A method and system for evaluating the matching of stadium tactical segments based on graphic similarity

Abstract

The present invention relates to the technical field of computer vision and image processing, and particularly relates to a method and system for evaluating the matching of court tactical segments based on graphic similarity. The method includes: obtaining a court video and dividing it into a plurality of court tactical segments through frame segmentation, constructing a coordinate system based on the court, and obtaining the position data of each moving target in each frame; converting the position data into a plurality of closed polygons, and determining the corner function of the closed polygons; calculating the corner function of any court tactical segment, corresponding and matching it with the tactical formations in the tactical template library, and determining the tactical graphic similarity between the court tactical segment and the corresponding and matched tactical formation to obtain the score of any court tactical segment; quantifying the graphic similarity index through the centroid algorithm and the corner function, and performing matching scoring, which solves the problem of low efficiency of traditional methods and eliminates subjective biases.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision and image processing, and particularly to a method and system for evaluating the matching of football tactical segments based on graphic similarity. Background Art

[0002] In the current field of football tactical analysis, it mainly relies on static formation template matching technology, which can calculate the similarity between different tactics by analyzing simple geometric features such as the position distribution of players and the center of gravity distance; however, football tactics actually have very strong temporal dynamic characteristics. For different execution stages, even the same tactic may show significant tactical formation evolutions such as the stretching and shrinking of the formation during the offense-defense transition process, which makes the existing static template matching methods unable to capture these dynamic features. Therefore, during the tactical execution process, especially when there are significant differences in stages such as offense-defense transition and formation stretching and shrinking, the accuracy of matching decreases; to solve this problem, usually, the calculation of graphic similarity is mostly based on shallow features such as Euclidean distance or contour overlap degree, which is difficult to depict the corner changes and side length dissimilarities of the formation structure; for example, traditional methods may misjudge a football field where there are essential differences in local corner features between a diamond midfield and a rectangular midfield as similar, unable to accurately match, resulting in a decrease in the accuracy of tactical analysis. Summary of the Invention

[0003] In order to solve the technical problems that the prior art cannot automatically implement tactical matching and there are misjudgments resulting in low accuracy of tactical matching, the purpose of the present invention is to provide a method for evaluating the matching of football tactical segments based on graphic similarity, and the specific technical solutions adopted are as follows:

[0004] Obtain a football field video, divide it into several football field tactical segments through frame segmentation, construct a coordinate system based on the football field, and obtain the position data of each moving target in each frame;

[0005] Convert the position data into several closed polygons, and determine the corner function of any closed polygon;

[0006] Establish a tactical template library, calculate the corner function of any football field tactical segment, perform corresponding matching with the tactical formations in the tactical template library, and determine the tactical graphic similarity between the football field tactical segment and the corresponding matching tactical formation, so as to obtain the score of any football field tactical segment.

[0007] Preferably, obtaining a football field video, dividing it into several football field tactical segments through frame segmentation, constructing a coordinate system based on the football field, and obtaining the position data of each moving target in each frame includes:

[0008] Obtain a football field video, determine the total number of frames and the number of frames for unit segmentation, and divide the football field video into several football field tactical segments. The corresponding calculation formula is:

[0009]

[0010] Among them, represents the total number of court tactical segments, that is , represents the th court tactical segment; represents the total number of frames; represents the unit split number of frames;

[0011] Based on the court, a plane rectangular coordinate system is constructed with the lower left corner of the top view as the origin, and the position data of any moving target of the player in each frame is collected, denoted as Among them, represents the coordinate position of the corresponding moving target in the th frame; represents the mark of the team to which the current moving target belongs.

[0012] Preferably, by converting the position data into a number of closed polygons, the corner function of any closed polygon is determined, including:

[0013] Based on adjacent moving targets to form sides, closed polygons are generated according to the position data of each moving target in each frame, and the closed polygons of all frames are obtained;

[0014] Calculate the lengths of all sides. The extension line of any side in any closed polygon forms an angle with the first adjacent side in the counterclockwise direction, and the corner function of the corresponding closed polygon is determined by combining the lengths of the sides corresponding to the corner;

[0015] The tactical graph similarity is obtained according to the norm of the difference between the corner functions of two closed polygons.

