A method and system for recognizing tactical graphics on a football field based on tactical relevance

Abstract

The present invention relates to the field of computer recognition technology, and particularly relates to a method and system for recognizing tactical graphics on a football field based on tactical relevance. The method includes: constructing a coordinate system based on the football field, and collecting the position data of any moving target of players and the ball in each frame; obtaining the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance to generate a relevance matrix; obtaining a weight matrix, and determining the tactical relevance between any two players in combination with the relevance matrix; constructing a tactical relevance graph, defining the tactical relevance as the weight of the corresponding edge in the tactical relevance graph, selecting any vertex in the edge with the largest weight as the starting point to obtain a plurality of polygons, and screening, and merging the screened polygons to obtain the tactical graphics of any frame; analyzing the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance between two players, comprehensively determining the tactical relevance of the players, and presenting it intuitively with tactical graphics, with higher accuracy.

Description

Technical Field

[0001] The present invention relates to the field of computer recognition technology, and particularly relates to a method and system for recognizing tactical graphics on a football field based on tactical relevance. Background Art

[0002] During the process of football training and matches, the positions of players and the ball are often complex and changeable. With the continuous progress of football match technology and the increasing demand for data analysis, traditional tactical analysis methods are difficult to automatically and intuitively display tactical graphics. Existing tactical recognition methods usually rely on manual experience analysis, relying on experts' experience to judge the tactics of real-time players, unable to effectively automatically identify the tactical connections between players. Moreover, manual analysis is inefficient and cannot quickly obtain the connection between the real-time changing positions of players and the trajectory of the moving ball, thus unable to accurately capture complex tactical intentions and the overall tactical layout of the team, and unable to meet the high requirements for tactical analysis in modern football matches. Summary of the Invention

[0003] In order to solve the technical problem that existing tactical recognition mainly relies on manual work, resulting in low recognition efficiency and accuracy and being unable to meet the high demand for tactical analysis in modern football matches, the object of the present invention is to provide a method for recognizing tactical graphics on a football field based on tactical relevance. The specific technical solutions adopted are as follows:

[0004] Construct a coordinate system based on the football field, and collect the position data of any moving target of players and the ball in each frame;

[0005] Obtain the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance in sequence according to the position data, and generate a relevance matrix;

[0006] Assign weights to the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance respectively to obtain a weight matrix, and combine the relevance matrix to determine the tactical relevance between any two players;

[0007] Construct a tactical relevance graph, define the tactical relevance as the weight of the corresponding edge in the tactical relevance graph, select any vertex in the edges with the largest weight as the starting point to obtain several polygons, and screen and merge the screened polygons to obtain the tactical graphics of each frame.

[0008] Preferably, constructing a coordinate system based on the football field and collecting the position data of any moving target of players and the ball includes:

[0009] Construct a plane rectangular coordinate system with the lower left corner of the top view of the football field as the origin, and use a camera to collect the position data of any moving target of players and the ball in each frame, denoted as ;

[0010] Among them, represents the number of frames, represents the total number of frames; represents the number of players, represents the total number of players; represents the team to which the players on the football field belong, represents the total number of teams.

[0011] Preferably, based on the position data, the spatial distance correlation, the speed and acceleration correlation, and the historical tactic correlation are obtained in sequence to generate a correlation matrix, including:

[0012] Based on the position data of the players, the straight-line distance between any two players is obtained, a judgment threshold is set to determine the authenticity of the interaction between the two players, and the spatial distance correlation is calculated;

[0013] Based on the position data of the players and combined with the frame interval of each frame, the speed and acceleration of the players are obtained, and based on the coordinate system, the speed angle and the acceleration angle are respectively obtained, and the speed correlation and the acceleration correlation are calculated correspondingly, and the speed correlation and the acceleration correlation are combined to obtain the speed and acceleration correlation;

[0014] Analyze the average tactic correlation between two players in the previous frame adjacent to the current frame, and obtain the historical tactic correlation through the contribution weight;

[0015] Construct a tactic correlation matrix through the spatial distance correlation, the speed and acceleration correlation, and the historical tactic correlation, calculate the average distance between two players and the ball to determine the correction coefficient, and based on the correction coefficient, correct the tactic correlation matrix to generate a correlation matrix.

[0016] Preferably, based on the position data of the players, the straight-line distance between any two players is obtained, a judgment threshold is set to determine the authenticity of the interaction between the two players, and the spatial distance correlation is calculated, including:

[0017] Based on the position data of the players, the straight-line distance between any two players is obtained, and the corresponding calculation formula is:

[0018]

[0019] Among them, represents the th player and the th player's straight-line distance;

[0020] Set a judgment threshold , when the straight-line distance is less than the judgment threshold, it means that the interaction between the two players is real; when the straight-line distance is greater than the judgment threshold, it means that the interaction between the two players is not real, and a decay coefficient is constructed;

[0021] Calculate the spatial distance correlation, and the corresponding calculation formula is:

[0022]

[0023] Wherein, represents the spatial distance correlation of the th frame; represents a constant; represents the attenuation coefficient.

