RU2599699C1 - Method of detecting and analysing competition game activities of athletes - Google Patents

Abstract

FIELD: sports.

SUBSTANCE: method relates to physical education and sports and is intended for detecting and analysis of competition game activities of athletes. Method of detecting and analysing competition game activities of athletes, involving using video cameras in real time recording objects (players, ball, disc), and subsequent rapid analysis of obtained image in data medium in central image processing module, including statistical, analytical and graphical modules, data from which are transmitted to an operational applications server, characterised by that, method includes recording and analysing competitive game action based on an integral parameter - command game space (CS) including generalised dynamic characteristics: rate of change of command game space (CS_v); figure of command game space (CS_f), and partial dynamic characteristics: command game space of attack (CS_i); command game space for defense (CS_p); overlap area (CS_ip); figure of command game space for attack (CS_fi); figure of command game space for defense (CS_fp); rate of change of command game space for attack (CS_vi), rate of change of command game space for defense (CS_vp), dynamic characteristics are processed in analytical module and converted into graphic game circuits, game situations, as well as in interaction and ratio of objects (players, ball, disc) in a graphic module, with further indexing, storage and search for statistical, analytical and graphic information on operational applications server.

EFFECT: technical result is recording competition game activities with subsequent determination of their efficiency on basis of integral parameter - command game space; score (comparison) of private dynamic characteristics of integral parameter to detect most effective for individual players, commands of various qualification, age and sex; possibility of operational development of recommendations during competition to improve competition game activities, interactions of athletes; possibility of development of situational exercises and new tactical combinations based on obtained data on integral parameter; creating animation with automatic graphic tracking of game, action of athletes in real time; creation of graphic images with 2D and 3D models of dynamic characteristics for mapping examination of game situations; creation of reports containing statistical, analytical and graphic information on competitive game activities of athletes; creation of a data base on basis of indexing, storage and retrieval of video events( automatic indexing and searching for a separate video fragment with game situation) with subsequent storage in a database on a web server or personal archive.

1 cl, 7 dwg

Description

The method relates to the field of physical education and sports and is intended for registration and analysis of competitive game actions of athletes.

Among modern means of recording the actions of athletes, optical devices are known that receive a signal from electromagnetic sensors attached to athletes. The image is processed by “passive” methods of analysis of non-real-time competition using video repetitions. These funds are presented in US patents (No. 5.513.854 - http://www.google.com/patents/US 5513854; No. 6124862 - http://www.google.com/patents/US 6124862; No. 6483511 - http: // www.google.com/patents/US 6483511).

A known system for tracking, identifying and capturing the movements of athletes and other objects in real time sports competitions using "non-intrusive" peripheral methods (obtaining images based on gamma radiation). The registration of actions is carried out using cameras located along the perimeter of the playing field, and the information obtained is quickly processed in a computer environment (US patent No. 20080192116 A1 - http://www.google.de/patents/US 20080192116, which is the prototype of the present invention).

The main disadvantage among the known means and registration systems is the lack of integrated parameters for evaluating game actions that allow you to quickly determine the effectiveness of a player and a sports team.

Among the known methods of analysis, calculation and rapid assessment of quantitative and qualitative parameters of competitive game actions, determining the structure of competitive activity, we can distinguish pairwise methods for comparing the real values of game indicators with their model characteristics (Amalin M.E. Methodology for evaluating competitive activity in sports games // Theory and the practice of physical culture. - 1980. - No. 9. - P. 19-22; Tsimbalyuk V.A. Evaluation of the effectiveness of competitive activity in the process of managing technical active training of a highly qualified team (for example, a major league team): abstract of thesis ... candidate of pedagogical sciences. - Omsk, 1985. - 24 pp .; Suvorov VV Technical training of young football players based on the structure of competitive activity: abstract of dissertation ... candidate of pedagogical sciences. - Krasnodar, 1996. - 24 p.). These techniques reveal the percentage of technical and tactical actions rejects per game, and when comparing them with the model indicators of team play, there is a difficulty with a qualitative assessment of individual and team actions in dynamics.

The Sands scale is known (Yakhontov E.R. Sands Scale // Sports Games. - 1971. - No. 5. - P. 32), based on the analysis of statistical indicators of game actions. In the calculation, each fixed game indicator is estimated by a unit. The effectiveness of free throws and shots from the game is compared with the average team score. However, in this case, it is not foreseen to identify the interconnections between private indicators of player actions, which excludes a qualitative analysis of competitive gaming activity.