[0016] Preferably, calculate the lengths of all sides. The extension line of any side in any closed polygon forms an angle with the first adjacent side in the counterclockwise direction, and the corner function of the corresponding closed polygon is determined by combining the lengths of the sides corresponding to the corner, including:

[0017] Based on any closed polygon the vertex set is obtained as , and the corresponding position data is Among them, represents the number of vertices of the closed polygon;

[0018] Determine the centroid of the closed polygon. Based on the centroid as the origin and the horizontal right as the positive direction of the polar axis, a polar coordinate system is constructed, and the polar angle corresponding to each vertex is calculated;

[0019] Sort the polar angles from smallest to largest, define the vertex corresponding to the smallest polar angle as the starting point, and sort counterclockwise along the starting point to determine an ordered set of closed polygons , and the corresponding position data is , where represents a rearrangement of

[0020] Obtain the set of edges of the closed polygon . The included angle between the extension line of any edge and the first adjacent edge in the counterclockwise direction forms a turning angle , and calculate the lengths of all edges of the closed polygon ;

[0021] Use the arccosine function to calculate the turning angle and obtain the length of the corresponding edge to determine the turning angle function

[0022] Preferably, use the arccosine function to calculate the turning angle, and the corresponding calculation formula is

[0023]

[0024] where represents the turning angle between the extension line of edge in the closed polygon and the first adjacent edge in the counterclockwise direction, represents the th vertex in the closed polygon, ; respectively represent the position data of vertices , and ;

[0025] Obtain the length of the corresponding edge, and the corresponding calculation formula is

[0026]

[0027]

[0028] where and respectively represent the lengths of edge and edge ;

[0029] Determine the turning angle function, and the corresponding calculation formula is

[0030]

[0031] where and are both independent variables, representing the turning angle and the side length respectively

[0032] Preferably, the similarity of tactical graphics is obtained according to the norm of the difference of the corner functions of pairwise closed polygons, and the corresponding calculation formula is:

[0033]

[0034] Wherein, represents the similarity of tactical graphics; and respectively represent the corner functions of the closed polygon and the closed polygon .

[0035] Preferably, a tactical template library is established, the corner function of any court tactical segment is calculated, corresponding matching is performed with the tactical formations in the tactical template library, and the similarity of the tactical graphics between the court tactical segment and the corresponding matching tactical formation is determined to obtain the score of any court tactical segment, including:

[0036] Establish a tactical template library according to professionals and classic tactics. Each set of tactical formations includes a tactical name, the number of formations, and formation information;

[0037] Calculate the corner function of the closed polygon formed in each frame of the court tactical segment, traverse the tactical template library, obtain the tactical formation that matches the closed polygon formed in each frame, judge whether the number between the unit segmentation frames and the number of formations is equal, and determine the similarity of the tactical graphics between the court tactical segment and the corresponding matching tactical formation according to whether the numbers are equal or not, and correspondingly calculate the score of the court tactical segment.

[0038] Preferably, if the number between the unit segmentation frames and the number of formations is equal, determine the similarity of the tactical graphics between each frame closed polygon in the court tactical segment and the corresponding matching tactical formation, and sum and obtain the average value to get the score of the court tactical segment. The corresponding calculation formula is:

[0039]

[0040] Wherein, represents the score of the court tactical segment; represents the unit segmentation frames; represents the th frame in the unit segmentation frames; represents the th frame;

[0041] Preferably, if the number of unit segmentation frames is not equal to the number of formation types, calculate the tactical graphic similarity between each tactical formation in the tactical template library and the closed polygon formed by all frames in the court tactical segment, and select the highest tactical graphic similarity to determine the score of the corresponding court tactical segment.

[0042] When the number of unit segmentation frames is not equal to the number of formation types, the corresponding calculation formula is:

[0043]

[0044] Among them, and respectively represent the closed polygon formed within the th frame in the court tactical segment, and the tactical formation with the highest tactical graphic similarity to the closed polygon formed within the th frame 's corner function; represents a non-decreasing sequence; represents the number of formation types of the tactical formation.

[0045] To solve the above problems, the present application also provides: A court tactical segment matching evaluation system based on graphic similarity, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a court tactical segment matching evaluation method based on graphic similarity as described in any one of the foregoing.