[0024] Preferably, based on the position data of the player and the frame interval of each frame, obtain the speed and acceleration of the player, and respectively obtain the speed angle and acceleration angle based on the coordinate system, calculate the speed correlation and acceleration correlation, and combine the speed correlation and acceleration correlation to obtain the speed-acceleration correlation, including:

[0025] Based on the position data of the player and the frame interval of each frame, obtain the speed and acceleration of the player, and respectively denote them as and ;

[0026] Based on the coordinate system, respectively obtain the speed angle and acceleration angle, and respectively denote them as and and ;

[0027] Calculate the speed correlation, and the corresponding calculation formula is:

[0028]

[0029] Wherein, represents the speed correlation of the th frame; represents the th player; represents the th player;

[0030] Calculate the acceleration correlation, and the corresponding calculation formula is:

[0031]

[0032] Wherein, represents the acceleration correlation of the th frame;

[0033] Combine the speed correlation and the acceleration correlation to obtain the speed-acceleration correlation, and the corresponding calculation formula is:

[0034]

[0035]

[0036] Among them, represents the speed and acceleration correlation degree of the th frame; represents the minimum value in .

[0037] Preferably, analyze the average tactical correlation degree of two players in the previous frame adjacent to the current frame, and obtain the historical tactical correlation degree through the contribution weight. The corresponding calculation formula is:

[0038]

[0039] Among them, represents the historical tactical correlation degree of the th frame; represents the average tactical correlation degree of two players in the th frame; represents the weight coefficient.

[0040] Preferably, construct a tactical correlation matrix through the spatial distance correlation degree, speed and acceleration correlation degree, and historical tactical correlation degree. Calculate the average distance between two players and the ball to determine the correction coefficient, and correct the tactical correlation matrix based on the correction coefficient to generate the correlation matrix, including:

[0041] Construct a tactical correlation matrix, denoted as ;

[0042] Calculate the average distance between two players and the ball. The corresponding calculation formula is: , where represents the average distance between two players and the ball, represents the straight-line distance between the th player and the ball; represents the straight-line distance between the th player and the ball;

[0043] Determine the correction coefficient. The corresponding calculation formula is: , where represents the correction coefficient;

[0044] Based on the correction coefficient, correct the tactical correlation matrix to generate the correlation matrix. The corresponding calculation formula is:

[0045]

[0046] Among them, represents the correlation matrix; , , respectively represent the corrected spatial distance correlation, speed and acceleration correlation, and historical tactics correlation.

[0047] Preferably, weights are respectively assigned to the spatial distance correlation, speed and acceleration correlation, and historical tactics correlation to obtain a weight matrix, and the tactics correlation between any two players is determined in combination with the correlation matrix, including:

[0048] Determine the weight matrix, denoted as , where , , respectively represent the weights corresponding to the spatial distance correlation, speed and acceleration correlation, and historical tactics correlation;

[0049] The tactics correlation between two players is determined in combination with the correlation matrix, and the corresponding calculation formula is:

[0050]

[0051] where represents the tactics correlation; represents the correlation matrix.

[0052] Preferably, a tactics correlation graph is constructed, and the tactics correlation is defined as the weight of the corresponding edge in the tactics correlation graph. Any vertex in the edges with the largest weight is selected as the starting point to obtain several polygons, and after screening, the screened polygons are merged to obtain the tactics graph of any frame, including:

[0053] Based on the first frame, construct a tactics correlation graph, denoted as , where represents the team on the football field; represents the total number of teams on the football field; represents the th team, which is the set of position data of all players in the team; represents the th team, which is the set of tactics correlations between any two players in the team, denoted as the edge set of the tactics correlation graph;

[0054] According to the edge set select any vertex in the largest edge as the starting point, denoted as vertex , set a screening threshold, and select the edges that contain vertex and have a number of vertices greater than the screening threshold from the edge set to obtain several polygons;

[0055] Calculate the average tactics correlation for any polygon, and the corresponding calculation formula is:

[0056]

[0057]

[0058] Among them, represents the average tactical relevance of the polygon; represents the increase coefficient; represents the total number of players and the ball in the currently analyzed team; represents the number of sides of the polygon; represents the th side of the polygon in terms of tactical relevance;

[0059] Sort the polygons in descending order according to the average tactical relevance, and sequentially obtain the first initial tactical graph until the initial tactical graph. Among them, the first initial tactical graph is denoted as , and delete the position data of the players in from , and delete the tactical relevance in from . Similarly, process all polygons until the number of points of the position data of all players in the rd team is less than or equal to 2; When the number of points of the position data in

[0060] Merge all the processed initial tactical graphs to obtain the tactical graph of any frame, denoted as . Among them, represents the tactical graph of the th frame for the th team; represents the set of the position data of the players in all the initial tactical graphs; represents the set of the tactical relevance between any two players in all the initial tactical graphs.

[0061] To solve the above problems, the present application also provides: A football field tactical graph recognition system based on tactical relevance, 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 football field tactical graph recognition method according to any one of the foregoing.

