Inventors: Yue WANG; Qiuzi XU
- Web page: https://patentgrep.com/patent/US20210343024A1

## CPC

- G06T 7/248: {involving reference images or patches}
- G06T 7/215: Motion-based segmentation
- G06T 7/292: Multi-camera tracking
- G06T 7/74: {involving reference images or patches}
- G06V 40/20: Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16)
- G06T 2200/04: involving 3D image data
- G06T 2207/10016: Video; Image sequence
- G06T 2207/10048: Infrared image
- G06T 2207/30204: Marker
- G06T 2207/30224: Ball; Puck
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

The present invention provides a capture-ball-based on-ball point distribution method, and a motion-posture recognition method, system, and apparatus. Because the on-ball point distribution method adopts a predetermined graphic combination to divide the surface of the ball, the surface of the ball demonstrates a uniform geometric distribution state, which is conducive to uniformly disposing reflective marking points to form a capture ball that can be used for optical capture, and helps the optical motion capture system recognize the motion-posture information of the capture ball according to the displacement changes of the reflective marking points. In addition, because the motion-posture recognition method adds a method for matching the recognizable reflective marking points and the capture ball in the process of acquiring the plurality of matching points on the capture ball, and introduces an iterative algorithm for calculating the coordinates of the spherical center in the process of acquiring the coordinates of the spherical center of the capture ball, misrecognition caused by matching errors of the reflective marking points and increased coordinate errors of the spherical center can be effectively avoided, conducive to improving the recognition degree of the capture ball in the system.

## Claims

### Claim 1

An on-ball point distribution method used for optical motion capture, comprising:
dividing a surface of a ball into a plurality of sub-regions; and
distributing reflective marking points in the plurality of sub-regions such that a distribution of distance values between any two reflective marking points is concentrated, wherein the reflective marking points are used for the optical motion capture of the ball.

### Claim 2 (depends on 1)

The on-ball point distribution method according to claim 1, wherein dividing the surface of the ball into the plurality of sub-regions includes:
according to a size of the ball, dividing the surface of the ball into a graphic combination of a first quantity of first geometric figures and a second quantity of geometric figures.

### Claim 3 (depends on 2)

The on-ball point distribution method according to claim 2, wherein:
the first geometric figures are regular hexagons, the second geometric figures are regular pentagons, and the regular hexagons and the regular pentagons have a same side length.

### Claim 4 (depends on 2)

The on-ball point distribution method according to claim 2, wherein:
the first quantity is twenty, and the second quantity is twelve.

### Claim 5 (depends on 3)

The on-ball point distribution method according to claim 3, wherein according to the size of the ball, dividing the surface of the ball into the first quantity of first geometric figures and the second quantity of second geometric figures includes:
according to a diameter or a radius of the ball, calculating a surface area of the ball; and
according to a total area of the graphic combination equal to the surface area of the ball, calculating the side length of the regular hexagons and the regular pentagons.

### Claim 6 (depends on 3)

The on-ball point distribution method according to claim 3, wherein according to the size of the ball, dividing the surface of the ball into the first quantity of first geometric figures and the second quantity of second geometric figures includes:
determining a reference point on an equatorial line of the surface of the ball, using the reference point as a center point of a first regular hexagon to determine a positional region of the first hexagon; and
forming a graphic combination in which three regular hexagons and three regular pentagons are alternately distributed on six edges of a regular hexagon by respectively dividing the regular pentagons based on three non-adjacent edges of the first regular hexagon, and dividing the regular hexagons again based on the other three non-adjacent edges of the first regular hexagon, and forming the twenty regular hexagons and the twelve regular pentagons until the surface of the ball is completely divided.

### Claim 7 (depends on 6)

The on-ball point distribution method according to claim 6, wherein distributing the reflective marking points in the plurality of sub-regions includes:
distributing the reflective marking points of the ball at apexes or edge-midpoints of each geometric figure.