[0046] The present invention has the following beneficial effects:

[0047] 1. This application obtains real-time player movement data through target recognition and tracking technology, constructs a coordinate system, and quantifies the position analysis of moving targets; combines the centroid algorithm and defines the corner function of a closed polygon to quickly generate a quantitative representation of the tactical formation, and through dynamic modeling of the collaboration relationship between moving targets on the field, quantifies the graphic similarity index, realizes the automatic extraction, modeling, and matching of court tactical segments, automatically matches and scores the court tactical segments and the tactical template library, significantly reducing the time cost of manual observation and video playback, and solving the problems of relying on manual experience and low efficiency in traditional methods; through the scoring mechanism, eliminates the subjective bias of human judgment, and can provide a more accurate evaluation of tactical completion through the provided mathematical differences, providing scientific and objective game or training feedback for professionals; among them, the tactical template library is compatible with professionals and classic techniques and tactics, can be flexibly extended, is conducive to subsequent integration of machine learning algorithms, and improves the ability of dynamic prediction and strategy recommendation, enabling the entire evaluation method to be used not only for real-time tactical analysis during the game, but also to provide technical support for player training to compare the differences between the matched tactical formation and the actual court tactics, enabling professionals to quickly identify tactical execution loopholes and then optimize training strategies; and can also provide a visual interpretation of the tactics for the audience, enhance the viewing experience, and promote the intelligent development of sports events.

[0048] 2. The present invention also provides a court tactical segment matching evaluation system based on graphic similarity for implementing the foregoing court tactical segment matching evaluation method based on graphic similarity. This system has the same beneficial effects as the foregoing court tactical segment matching evaluation method based on graphic similarity and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is the implementation flowchart of a court tactical segment matching evaluation method based on graphic similarity provided by an embodiment of the present invention;

[0051] Figure 2 It is the graphic difference comparison diagram of converting a closed polygon into a corner function in a court tactical segment matching evaluation method based on graphic similarity provided by an embodiment of the present invention;

[0052] Figure 3Schematic diagram for quantifying the difference between pairwise corner functions of a method for evaluating the matching of court tactical segments based on graphic similarity provided by an embodiment of the present invention;

[0053] Figure 4 Schematic diagram of a tactical template library of a method for evaluating the matching of court tactical segments based on graphic similarity provided by an embodiment of the present invention. Detailed implementation manners

[0054] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features and effects of a method and system for evaluating the matching of court tactical segments based on graphic similarity proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0056] The following specifically describes the specific solutions of a method and system for evaluating the matching of court tactical segments based on graphic similarity provided by the present invention in combination with the accompanying drawings.

[0057] In the prior art, traditional static template matching and relying on manual experience analysis cannot capture the timing dynamic characteristics of the conversion of techniques and tactics in real time for different execution stages, resulting in a low matching accuracy between the techniques and tactics and the tactical template library during the tactical execution process, and reducing the accuracy of tactical analysis; the first embodiment of the present invention provides a method for evaluating the matching of court tactical segments based on graphic similarity, which automatically matches the tactical map and gives a score evaluation by dynamically modeling the cooperation relationship between players and quantifying the graphic similarity index, intuitively quantifying the techniques and tactics executed by players and providing real-time dynamic feedback on tactical changes; to solve a method for evaluating the matching of court tactical segments based on graphic similarity, the second embodiment of the present invention provides a system for evaluating the matching of court tactical segments based on graphic similarity, which is essentially a software system composed of units that implement corresponding functions. The specific steps in the method are introduced in detail below.

[0058] Please refer to Figure 1 , which shows the implementation flowchart of a method for evaluating the matching of court tactical segments based on graphic similarity provided by an embodiment of the present invention. The method includes:

[0059] Step S1: Obtain the stadium video, segment it into several stadium tactical segments through frame segmentation, construct a coordinate system based on the stadium, and obtain the position data of each moving target in each frame;

[0060] Step S2: Convert the position data into several closed polygons, and determine the corner function of any closed polygon;

[0061] Step S3: Establish a tactical template library, calculate the corner function of any stadium tactical segment, perform corresponding matching with the tactical formations in the tactical template library, and determine the tactical graphic similarity between the stadium tactical segment and the corresponding matching tactical formation to obtain the score of any stadium tactical segment.