[0062] The present invention has the following beneficial effects:

[0063] 1. First, this application constructs a coordinate system, obtains the position data of moving targets, analyzes the spatial positions, speeds, accelerations, and the relevance of historical cooperation relationships between any two players, and determines the tactical relevance between players by integrating the position data of the ball. By constructing an initial tactical graph, it can accurately identify all polygons with the maximum average tactical relevance, and on this basis, merge them to obtain the actual tactical graph of the corresponding team in each frame, which can effectively identify the tactics that players may execute on the field and display them in an intuitive way. It is applicable to football training and competitions with complex standing positions, formations, and tactics, can deeply explore the tactical connections between players, accurately identify the tactical content at each moment, that is, each frame, and display it intuitively through a graphical method, providing a more explicit display of on-field features.

[0064] 2. The present invention also provides a football field tactical graph recognition system based on tactical relevance for implementing the foregoing provided football field tactical graph recognition method based on tactical relevance. This system has the same beneficial effects as the foregoing football field tactical graph recognition method based on tactical relevance and will not be elaborated herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] 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 use in the description of the embodiments or the prior art. Obviously, the following 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.

[0066] Figure 1 It is the flowchart of the implementation of a football field tactical graph recognition method based on tactical relevance provided by an embodiment of the present invention;

[0067] Figure 2 It is a schematic diagram of the spatial distance relevance of a football field tactical graph recognition method based on tactical relevance provided by an embodiment of the present invention;

[0068] Figure 3 It is a schematic diagram of the speed and acceleration relevance of a football field tactical graph recognition method based on tactical relevance provided by an embodiment of the present invention;

[0069] Figure 4 It is a schematic diagram of any possible initial tactical graph of a football field tactical graph recognition method based on tactical relevance provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0070] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, elaborate in detail on a method and system for identifying tactical patterns on a football field based on tactical relevance proposed according to the present invention, including its specific implementation manner, structure, features, and effects. 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.

[0071] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0072] The following will specifically describe the specific solution of a method and system for identifying tactical patterns on a football field based on tactical relevance provided by the present invention in conjunction with the accompanying drawings.

[0073] Since the existing identification of tactics on a football field mostly relies on manual acquisition by experts, professors, etc. with rich experience, it is inefficient and has low accuracy. It cannot accurately and real-time obtain the tactical connection between players and the ball, capture complex tactical intentions and the overall tactical layout of the team, and cannot meet the high requirements for tactical analysis in modern football games. The first embodiment of the present invention provides a method for identifying tactical patterns on a football field based on tactical relevance. This method analyzes the position data of moving targets by constructing a coordinate system, sequentially obtains the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance, determines the tactical relevance, constructs an initial tactical pattern, analyzes it, and obtains the tactical pattern actually existing for the team corresponding to each frame, which can effectively identify the tactics that players may execute on the field and present them in an intuitive manner. To solve the method for identifying tactical patterns on a football field based on tactical relevance, the second embodiment of the present invention provides a system for identifying tactical patterns on a football field based on tactical relevance. This system is essentially a software system composed of units that implement corresponding functions. Now, the specific steps in this method will be introduced in detail.

[0074] Please refer to Figure 1 , which shows the implementation flowchart of a method for identifying tactical patterns on a football field based on tactical relevance provided by an embodiment of the present invention. The method includes:

[0075] Step S1: Based on the football field, construct a coordinate system and collect the position data of any moving target of players and the ball in each frame;

[0076] Step S2: According to the position data, sequentially obtain the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance, and generate a relevance matrix;

[0077] Step S3: Assign weights to the spatial distance relevance, speed and acceleration relevance, and historical tactics relevance respectively to obtain a weight matrix, and determine the tactics relevance between any two players in combination with the relevance matrix;

[0078] Step S4: Construct a tactics relevance graph, and define the tactics relevance as the weight of the corresponding edge in the tactics relevance graph. Select any vertex in the edges with the largest weights as the starting point to obtain several polygons, and perform screening. Merge the screened polygons to obtain the tactics graph for any frame.

[0079] For better illustration, currently, the recognition of tactics on the football field mostly relies on the manual acquisition by experts, professors and other experienced people, which is inefficient and has low accuracy. It cannot accurately obtain the tactical connection between players and the ball in real time, cannot capture complex tactical intentions and the overall tactical layout of the team, and cannot meet the high requirements for tactical analysis in modern football games. In response to this, this application proposes a method for recognizing tactics graphs on the football field based on tactics relevance, constructs tactics graphs, so as to be able to automatically recognize the complex formations and tactics on the football field based on the connection during the real-time movement of moving targets, with high reliability and accuracy, and is presented in the form of tactics graphs, which is clear and intuitive and can adapt to complex football training and game scenarios.

[0080] Further, in step S1, it includes:

[0081] Construct a plane rectangular coordinate system with the lower left corner of the top view of the football field as the origin, and use a camera to collect the position data of any moving target of players and the ball in each frame, denoted as ;

[0082] Among them, represents the number of frames, represents the total number of frames; represents the number of players, represents the total number of players; represents the team to which the players on the football field belong, represents the total number of teams.