There are known methods of analysis and evaluation of game actions according to Frank Mac Guire (Lindeberg F. Basketball. - M .: FiS, 1971. - P. 82-85), as well as according to Oliver, Dean. Basketball on paper: rules and tools for performance analysis. - Brassey′s, Inc. - 376 p). These methods have much in common with the Furukawa index (Tailors Yu.M. Furukawa index // Sports games. - 1982. - No. 6. - P. 24-25) and the method of determining the reliability index (Lyapin N. Reliability index // Sports games . - 1983. - No. 4. - S. 20-21). These approaches to determine the effectiveness of the actions of athletes and teams summarize the importance of private gaming indicators and confirm the final results of the games.

In the presented methods of analysis and evaluation of competitive game actions, the lack of integral parameters taking into account the specifics and dynamics of the transformation of game situations negatively affects the objectivity of the interpretation of the results, which does not allow us to fully establish causal relationships between the activities of athletes and the result. As a result, it is not known which of the parameters is more important at a certain moment of the game and what is the value of their contribution to the effectiveness of competitive gaming activity.

Technical task: increasing the effectiveness of competitive gaming actions of each player and sports team based on their registration and analysis.

The essence of the invention lies in the fact that the method of registration and analysis of competitive gaming actions of athletes, including the use of real-time video cameras recording objects (players, a ball, puck), and subsequent operational analysis of the received image in the information environment in the central image processing module, which includes statistical, analytical and graphical modules, the data from which are sent to the server of operational applications, where indexing, storage and search for video files with games takes place or in separate game episodes, it is distinguished by the fact that competitive game actions are recorded and analyzed on the basis of an integral parameter - the command area of the game (hereinafter referred to as CS), including generalized dynamic characteristics: the rate of change of the command area of the game (hereinafter CS _v ); the figure of the command area of the game (hereinafter referred to as CS _f ), and particular dynamic characteristics: the command area of the attack game (hereinafter referred to as CS _i ); command area of the defense game (hereinafter referred to as CS _p ); overlap area (hereinafter referred to as CS _ip ); rate of change of the command area of the attack game (hereinafter referred to as CS _vi ); rate of change of the command area of the defense game (hereinafter referred to as CS _vp ); the figure of the command area of the attack game (hereinafter - CS _fi ); the figure of the command area of the defense game (hereinafter referred to as CS _fp ), then the dynamic characteristics are converted into graphic game schemes, game situations, as well as into the interactions and relations of objects (players, ball, puck) in the graphic module with subsequent indexing, storage and search statistical, analytical and graphical information on the server of operational applications.

A brief description of the drawings and diagrams.

In FIG. 1 is a diagram of the location of the cameras above the basketball court, where 1 is camera No. 1; 2 - camera No. 2; 3 - camera No. 3; 4 - camera number 4.

In FIG. 2 shows a diagram of the information environment, which consists of the following modules: 5 - a central image processing module; 6 - statistical; 7 - analytical; 8 - graphic; 9 - server operational applications. The diagram outlines the sequence of processing data obtained as a result of registration by cameras of competitive game actions of athletes.

In FIG. 3 shows CS _i in projection onto the surface of the site, where i1; i2; i3; i4; i5 - attacking players; 2 - video camera No. 2; 10 - CS _i ; 11 - line of attack, drawn through the player inside CS _i from the side of the ring; 12 - basketball hoop.

In FIG. 4 shows the interaction of CS _i and CS _p to form CS _ip , where i1; i2; i3; i4; i5 - attacking players; p1, p2, p3, p4, p5 - defense players; 2 - video camera No. 2; 10 - CS _i ; 13 - CS _p ; 14 - CS _ip .

In FIG. 5 shows typical CS _fi : 15 - triangle; 16 - rectangle; 17 - trapezoid; 18 - pentagon.

In FIG. Figure 6 shows a graph according to which you can track the change in CS speed of teams 1 and 2 (K1, K2) with the display of technical and tactical parameters in the form of conditional information markers.

In FIG. 7 shows the sequence for determining the effectiveness of competitive gaming actions of athletes in the analytical module of the information environment.

The present method can be used for the following team games: basketball, handball, futsal, soccer, hockey, ball hockey, etc. The description given hereinafter is made using basketball as an example.