[0062] As an optional implementation manner, in this embodiment, the provided method for matching and evaluating stadium tactical segments is mainly described for a football field, and it is equally applicable to sports events such as basketball courts or volleyball that require close technical and tactical cooperation among players, that is, it is generally applicable to projects with complex standing positions, formations, and tactics; by performing matching and evaluation on tactics, a real-time and effective tactical analysis method is provided for professionals such as referees, coaches, or players to better understand the tactical application in the game and improve the overall competitive level.

[0063] Further, in step S1, it includes:

[0064] Step S11: Obtain the stadium video, determine the total number of frames and the unit segmentation number of frames, and segment the stadium video into several stadium tactical segments. The corresponding calculation formula is:

[0065]

[0066] Wherein, represents the total number of stadium tactical segments, that is , represents the th stadium tactical segment; represents the total number of frames; represents the unit segmentation number of frames.

[0067] Specifically, obtain the stadium video to be analyzed, and determine its total number of frames and the unit segmentation number of frames , wherein, that is, frames are used as a unit to divide the total number of frames of the stadium video, and a total of groups of stadium tactical segments formed based on a certain number of frames are obtained.

[0068] Step S12: Based on the stadium, construct a plane rectangular coordinate system with the lower left corner of the top view as the origin, collect the position data of any moving target of the player in each frame, denoted as , wherein, Indicates the coordinate position of the corresponding moving target in the frame; Indicates the label of the team to which the current moving target belongs.

[0069] Explanation is made that a plane rectangular coordinate system is constructed with the lower left corner of the top view of the stadium as the origin through a computer vision recognition tool; perspective transformation is performed through a dimension transformation matrix to convert the pixel coordinates in the stadium video into the plane rectangular coordinate system, and the position data of each player on the stadium in the plane rectangular coordinate system constructed for each frame is obtained.

[0070] Please refer to Figure 2 and Figure 3 which respectively show a schematic diagram of the graphical difference comparison of converting a closed polygon into a corner function and a schematic diagram of the quantification of the difference between two corner functions in a method for evaluating the matching of stadium tactical segments based on graphical similarity provided by an embodiment of the present invention; wherein, the corner function for the football tactical segment determination and the corner function determined by the tactical formation have the same change trend and are relatively close in difference, that is, they are the matching football tactical segment and tactical formation; and the difference curve is the quantification of the difference between two corner functions to reflect the difference degree between the football tactical segment and the tactical formation.

[0071] Further, in step S2, it includes:

[0072] Step S21: Form edges based on adjacent moving targets, generate a closed polygon according to the position data of each moving target in each frame, and obtain the closed polygons of all frames.

[0073] Specifically, according to the different teams to which the players belong, an edge is formed between two adjacent players in the corresponding team, and the players in the same team in each frame are connected to form a closed polygon.

[0074] Step S22: Calculate the lengths of all edges, form a corner by the extension line of any edge in any closed polygon and the first adjacent edge in the counterclockwise direction, and determine the corner function of the corresponding closed polygon in combination with the lengths of the edges corresponding to the formed corner.

[0075] Further, in step S22, it includes:

[0076] Step S221: Based on any closed polygon obtain a vertex set as , and the corresponding position data is , where represents the number of vertices of the closed polygon.

[0077] Step S222: Determine the centroid of the closed polygon. With the centroid as the origin and the horizontal rightward direction as the positive direction of the polar axis, construct a polar coordinate system and calculate the polar angle corresponding to each vertex; make an explanation. Assume that the selected centroid is , then , .

[0078] Step S223: Sort the polar angles from smallest to largest, define the vertex corresponding to the smallest polar angle as the starting point, and sort counterclockwise along the starting point to determine the ordered set of the closed polygon , and the corresponding position data is , where represents rearrangement of.

[0079] Step S224: Obtain the set of sides of the closed polygon . The included angle between the extension line of any side and the first adjacent side in the counterclockwise direction forms a turning angle , and calculate the lengths of all sides of the closed polygon .

[0080] It can be explained that according to the logic of the turning angle composition, similarly, the turning angles of the closed polygon at the starting point and the end point obtained after counterclockwise sorting can be obtained, which are respectively and , that is is the extension line of and the first adjacent side in the counterclockwise direction formed; is the extension line of and formed, represents the starting point, represents the end point.