[0083] It should be noted that the camera is used to record the field conditions on the football field in real time, that is, the position information of the players and the ball on the field is obtained by means of various existing perspective conversion, object detection, object tracking and identity matching methods. According to the constructed plane rectangular coordinate system, the position data of each player and the ball, that is, the corresponding position coordinates, are determined.

[0084] It can be explained that the tactics relevance refers to the degree of cooperation and coordination of players in the football field according to the tactical intentions formulated by the coach and the actions of their teammates, and can reflect how players cooperate, understand and respond to each other's actions on the football field, so as to accurately reflect the tactical layout.

[0085] Furthermore, in step S2, it includes:

[0086] Step S21: Obtain the straight-line distance between any two players based on the position data of the players, set a judgment threshold to determine the authenticity of the interaction between the two players, and calculate the spatial distance correlation.

[0087] It should be noted that the spatial distance correlation refers to a correlation index of how players effectively occupy the field space and cooperate with teammates through standing positions, running positions, and passing the ball, that is, by analyzing the straight-line distance between players, the position changes of players are analyzed; the farther the spatial distance between two players, the lower the degree of cooperation between them, and when the spatial distance is lower than a certain threshold, the degree of cooperation is more stable.

[0088] Please refer to Figure 2 , which shows a schematic diagram of the spatial distance correlation of a method for recognizing tactical graphics on a football field based on tactical correlation provided by an embodiment of the present invention; among them, a high spatial distance correlation, that is, a short straight-line distance between two players, is represented by a solid line; a low spatial distance correlation, that is, a long straight-line distance between two players, is represented by a dashed line.

[0089] Furthermore, in step S21, it includes:

[0090] Obtain the straight-line distance between any two players based on the position data of the players, and the corresponding calculation formula is:

[0091]

[0092] Wherein, represents the straight-line distance between the th player and the th player; represents the total number of players;

[0093] Set a judgment threshold , when the straight-line distance is less than the judgment threshold, it indicates that the interaction between the two players is real; when the straight-line distance is greater than the judgment threshold, it indicates that the interaction between the two players is not real, and a decay coefficient is constructed;

[0094] Calculate the spatial distance correlation, and the corresponding calculation formula is:

[0095]

[0096] Wherein, represents the spatial distance correlation of the th frame; represents a constant; represents the decay coefficient.

[0097] Understandably, a preset judgment threshold , and the value of the judgment threshold is set according to actual requirements; when the straight-line distance is less than the judgment threshold, it indicates that the interaction between the two players is real, that is, the interaction and coordinated defense between the players are relatively stable, and the degree of cooperation between the two is not affected by the change of the distance interval. At this time, the corresponding spatial distance correlation is a constant value; when the straight-line distance is greater than the judgment threshold, it indicates that the interaction between the two players is not real, that is, the straight-line distance between the two players is relatively far. At this time, the lower the spatial distance correlation, the overall shows a negative correlation state. It is difficult for the two players to have an effective interaction due to the distance problem, and the relationship between the spatial distance correlation and the straight-line distance shows a rapid attenuation characteristic to construct an attenuation coefficient, and a negative exponential function is used to describe the situation where the interaction is not real.

[0098] Step S22: Obtain the speed and acceleration of the players according to the position data of the players combined with the frame interval of each frame, and respectively obtain the speed angle and acceleration angle based on the coordinate system, calculate the speed correlation and acceleration correlation correspondingly, and combine the speed correlation and acceleration correlation to obtain the speed-acceleration correlation.

[0099] It should be noted that the speed-acceleration correlation refers to the relationship and the degree of interaction between the speed and acceleration of the players during the movement process. Among them, the speed is used to describe the rate of change of the displacement of the players per unit time; the acceleration is the rate of change of the speed per unit time; that is, obtain the difference between the speed and acceleration vectors of the two players. The greater the difference, the lower the cooperation relationship; on the contrary, the higher the cooperation relationship.

[0100] Understandably, the speed and acceleration are collectively referred to as the motion vectors. When the difference in the magnitude and angle of the motion vectors between the two players is greater, the corresponding speed correlation and acceleration correlation are smaller; in actual motion, when the difference of either the speed or the acceleration tends to be extremely large, the speed-acceleration correlation will tend to zero. Therefore, it is necessary to consider the influence of these two factors on the finally obtained correlation at the same time; when the difference in the magnitude and direction of the motion vectors is zero, that is, when the motion states of the two players are exactly the same, it can be regarded that the cooperation degree between the two reaches the highest and the tactical effect is the best; and the speed-acceleration correlation decays rapidly with the increase of the difference through exponential operation, and the trigonometric function converts the difference between the speed and acceleration into a numerical value, effectively reflecting the influence of the angle factor on the speed-acceleration correlation.

[0101] Please refer to Figure 3 , which shows a schematic diagram of the speed-acceleration correlation of a method for recognizing tactical graphics on a football field based on tactical correlation provided by an embodiment of the present invention.