The method is as follows.

The registration of competitive gaming activities is carried out in real time using four cameras located above the playground along the perimeter of the sports facility. Two cameras (1) and (2) are located above the basketball hoops and two (3) and (4) in the zone of intersection of the central and side lines on different sides of the playing court (Fig. 1). This allows you to capture all the actions of athletes in the frame.

After registration of competitive gaming actions by athletes with video cameras, they are analyzed based on an integral parameter that includes generalized and particular dynamic characteristics. This analysis is aimed at determining the effectiveness of competitive gaming actions of athletes.

Registration and analysis of competitive gaming actions of athletes is carried out in real time during the match and (or) after it.

Integral parameter - CS includes generalized dynamic characteristics: CS _v ; CS _f , and particular dynamic characteristics: CS _i ; CS _p ; CS _ip ; CS _vi ; CS _vp ; CS _fi CS _fp . Thus, in the claimed method for registering and analyzing competitive gaming actions, they proceed from a continuum of particular dynamic characteristics of the integral parameter.

By CS is meant the area of a figure formed by a polygon whose vertices correspond to the positions of five attack or defense players in projection onto the surface of the playing court (Fig. 4). In this case, the lines of attack or defense (11) (Fig. 3) from the side of the ring (12) are always drawn through a player located inside the perimeter of the figure of the remaining four players. Acceptance of this condition is dictated by the fact that up to 60-70% of attacks, of which 40-45%) occur in a limited area of the court (three-second zone) are successful (Kozin V. Simulation of competitive activity basketball players // Materials for 8-a international Scientific and practical conference, “Inquire of the World Widely Asked for Science”, 2012. - P. 41-46). Accordingly, for defense players, this zone is the most “vulnerable” and priority, which encourages coaches to build tactical defense schemes with a focus on the concentration of players in a limited area of the site.

As a result of players moving around the court, CS _i (10) and CS _p (13) “overlap” each other, forming CS _ip (14) - the total area for CS _i and CS _p (Fig. 4). With a decrease in CS _ip , a conclusion is drawn about a decrease in the activity of defensive and attacking actions and interactions, and, on the contrary, with an increase in CS _ip , an increase in activity along with the number of game tricks used. A decrease in CS _ip leads to an increase in free (without resistance) throws on the ring.

When players move around the court, CS _i and CS _p form geometric shapes - a triangle (15), a rectangle (16), a trapezoid (17), a pentagon (18) (Fig. 5). Hence CS _f is the relative position of the borders (lines) of the figure, as well as the relative position of the points of the line. Non-standard CS is considered when it is unlike any of the simple geometric shapes.

When registering CS _f in the form of a triangle, the low effectiveness of competitive gaming actions is noted. At the same time, players more often use group selection of the ball and the safety net of another defender. CS _fs such as trapezoid, rectangle and polygon provide rational use and a combination of individual and team interactions. Non-standard CS _{f are} considered as special cases and are grouped separately as the database is formed and information about them is systematized. Simultaneously with the registration of CS _fi and CS _fp in the information environment, their values are calculated.

With a constant value of CS _p and a simultaneous increase in CS _{i, the} effectiveness of protective actions decreases. On the contrary, with a constant value of CS _i and a simultaneous increase in CS _{p, the} effectiveness of protective actions increases. Moving away from the player’s ring an attack or defense that is passive within the perimeter of a four-player piece reduces the CS size and negatively affects the effectiveness of the team’s actions.

According to the data (http://stats.nba.com/tracking/#!/player/?ls=iref:nba:gnav) presented on the website of the National Basketball Association, the average speed of a player moving around the court is 1.4-2, 3 m / s, while the maximum values reach 8.3 m / s. Hence, the possible ranges of variation of CS _v are from 0 m ² / s (players are in a static position, do not move around the playing court) to 27 m ² / s (5 players equally spaced on the center circle line of the team move from the center of the circle in the direction of the corresponding radius at the same speed of 2 m / s) or up to 102 m ² / s (5 players equally spaced on the center circle line of the team move from the center of the circle in the direction of the corresponding radius with the same speed of 5 m / s). After carrying out the necessary calculations, the computer builds graphs of changes in the indicators of competitive activity, on which quantitative indicators of competitive game activity (the number of transfers, throws, stops, etc.) are displayed using information markers, which allows us to study the effect of CS _v on their implementation (Fig. 6). Teams with high CS _v during the game show great performance and more often rivals win. It should be noted the correlation between CS _v and the effectiveness of individual, group and team competitive game actions.