[0081] Step S225: Use the arccosine function to calculate the turning angle and obtain the length of the corresponding side to determine the turning angle function.

[0082] Furthermore, in step S225, the arccosine function is used to calculate the turning angle, and the corresponding calculation formula is:

[0083]

[0084] where represents the turning angle between the extension line of side in the closed polygon and the first adjacent side in the counterclockwise direction , represents the th vertex in the closed polygon, ; respectively represent the positions of vertices , and position data.

[0085] Preferably, based on the corner calculation logic, similarly, the corners formed at the starting point and the ending point of the closed polygon can also be calculated to determine the feasibility of the overall technical solution. Among them, the formula for calculating the corner at the starting point is as follows:

[0086]

[0087] Similarly, the formula for calculating the corner formed at the ending point is as follows:

[0088]

[0089] The length of the corresponding side is obtained, and the corresponding calculation formula is:

[0090]

[0091]

[0092] Among them, and respectively represent the lengths of side and side ;

[0093] Optionally, the lengths of the two sides forming the corner are calculated by the Euclidean distance formula, that is, the straight-line distance between the two vertices forming the side is determined to obtain the length of the side.

[0094] The corner function is determined, and the corresponding calculation formula is:

[0095]

[0096] Among them, and are both independent variables, respectively representing the corner and the side length.

[0097] It can be explained that in this embodiment, the corner function describes the angle change of the closed polygon from one side to another side, and is used to quantify the tactical behavior of the player for the formed closed polygon, facilitating the accurate analysis of technical and tactical skills; and similarly, based on the verification of the corners at the starting point and the ending point respectively, the corresponding lengths can be calculated for all sides of the entire closed polygon, and then the corner function of the closed polygon can be determined according to any side.

[0098] Step S23: Obtain the tactical graph similarity based on the norm of the difference between the corner functions of pairwise closed polygons.

[0099] Specifically, the difference degree between two tactical graphs is represented by defining the norm of the difference between the corner functions of two closed polygons, and then this difference degree is used as the tactical graph similarity.

[0100] Furthermore, in step S23, the corresponding calculation formula is:

[0101]

[0102] where, represents the tactical graph similarity; and respectively represent the corner functions of the closed polygon and the closed polygon .

[0103] It should be noted that when the value of the tactical graph similarity is larger, it indicates that the similarity between pairwise tactical graphs is higher; on the contrary, when the value of the tactical graph similarity is smaller, it indicates that the similarity between tactical graphs is lower, indicating less commonality in vision.

[0104] Please refer to Figure 4 , which shows a schematic diagram of the tactical template library of a method for evaluating the matching of football tactical segments based on graph similarity provided by an embodiment of the present invention.

[0105] Furthermore, in step S3, it includes:

[0106] Step S31: Establish a tactical template library according to professionals and classic tactics. Each set of tactical formations includes a tactical name, the number of formations, and formation information.

[0107] It should be noted that in this embodiment, the professionals refer to professional football coaches; specifically, according to professional football coaches and classic football tactics, such as different types of tactical formations like 4-4-2, 4-3-3, or total football, pressing defense, etc., a tactical template library is established based on two-dimensional atlases, denoted as , which can provide a benchmark for real-time football tactical segments, ensure the accuracy and objectivity of subsequent scoring, and improve the scientificity and practicality of football evaluation. Its structure is defined as , where, represents the total number of tactical formations in the tactical template library; for any tactical formation, it is expressed as , represents the th set of tactical formations; represents the tactical name of the th set of tactical formations; Indicates the number of formation patterns in the tactical formation set; Indicates the set of formation information executed in sequence for this set of tactical formations, that is, any formation graphic corresponding to each set of tactical formations Stored in a graph structure, that is , Indicates the all vertices of the Indicates the all edges of the

[0108] Step S32: Calculate the turning function of the closed polygon formed within each frame of the court tactical segment, traverse the tactical template library, obtain the tactical formation that matches the closed polygon formed within each frame, determine whether the quantity between the unit segmentation frames and the number of formation patterns is equal, and respectively determine the tactical graphic similarity between the court tactical segment and the corresponding matching tactical formation according to whether the quantity is equal or not, and correspondingly calculate the court tactical segment score.