[0102] Furthermore, in step S22, it includes:

[0103] The speed and acceleration of a player are obtained based on the player's position data in combination with the inter-frame interval of each frame, and are denoted as and ;

[0104] Based on the coordinate system, the speed angle and the acceleration angle are obtained respectively, and are denoted as and respectively, and ;

[0105] Calculate the speed correlation, and the corresponding calculation formula is:

[0106]

[0107] where represents the speed correlation of the th frame; represents the th player; represents the th player;

[0108] Calculate the acceleration correlation, and the corresponding calculation formula is:

[0109]

[0110] where represents the acceleration correlation of the th frame;

[0111] Combine the speed correlation and the acceleration correlation to obtain the speed-acceleration correlation, and the corresponding calculation formula is:

[0112]

[0113]

[0114] where represents the speed-acceleration correlation of the th frame; represents the minimum value in , that is .

[0115] Specifically, for the position data of the first frame and the last frame in the acquisition process, backward difference quotient and forward difference quotient are respectively used, while the middle difference quotient is used for the middle frames, that is, the speed and acceleration of the corresponding analysis frame are obtained from the changing position data in two frames; the speed and acceleration of the player are obtained based on the player's position data in combination with the inter-frame interval of each frame; by analyzing the coordinate system, the included angle between the speed and acceleration in the current frame can be obtained, so as to calculate the speed correlation and the acceleration correlation respectively, and jointly determine the speed-acceleration correlation.

[0116] Step S23: Analyze the average tactical relevance between two players in the previous frame adjacent to the current frame, and obtain the historical tactical relevance through the contribution weight.

[0117] It should be noted that the historical tactical relevance refers to the tactical relevance of the aforementioned number of frames corresponding to the current frame, that is, the past tactical relevance between players; the contribution weight is used to measure the role of the average tactical relevance between two players in the previous frame in the historical tactical relevance of the current analysis frame.

[0118] Furthermore, in step S23, the corresponding calculation formula is:

[0119]

[0120] where represents the historical tactical relevance of the th frame; represents the average tactical relevance between two players in the th frame; represents the weight coefficient.

[0121] Specifically, for the analysis of the historical tactical relevance of the current th frame, to a certain extent, it depends on the average tactical relevance between two players in the previous adjacent th frame, that is . Assuming that has a high result value, it indicates that there has been a close cooperation relationship between the two players. In the next frame within a short period, the tactical cooperation relationship between the two players may remain unchanged. Therefore, at the th frame, it is considered that there is still a high tactical relevance between the two players; that is, it means that the tactical relevance of the previous adjacent frame makes a certain contribution to the tactical relevance of the subsequent frame. Set the contribution weight to numerically represent the influence of the previous frame on the subsequent frame, and combine the contribution weight with the average tactical relevance to obtain the historical tactical relevance between players. Among them, the greater the contribution weight, the more important the historical tactical relevance.

[0122] Step S24: Construct a tactical relevance matrix through the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance, calculate the average distance between two players and the ball to determine the correction coefficient, and generate a relevance matrix based on the correction coefficient to correct the tactical relevance matrix.

[0123] It can be understood that analysis is made based on the three data dimensions of spatial distance relevance, speed and acceleration relevance, and historical tactical relevance to flexibly adjust the importance of different dimensions in evaluating the cooperation relationship between players, and more comprehensively and systematically obtain the cooperation relationship between players and the ball to adapt to the changing on-field situations.

[0124] Further, in step S24, it includes:

[0125] Construct a tactical relevance matrix, denoted as ;

[0126] Calculate the average distance between two players and the ball. The corresponding calculation formula is: , where represents the average distance between two players and the ball, represents the straight-line distance between the th player and the ball; represents the straight-line distance between the th player and the ball;

[0127] Determine the correction coefficient. The corresponding calculation formula is: , where represents the correction coefficient;

[0128] Based on the correction coefficient, correct the tactical relevance matrix to generate a relevance matrix. The corresponding calculation formula is:

[0129]

[0130] where represents the relevance matrix; , , respectively represent the corrected spatial distance relevance, speed and acceleration relevance, and historical tactical relevance.

[0131] Specifically, first construct a tactical relevance matrix based on the three data of the previously obtained spatial distance relevance, speed and acceleration relevance, and historical tactical relevance; then obtain the straight-line distances between the two currently analyzed players and the ball respectively to obtain the average distance between them; secondly, if the distance between the player and the ball is far, the role in tactical play is smaller. Therefore, analyze the distance between the player and the ball comprehensively, and assist with a correction coefficient to correct the tactical relevance in three dimensions to evaluate the impact of the distance between the player and the ball on the overall tactical effect. When the average distance between the player and the ball increases linearly, the correction coefficient will also decrease linearly. Furthermore, correct the tactical relevance matrix through the correction coefficient to obtain the relevance matrix of the two players.

[0132] Further, in step S3, it includes:

[0133] Step S31: Determine the weight matrix, denoted as , where , , respectively represent the weights corresponding to the spatial distance relevance, speed and acceleration relevance, and historical tactical relevance.