### Claim 8 (depends on 7)

The on-ball point distribution method according to claim 7, wherein distributing the reflective marking points of the ball at the apexes or edge-midpoints of each geometric figure includes one of:
distributing the reflective marking points of the ball at the apexes of the regular hexagons; and
distributing the reflective marking points of the ball at the midpoints of non-adjacent edges of the hexagons.

### Claim 9 (depends on 6)

A capture ball for optical motion capture, wherein a plurality of reflective marking points is distributed on the capture ball, and each reflective marking point is obtained through the on-ball point distribution method according to claim 6.

### Claim 10 (depends on 9)

An optical motion capture system, comprising:
a capture ball including the capture ball of claim 9;
a plurality of cameras, configured to capture motion images of the capture ball in a motion space during a movement process; and
a motion-posture recognition apparatus, communicatively coupled to the plurality of cameras, and configured to perform motion-posture recognition on the capture ball based on the motion images.

### Claim 11

A motion-posture recognition method for a capture ball, wherein a surface of the capture ball is configured to include a plurality of reflective marking points that is uniformly distributed, a distribution of distance values between any two reflective marking points is concentrated, and the method includes:
acquiring three-dimensional coordinates of all the reflective marking points of the capture ball in a motion space, and identifying a plurality of matching points of the capture ball, wherein the matching points are reflective marking points on the capture ball;
acquiring coordinates of a spherical center of the capture ball according to three-dimensional coordinates of the plurality of matching points; and
acquiring motion-posture information of the capture ball in the motion space according to the three-dimensional coordinates of the plurality of matching points and the coordinates of the spherical center of the capture ball.

### Claim 12 (depends on 11)

The motion-posture recognition method according to claim 11, wherein acquiring the three-dimensional coordinates of all the reflective marking points of the capture ball in the motion space, and identifying the plurality of matching points of the capture ball include:
acquiring the three-dimensional coordinates of all the reflective marking points according to two-dimensional coordinates of all the reflective marking points in the motion space, wherein the motion space includes other identifiable objects different from the capture ball;
calculating distances between any two reflective marking points among all the reflective marking points according to the three-dimensional coordinates of all the reflective marking points; and
when a distance between two reflective marking points is equal to or close to a standard value of distance, determining the two reflective marking points as marching points of the capture ball until a number of the marching points identified on the capture ball reaches a predetermined number.

### Claim 13 (depends on 12)

The motion-posture recognition method according to claim 12, wherein the standard value of distance is obtained by:
acquiring the three-dimensional coordinates of all reflective marking points in a testing stage, wherein identifiable objects in the motion space in the testing stage only include the capture ball;
calculating distances between any two reflective marking points according to the three-dimensional coordinates of all the reflective marking points on the capture ball; and
counting all obtained distance values, and setting a distance value at which a result of distance distribution is concentrated as the standard value of distance.

### Claim 14 (depends on 11)

The motion-posture recognition method according to claim 11, wherein acquiring the coordinates of the spherical center of the capture ball according to the three-dimensional coordinates of the plurality of matching points includes:
obtaining theoretical coordinates of the spherical center and a theoretical radius of the capture ball;
calculating a corrected radius of the capture ball and corrected coordinates of the spherical center according to the theoretical coordinates of the spherical center;
determining whether a difference between the corrected coordinates of the spherical center and the theoretical coordinates of the spherical center is within a preset error range, and determining whether a difference between the corrected radius of the capture ball and the theoretical radius of the capture ball is in a present error range;
when results of both determinations are yes, determining the corrected coordinates of the spherical center as the spherical center of the capture ball; and
when a result of any determination is no, updating the theoretical coordinates of the spherical center to the corrected coordinates of the spherical center, updating the theoretical radius of the capture ball to the corrected radius, and then according to the theoretical coordinates of the spherical center, entering a step of calculating a corrected radius of the capture ball and corrected coordinates of the spherical center.