[0109] It can be explained that judging whether the quantity between the unit segmentation frames of the court tactical segment and the number of formation patterns in the tactical template library is equal can evaluate the integrity and coherence of the overall tactical execution. That is, if the unit segmentation frames are equal to the number of formation patterns, it means that the tactical execution in the court tactical segment is completely matched with the tactical formation, and sufficient tactics are obtained without omission or jump, and the integrity of the tactical execution process is high; on the contrary, if the quantity between the two is not equal, it means that there are mistakes in the tactical execution and corresponding matching is required.

[0110] Furthermore, if the quantity between the unit segmentation frames and the number of formation patterns is equal, determine the tactical graphic similarity between each frame's closed polygon in the court tactical segment and the corresponding matching tactical formation, and sum and obtain the average value to get the score of the court tactical segment. The corresponding calculation formula is:

[0111]

[0112] Wherein, Indicates the score of the court tactical segment; Indicates the unit segmentation frames; Indicates the th frame in the unit segmentation frames; Indicates the tactical graphic similarity between the closed polygon formed within the

[0113] Specifically, construct the closed polygon formed by the players within the th frame in the same way as in Step S2, that is, for the position data set of the players within this frame , taking the centroid as the origin and the positive horizontal direction as the polar axis to construct a polar coordinate system, calculating the polar angle, determining the starting point, and sorting in counterclockwise order from the starting point to obtain the ordered set after sorting as , the sequence is a rearrangement from 1 to . Connect each adjacent point in , and connect the last point to the starting point to form a closed polygon , whose edge set is denoted as , and then determine the corresponding turning angle function as , and match it with the tactical template library . That is, traverse the initial formation of each tactical formation in the tactical template library as , and its corresponding turning angle function is , to obtain the tactical formation with the highest tactical graphic similarity to the closed polygon in the first frame of the court tactical segment as , and the corresponding logical formula is:

[0114]

[0115]

[0116] Then judge whether the number of unit segmentation frames is equal to the number of formations. If so, that is , determine the tactical graphic similarity between the closed polygons of all frames in the currently analyzed court tactical segment and the tactical formations they match, and sum and average to obtain the score corresponding to the currently analyzed court tactical segment.

[0117] Furthermore, if the number of unit segmentation frames is not equal to the number of formations, calculate the tactical graphic similarity between each tactical formation in the tactical template library and the closed polygons formed by all frames in the court tactical segment, and screen the highest tactical graphic similarity to determine the score corresponding to the court tactical segment;

[0118] When the number of unit segmentation frames is not equal to the number of formations, the corresponding calculation formula is:

[0119]

[0120] Among them, and respectively represent the closed polygon formed within the th frame in the court tactical segment, and the tactical formation with the highest tactical graphic similarity to the closed polygon formed within the th frame 's turning angle function; represents a non-decreasing sequence; represents the number of formations of the tactical formation.

[0121] Understandably, when the number of unit segmentation frames is not equal to the number of formation patterns, there are two cases for the data between the two, which are respectively or When , it indicates that the current executed technique and tactic is relatively simple and belongs to a formation pattern that can be fully preset by the tactical formation patterns in the tactical template library. At this time, the highest tactical graph similarity is selected to determine the score of the corresponding stadium tactical segment; conversely, when , it means that the techniques and tactics covered in the stadium tactical segment exceed the scope covered by the tactical formation patterns in the tactical template library. Therefore, in the process of calculating the score of the stadium tactical segment, when the number of unit segmentation frames is greater than the number of formation patterns, although there are redundant frames, due to the coherence of the techniques and tactics, the stadium tactical segment still corresponds to a tactical formation pattern, that is, the stadium tactical segment can still select a matching tactical formation pattern from the tactical template library. Therefore, based on the case where the number of unit segmentation frames is equal to the number of formation patterns, the score of the stadium tactical segment is obtained in the same way.