[0134] As an alternative implementation, in this embodiment, the analytic hierarchy process is used to calculate the corresponding weights for the relevance matrix, that is, by constructing a hierarchical structure model, a complex decision-making problem is decomposed into multiple levels and factors. Specifically, first, a judgment matrix of the relevance matrix is constructed to evaluate the relative importance between different data elements. The corresponding logical expression is:

[0135]

[0136] where, represents the judgment matrix, and each row and column of it are the relative weight values of the spatial distance relevance, speed and acceleration relevance, and historical tactics relevance in sequence. It should be noted that, represents the importance degree of the index relevance of the first row to the index relevance of the second row, which is represented by integers from 1 to 9. The same applies to the other data elements. Then, a consistency check is performed on the judgment matrix , that is, to check whether there are logical contradictions in the judgment matrix to ensure the rationality of the judgment matrix . Finally, the judgment matrix is normalized to obtain the maximum eigenvalue and the corresponding weight matrix of the relevance matrix to reflect the relative importance between each data element in the relevance matrix, denoted as .

[0137] Step S32: Determine the tactical relevance between two players in combination with the relevance matrix. The corresponding calculation formula is:

[0138]

[0139] where, represents the tactical relevance; represents the relevance matrix.

[0140] It can be understood that through the foregoing steps, the influences of the three data dimensions of spatial distance relevance, speed and acceleration relevance, and historical tactics relevance on the dynamic changes and tactical layouts on the football field are respectively obtained. In order to more intuitively display the tactical control, it can also be represented by a graph, converting the abstract data into an intuitive visual image, that is, constructing a tactical graph to help coaches, players, and analysts better understand the tactical intentions and the on-field situation, and then optimize the tactical arrangement and execution.

[0141] Please refer to Figure 4 , which shows a schematic diagram of any possible initial tactical graph provided by an embodiment of the present invention for a method for identifying tactical graphs on a football field based on tactical relevance.

[0142] Further, in step S4, it includes:

[0143] Step S41: Construct a tactical relevance graph based on the first frame, denoted as , where represents the teams on the football field; represents the total number of teams on the football field; represents the - th set of position data of all players in the - th team; represents the set of tactical relevance degrees between any two players in the

[0144] - th team, denoted as the edge set of the tactical relevance graph.

[0145] Specifically, according to the analyzed number of frames, start constructing the tactical relevance graph from the first frame, and then obtain the initially possible tactical relevance graphs in each frame. Among them, two players form an edge, and the tactical relevance degree between the two players is used as the weight of the corresponding edge. Step S42: Select any vertex in the largest edge in the edge set as the starting point, denoted as vertex , set a screening threshold, and select from the edge set the edges that contain vertex

[0146] and whose number of vertices is greater than the screening threshold to obtain several polygons. Specifically, according to the steps of the foregoing S41, select any vertex of the two vertices in the edge with the largest tactical relevance degree in the obtained edge set as the starting point, that is, vertex , which also represents the starting point ; select more vertices to construct a polygon according to the starting point . In this embodiment, the screening threshold is 3. That is, taking vertex as the starting point, it is necessary to additionally select greater than or equal to 3 vertices to form a polygon. Conversely, if the number of selected vertices is less than 3, no polygon is formed; then sort all the selected vertices counterclockwise according to the polar angle, that is, define every three consecutive points among the selected vertices as , calculate the cross product of the vector . If these three cross products have the same sign, it means that the polygon is a convex polygon, denoted as

[0147] Step S43: Calculate the average tactical relevance degree for any polygon, and the corresponding calculation formula is:

[0148]

[0149]

[0150] where Represents the average tactical relevance of the polygon; Represents the increase coefficient; Represents the total number of players and the ball in the currently analyzed team; Represents the number of sides of the polygon; Represents the tactical relevance of the

[0151] It is explained that when the number of sides forming the polygon is large, the difficulty of analysis at this time is large, and the tactical relevance is also higher. Therefore, the increase coefficient is proposed to correct the average tactical relevance so that the obtained average tactical relevance is more accurate and it is convenient to improve the understanding ability of the subsequent constructed tactical graph.

[0152] Step S44: Sort the polygons in descending order according to the average tactical relevance, and sequentially obtain the first initial tactical graph until the initial tactical graph, where the first initial tactical graph is denoted as , and delete the position data of the players in from , and delete the tactical relevance in from . Similarly, process all polygons until the number of position data points in the set of all players in the th team is less than or equal to 2.

[0153] Specifically, based on the previously obtained average tactical relevance, sort from large to small. The polygon with the largest average tactical relevance is used as the first initial tactical graph, denoted as , and delete the position data of the players included therein and the included tactical relevance from the data elements in the total tactical relevance graph; the polygon with an average tactical relevance slightly inferior to the first initial tactical graph is denoted as the second initial tactical graph, that is , repeat the deletion step, and so on, until the position data of all players in the analyzed team in the total tactical relevance graph is less than or equal to 2 data elements, and stop the operation of step S44, that is, several initial tactical graphs are obtained, denoted as , where represents the total number of the finally obtained initial tactical graphs.

[0154] Step S45: Merge all the processed initial tactical graphs to obtain the tactical graph of any frame, denoted as , where represents the th tactical graph of the Represents the set of position data of players in all initial tactical graphics; Represents the set of tactical relevance between any two players in all initial tactical graphics.