### Claim 15 (depends on 14)

The motion-posture recognition method according to claim 14, wherein calculating the corrected radius of the capture ball and the corrected coordinates of the spherical center according to the theoretical coordinates of the spherical center includes:
calculating a distance from each matching point to the theoretical spherical center based on the three-dimensional coordinates of the plurality of matching points and the theoretical coordinates of the spherical center of the capture ball;
calculating an average value of the distance from each matching point to the theoretical spherical center, and taking the average value as the corrected radius of the capture ball;
obtaining a coordinate offset M from each matching point to the theoretical spherical center based on the three-dimensional coordinates of the plurality of matching points and the theoretical coordinates of the spherical center;
respectively calculating a difference W between the distance from each matching point to the theoretical center and a preset radius of the capture ball, wherein the preset radius includes the theoretical radius or the corrected radius of the capture ball;
multiplying the coordinate offset M from each matching point to the theoretical spherical center and the difference W between the distance from each matching point to the theoretical spherical center and the preset radius of the capture ball to obtain a corrected coordinate offset of each matching point, and calculating an average value of corrected coordinate offsets of all the matching points; and
obtaining the corrected coordinates of the spherical center according to the average value of the corrected coordinate offsets of all the matching points and the theoretical coordinates of the spherical center.

### Claim 16 (depends on 11)

The motion-posture recognition method according to claim 11, wherein acquiring the motion-posture information of the capture ball in the motion space according to the three-dimensional coordinates of the plurality of matching points and the coordinates of the spherical center of the capture ball includes:
comparing coordinates of the spherical center of the capture ball of a current frame with coordinates of the spherical center of the capture ball of a previous frame to obtain displacement information of the capture ball;
comparing three-dimensional coordinates of each matching point of the current frame with three-dimensional coordinates of a corresponding matching point of the previous frame to obtain posture information of the capture ball; and
using the displacement information and the posture information of the capture ball as motion-posture information of the capture ball in the motion space.

### Claim 17 (depends on 6, 11)

The motion-posture recognition method according to claim 11, wherein:
the plurality of reflective marking points on the surface of the capture ball is distributed through the on-ball point distribution method according to claim 6.

### Claim 18

A motion-posture recognition apparatus for a capture ball, wherein a surface of the capture ball is configured to include a plurality of reflective marking points, and the apparatus includes:
a matching-point acquisition unit, configured to identify a plurality of matching points of the capture ball, and acquire three-dimensional coordinates of each matching point, wherein the matching point is a reflective marking point belonging to the capture ball;
a spherical-center-coordinates acquisition unit, configured to acquire coordinates of a spherical center of the capture ball according to three-dimensional coordinates of the plurality of matching points; and
a posture-information acquisition unit, configured to acquire motion-posture information of the capture ball in a motion space according to the three-dimensional coordinates of the plurality of matching points and the coordinates of the spherical center of the capture ball.

### Claim 19 (depends on 18)

The motion-posture recognition apparatus according to claim 18, wherein acquiring the coordinates of the spherical center of the capture ball according to the three-dimensional coordinates of the plurality of matching points includes:
obtaining theoretical coordinates of the spherical center and a theoretic radius of the capture ball;
calculating a corrected radius of the capture ball and corrected coordinates of the spherical center according to the theoretical coordinates of the spherical center;
determining whether a difference between the corrected coordinates of the spherical center and the theoretical coordinates of the spherical center is within a preset error range, and whether a difference between the corrected radius of the capture ball and the theoretical radius of the capture ball is in a present error range;
when results of both determinations are yes, determining the corrected coordinates of the spherical center as the spherical center of the capture ball; and
when a result of any determination is no, updating the theoretical coordinates of the spherical center to the corrected coordinates of the spherical center, updating the theoretical radius of the capture ball to the corrected radius, and then according to the theoretical coordinates of the spherical center, entering a step of calculating a corrected radius of the capture ball and corrected coordinates of the spherical center.

### Claim 20 (depends on 11)

A computer storage medium, including a program, wherein the program is executed by a processor to implement the method according to claim 11.