[0122] Understandably, this application uses target recognition and tracking technology to obtain player movement data in real time, constructs a coordinate system, and quantifies the position analysis of moving targets; combines the centroid algorithm and defines the corner function of a closed polygon to quickly generate a quantitative representation of the tactical formation pattern, and quantifies the graph similarity index by dynamically modeling the cooperation relationship between moving targets on the stadium, realizing the automatic extraction, modeling and matching of stadium tactical segments, so as to automatically match and score the stadium tactical segments and the tactical template library, significantly reducing the time cost of manual observation and video playback, and solving the problems of relying on manual experience and low efficiency in traditional methods; through the scoring mechanism, the subjective deviation of human judgment is eliminated, and through the provided mathematical differences, a more accurate evaluation of tactical completion can be provided, providing scientific and objective game or training feedback for professionals; among them, the tactical template library is compatible with the establishment of professionals and classic techniques and tactics, can be flexibly expanded, is conducive to subsequent integration of machine learning algorithms, and improves the dynamic prediction and strategy recommendation capabilities, making the entire evaluation method not only applicable to real-time tactical analysis during the game, but also providing technical support for player training to compare the differences between the matched tactical formation pattern and the actual stadium tactics, enabling professionals to quickly identify tactical execution loopholes and then optimize training strategies; and it can also provide a visual interpretation of the tactics for the audience, enhance the viewing experience, and promote the intelligent development of sports events.

[0123] The second embodiment of the present invention also proposes a stadium tactical segment matching and evaluation system based on graph similarity, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of a stadium tactical segment matching and evaluation method according to any one of the foregoing embodiments.

[0124] It should be noted that a court tactical segment matching evaluation system based on graphic similarity has the same beneficial effects as the previously provided court tactical segment matching evaluation method based on graphic similarity, which will not be elaborated here.

[0125] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0126] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for evaluating the matching of court tactical segments based on graphic similarity, characterized in that The method includes: Obtaining a stadium video, segmenting it into a number of stadium tactical segments by frame number, constructing a coordinate system based on the stadium, and obtaining the position data of each moving object in each frame; Converting the position data into a number of closed polygons, and determining the corner function of any closed polygon, including: Forming sides based on adjacent moving objects, generating closed polygons according to the position data of each moving object in each frame, and obtaining the closed polygons of all frames; Calculating the lengths of all sides, forming a corner by the extension line of any side in any closed polygon and the first adjacent side in the counterclockwise direction, and determining the corner function of the corresponding closed polygon in combination with the lengths of the sides constituting the corner; Obtaining the tactical graph similarity according to the norm of the difference between the corner functions of pairwise closed polygons; Among them, the tactical graph similarity is obtained according to the norm of the difference between the corner functions of pairwise closed polygons, and the corresponding calculation formula is: Among them, ξ represents the tactical graph similarity; and respectively represent the corner functions of the closed polygon Ψ1 and the closed polygon Ψ2; Establishing a tactical template library, calculating the corner function of any stadium tactical segment, corresponding matching with the tactical formations in the tactical template library, and determining the tactical graph similarity between the stadium tactical segment and the corresponding matching tactical formation, so as to obtain the score of any stadium tactical segment.

2. The evaluation method for matching and evaluating court tactical segments based on graphic similarity according to claim 1, wherein Obtaining a stadium video, segmenting it into a number of stadium tactical segments by frame number, constructing a coordinate system based on the stadium, and obtaining the position data of each moving object in each frame, including: Obtaining a stadium video, determining the total number of frames and the unit segmentation number of frames, and segmenting the stadium video into a number of stadium tactical segments, and the corresponding calculation formula is: r = Frame / N where r represents the total number of court tactical segments, i.e., T i , i = 1, 2, …, r, T i represents the i-th court tactical segment; Frame represents the total number of frames; N represents the number of frames per unit segment; Based on the court, a plane rectangular coordinate system is constructed with the lower left corner of the top view as the origin, and the position data of any moving target of the player in each frame is collected, denoted as ((x1,y1),(x2,y2),…,(x n ,y n ),sign(x)), where (x n ,y n ) represents the coordinate position of the corresponding moving target in the nth frame; sign(x) represents the mark of the team to which the current moving target belongs.