[0155] Understandably, the foregoing steps S41 to S44 are for the initial tactical graphics obtained for the current analysis frame. All the initial tactical graphics obtained are merged to obtain the tactical graphics for the corresponding frame, and then the tactical graphics for all frames are obtained, that is , where ; It can be stated that by constructing a tactical relevance graph and performing screening, all polygons with the largest average tactical relevance can be accurately identified, denoted as the initial tactical graphics, and on this basis, they are merged to obtain the tactical graphics actually existing at each moment on the field.

[0156] Understandably, first, a coordinate system is constructed to obtain the position data of moving targets. The relevance of the spatial positions, speeds, accelerations, and historical cooperation relationships between any two players is analyzed, and the tactical relevance between players is determined by integrating the position data of the ball; by constructing the initial tactical graphics, all polygons with the largest average tactical relevance can be accurately identified, and on this basis, they are merged to obtain the tactical graphics actually existing for the corresponding team in each frame, which can effectively identify the tactics that players may execute on the field and present them in an intuitive way; it is applicable to football training and competitions with complex standings, formations, and tactics, can deeply explore the tactical connections between players, accurately identify the tactical content at each moment, that is, in each frame, and present it intuitively in a graphical way, providing a more explicit display of the on-field characteristics.

[0157] An embodiment of the present invention also proposes a football field tactical graphic recognition system based on tactical relevance, 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 football field tactical graphic recognition method based on tactical relevance as described in any one of the foregoing embodiments.

[0158] It is explained that a football field tactical graphic recognition system based on tactical relevance has the same beneficial effects as the foregoing provided football field tactical graphic recognition method based on tactical relevance, and will not be elaborated here.

[0159] 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 result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0160] 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, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for identifying tactical graphics on a football field based on tactical relevance, characterized in that: The method comprises: Construct a coordinate system based on the football field and collect the position data of any moving target of the player and the ball in each frame; According to the position data, spatial distance correlation, speed acceleration correlation and historical tactical correlation are obtained in turn to generate a correlation matrix; including: Based on the players' position data, the straight-line distance between any two players is obtained, a judgment threshold is set to determine the authenticity of the interaction between the two players, and the spatial distance correlation is calculated; The player's speed and acceleration are obtained according to the player's position data combined with the frame interval of each frame, and the speed angle and acceleration angle are respectively obtained based on the coordinate system, and the speed correlation and acceleration correlation are correspondingly calculated, and the speed correlation and acceleration correlation are combined to obtain the speed-acceleration correlation; Analyze the average tactical relevance of two players in the previous frame adjacent to the current frame, and obtain the historical tactical relevance through the contribution weight; A tactical correlation matrix is ​​constructed through spatial distance correlation, speed acceleration correlation and historical tactical correlation, the average distance between two players and the ball is calculated to determine the correction coefficient, and the tactical correlation matrix is ​​corrected based on the correction coefficient to generate a correlation matrix; Among them, the straight-line distance between any two players is obtained based on the position data of the players, a judgment threshold is set to determine the authenticity of the interaction between the two players, and the spatial distance correlation is calculated, including: Based on the players' position data, the straight-line distance between any two players is obtained. The corresponding calculation formula is: Where d represents the straight-line distance between the Ath player and the Bth player; m represents the total number of players; Set a judgment threshold l. When the straight-line distance is less than the judgment threshold, it means that the interaction between the two players is real; when the straight-line distance is greater than the judgment threshold, it means that the interaction between the two players is not real. Construct an attenuation coefficient. Calculate the spatial distance correlation, the corresponding calculation formula is: Where s1(n) represents the spatial distance correlation of the nth frame; C represents a constant; α represents the attenuation coefficient; The player's speed and acceleration are obtained according to the player's position data combined with the frame interval of each frame, and the speed angle and acceleration angle are obtained based on the coordinate system, and the speed correlation and acceleration correlation are calculated accordingly. The speed correlation and acceleration correlation are combined to obtain the speed acceleration correlation, including: According to the player's position data and the interval between each frame, the player's speed and acceleration are obtained, which are recorded as v A (n) and a A (n); Based on the coordinate system, the velocity angle and acceleration angle are obtained respectively, which are denoted as θ and β, and θ, β∈(0,π); Calculate the speed correlation, the corresponding calculation formula is: Where η1(n) represents the speed correlation of the nth frame; A represents the Ath player; B represents the Bth player; Calculate the acceleration correlation, the corresponding calculation formula is: Among them, η2(n) represents the acceleration correlation of the nth frame; The velocity correlation and acceleration correlation are combined to obtain the velocity-acceleration correlation. The corresponding calculation formula is: Among them, s2(n) represents the velocity acceleration correlation of the nth frame; λ1 represents the minimum value in s2(n), that is, s2(n)∈[λ1,1]; Analyze the average tactical relevance of the two players in the previous frame adjacent to the current frame, and obtain the historical tactical relevance through the contribution weight. The corresponding calculation formula is: s3(n)=ε×s′3(n) Among them, s3(n) represents the historical tactical relevance of the nth frame; s3 ′ (n) represents the average tactical correlation between the two players in the n-1th frame; ε represents the weight coefficient; Assign weights to the spatial distance correlation, speed acceleration correlation and historical tactical correlation respectively to obtain a weight matrix, and determine the tactical correlation between any two players in combination with the correlation matrix; Construct a tactical correlation graph, and define the tactical correlation as the weight of the corresponding edge in the tactical correlation graph. Select any vertex in the edge with the largest weight as the starting point to obtain several polygons, and filter them. Merge the filtered polygons to obtain the tactical graph of any frame.

2. A method for identifying tactical graphics on a football field based on tactical relevance according to claim 1, characterized in that: Construct a coordinate system based on the football field to collect the position data of any moving target of the players and the ball, including: Based on the football field, a plane rectangular coordinate system is constructed with the lower left corner of the top view as the origin. The camera is used to collect the position data of any moving target of the player and the ball in each frame, which is recorded as (x A (n),y A (n),sign); Among them, n=1,2,…,N represents the number of frames, and N represents the total number of frames; A=1,2,…,m represents the number of players, and m represents the total number of players; sign=0,1,…,P represents the team to which the player on the football field belongs, and P represents the total number of teams.

3. The method for identifying tactical graphics on a football field based on tactical relevance according to claim 1, characterized in that: The tactical correlation matrix is ​​constructed through spatial distance correlation, speed acceleration correlation and historical tactical correlation, the average distance between two players and the ball is calculated to determine the correction coefficient, and the tactical correlation matrix is ​​corrected based on the correction coefficient to generate a correlation matrix, including: Construct a tactical correlation matrix, recorded as Calculate the average distance between the two players and the ball. The corresponding calculation formula is: in, represents the average distance between the two players and the ball, d A represents the straight-line distance between the Ath player and the ball; d B represents the straight-line distance between the Bth player and the ball; Determine the correction coefficient, the corresponding calculation formula is: Where λ represents the correction factor; The tactical correlation matrix is ​​corrected based on the correction coefficient to generate the correlation matrix. The corresponding calculation formula is: Among them, M2 represents the correlation matrix; S1(n), S2(n), and S3(n) represent the corrected spatial distance correlation, speed acceleration correlation, and historical tactical correlation, respectively.

4. The method for identifying tactical graphics on a football field based on tactical relevance according to claim 1, characterized in that: Weights are assigned to the spatial distance correlation, speed acceleration correlation and historical tactical correlation respectively to obtain a weight matrix. The tactical correlation between any two players is determined by combining the correlation matrix, including: Determine the weight matrix, recorded as W = (w1, w2, w3), where w1, w2, w3 represent the weights corresponding to the spatial distance correlation, speed acceleration correlation, and historical tactical correlation, respectively; Combined with the correlation matrix, the tactical correlation between two players is determined. The corresponding calculation formula is: S=W·M2 Among them, S represents tactical relevance; M2 represents the relevance matrix.

5. The method for identifying tactical graphics on a football field based on tactical relevance according to claim 1, characterized in that: Construct a tactical correlation graph, and define the tactical correlation as the weight of the corresponding edge in the tactical correlation graph. Select any vertex in the edge with the largest weight as the starting point to obtain several polygons, and filter them. Merge the filtered polygons to obtain the tactical graph of any frame, including: Construct a tactical correlation graph based on the first frame, denoted as G p (1) = (V, E), p = 1, 2, ..., P, where p represents the team on the football field; P represents the total number of teams on the football field; V represents the set of position data of all players in the p-th team; E represents the set of tactical correlations between any two players in the p-th team, recorded as the edge set of the tactical correlation graph; According to the edge set E, select any vertex in the largest edge as the starting point, record it as vertex A, set the screening threshold, and select the edges containing vertex A and whose number of vertices is greater than the screening threshold from the edge set E to obtain several polygons; Calculate the average tactical relevance for any polygon, and the corresponding calculation formula is: in, represents the average tactical relevance of the polygon; μ represents the increase coefficient; m represents the total number of players and balls in the current analyzed team; I represents the number of edges of the polygon; S h represents the tactical relevance of the hth edge in the polygon; According to the average tactical relevance, the polygons are sorted in descending order, and the first initial tactical graph is obtained in sequence, until the Qth initial tactical graph, where the first initial tactical graph is recorded as Delete the position data of the players in V1 from V, delete the tactical relevance in E1 from E, and process all polygons in the same way until the number of position data points in the set V of the position data of all players in the p-th team is less than or equal to 2; All the processed initial tactical graphs are merged to obtain the tactical graph of any frame, which is denoted as G com (n)=(V com ,E com ), where G com (n) represents the tactical graph of the p-th team in the n-th frame; V com Represents the set of position data of all players in the initial tactical graph; E com Represents the set of tactical correlations between any two players in all initial tactical graphs.

6. A tactical pattern recognition system on a football field based on tactical relevance, 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, the steps of the method for recognizing tactical graphics on a football field based on tactical relevance as described in any one of claims 1 to 5 are implemented.

Images