3. A method for evaluating the matching of a court tactical segment based on graphic similarity according to claim 1, characterized in that, Calculating the lengths of all sides, forming a corner by the extension line of any side in any closed polygon and the first adjacent side in the counterclockwise direction, and determining the corner function of the corresponding closed polygon in combination with the lengths of the sides constituting the corner, including: Based on any closed polygon Ψ, the vertex set is R = {O1, O2, …, O S}, and the corresponding position data is P = {(x1, y1), (x2, y2), …, (x S , y S ), where S represents the number of vertices of the closed polygon; Determining the centroid of the closed polygon, constructing a polar coordinate system with the centroid as the origin and the horizontal right direction as the positive direction of the polar axis, and calculating the polar angle corresponding to each vertex; Sort the polar angles from smallest to largest, define the vertex corresponding to the smallest polar angle as the starting point, sort counterclockwise along the starting point, and determine an ordered set of closed polygons The corresponding position data is Among them, k S represents a rearrangement of 1 to S; Obtain the set of edges of the closed polygon Ψ The included angle between the extension line of any edge and the first adjacent edge in the counterclockwise direction forms a turning angle And calculate the lengths of all the edges of the closed polygon Ψ Using the arccosine function to calculate the corner and obtaining the length of the corresponding side, and determining the corner function.

4. A method for evaluating the matching of court tactical segments based on graphic similarity according to claim 3, characterized in that, Using the arccosine function to calculate the corner, and the corresponding calculation formula is: Among them, represents the turning angle between the extension line of side in the closed polygon and the first adjacent side in the counterclockwise direction where i represents the i-th vertex in the closed polygon, and i = 2, 3, …, S - 1; respectively represent the position data of vertices and ; Obtaining the length of the corresponding side, and the corresponding calculation formula is: Among them, and respectively represent the lengths of edge and edge ; Determining the corner function, and the corresponding calculation formula is: Among them, both θ and L are independent variables, representing the corner and the side length respectively.

5. The evaluation method for matching and evaluating court tactical segments based on graphic similarity according to claim 1, characterized in that, Establishing a tactical template library, calculating the corner function of any stadium tactical segment, corresponding matching with the tactical formations in the tactical template library, and determining the tactical graph similarity between the stadium tactical segment and the corresponding matching tactical formation, so as to obtain the score of any stadium tactical segment, including: Establishing a tactical template library according to professionals and classic tactics, and each set of tactical formations includes a tactical name, the number of formations, and formation information; Calculating the corner function of the closed polygon formed in each frame of the stadium tactical segment, traversing the tactical template library, obtaining the tactical formation that matches the closed polygon formed in each frame, judging whether the number between the unit segmentation number of frames and the number of formations is equal, and respectively determining the tactical graph similarity between the stadium tactical segment and the corresponding matching tactical formation according to whether the numbers are equal or not, and correspondingly calculating the score of the stadium tactical segment.

6. The evaluation method for matching and evaluating court tactical segments based on graphic similarity according to claim 5, characterized in that, If the quantity between the unit segmentation number of frames and the number of formation patterns is equal, determine the tactical graphic similarity between each closed polygon of each frame within the stadium tactical segment and the corresponding matching tactical formation pattern, and sum them up to obtain the average value as the score of the stadium tactical segment. The corresponding calculation formula is as follows: Among them, Score represents the score of the tactical segment of the stadium; N represents the number of frames for unit segmentation; n represents the nth frame in the unit segmentation frames; ξ n represents the similarity between the closed polygon formed within the nth frame and the tactical graph of the corresponding matching tactical formation.

7. A method for evaluating the matching of a court tactical segment based on graphic similarity according to claim 6, characterized in that If the quantity between the unit segmentation number of frames and the number of formation patterns is not equal, calculate the tactical graphic similarity between each tactical formation pattern in the tactical template library and the closed polygon formed by all the frames in the stadium tactical segment, and screen the highest tactical graphic similarity to determine the score corresponding to the stadium tactical segment; When the unit segmentation number of frames is not equal to the number of formation patterns, the corresponding calculation formula is as follows: Among them, f n (θ, L) and f ti (θ, L) respectively represent the closed polygon formed within the nth frame in the stadium tactical segment, and the corner function of the tactical formation t with the highest similarity to the tactical graphic of the closed polygon formed within the nth frame; a n represents a non-decreasing sequence; K i represents the number of formations of the tactical formation.

8. A court tactical segment matching and evaluation system based on graphic similarity, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a method for matching and evaluating stadium tactical segments based on graphic similarity as described in any one of claims 1 to 7.

Images