Data obtained as a result of registration of competitive gaming actions of athletes is processed in the central (5), statistical (6), analytical (7) and graphic (8) modules of the information environment (Fig. 2). The data obtained is used to identify the effectiveness and improve the competitive game actions of athletes, teams; developing and improving situational exercises, creating game schemes in a graphical environment, as well as creating a database based on indexing, storing and searching for video events, followed by saving them in a database on a web server or in a personal archive.

In the process of observation, the video signal from the cameras enters a personal computer, namely the central image processing module (5), in parallel there is segmentation, localization and identification of 2D and 3D objects (Fig. 2). Segmentation and localization is carried out using algorithms known and described in the research papers and references therein (J. Pers and S. Kovacic, "A system for tracking players in sports games by computer vision", Electrotechnical Review 67: 281-288, 2000 ; T. Matsuyama and N. Ukita, "Real time multi target tracking by a cooperative distributed vision system", Dept. of Intelligent Science and Technology, Kyoto University, Japan). The location and real-time tracking of the position of the head, hands and other parts of the body, as well as the recognition of gestures of an individual person based on video image analysis, are described in studies and references therein (C. Wren et al. “Pfinder: real time tracking of the human body ", IEEE Transactions on Pattern Analysis and Machine Intelligence, 19: 780-785, 1997; A. Aagarwal and B. Triggs" 3D human pose from silhouettes by relevance vector regression ", International Conference on Computer Vision & Pattern Recognition, pages II 882 -888, 2004). Players are identified automatically or with the help of an operator (in case of technical problems), based on the rules of basketball and the logic of recognizing objects, numbers and other applied patterns, logos on the player’s uniform using the methods described in the following US patents (No. 5353392 - http : //www.google.com/patents/ US 5353392; No. 5264933 - http://www.google.com/patents/US 5264933). In the claimed method, in addition to identifying players, they track the movement in space of such material objects as a ball, puck.

From the central module (5), the processed data arrives at the statistical (6), analytical (7) and graphic (8) modules. In the statistical module (6), based on the results of identification, segmentation and localization, the game indicators are archived - ball throw, method of throw, throw distance, distance between the attacker and the defender, speed of movement of players, direction of movement of players, etc., described in detail in ( VV Kozin. Prerequisites for the creation of visual software for competitive activities in basketball. Materials of the conference "Young Russia: advanced technologies in industry", S. 263-266, 2010) (Fig. 2).

In the analytical module (7), the effectiveness of competitive gaming actions of athletes is determined based on the analysis of CS. The sequence for determining the effectiveness of competitive gaming actions of athletes is shown in FIG. 7.

In module (8), a graphic image is generated (Fig. 2) in real time using well-known algorithms presented in (N. Sidenbladh, M. Black and D. Fleet, “Stochastic tracking of 3D human figures using 2D image motion "In Proc. Of the European Conference On Computer Vision, pages 702-718, 2000). This allows you to get a graphical 2D and / or 3D image and schematically represent the playing area with particular dynamic characteristics, the location of the players, and also the ball or puck in the form of icons and other spatial objects (lines, points).

The received data comes from each module to the operational application server (9), where indexing, storage and retrieval of reports, statistical, analytical and graphical information, video files with games or individual game episodes take place.

The technical result is: registration of competitive game actions with the subsequent determination of their effectiveness on the basis of an integral parameter - the command area of the game; assessment (comparison) of the partial dynamic characteristics of the integral parameter with the identification of the most effective for individual players, teams of various qualifications, age and gender; the ability to quickly develop recommendations during the competition to improve competitive gaming actions, the interactions of athletes; the possibility of developing situational exercises and new tactical combinations based on the obtained data on the integral parameter; creation of animations with automatic graphic accompaniment of the game, the actions of athletes in real time; creation of graphic images with 2D and 3D models of dynamic characteristics for a schematic analysis of game situations; creation of reports containing statistical, analytical and graphical information about the competitive gaming activity of athletes; creating a database based on indexing, storing and searching for video events (automatic indexing and searching for a particular video fragment with a game situation), followed by saving it in a database on a web server or in a personal archive.

Claims (1)

A method for recording and analyzing competitive gaming actions of athletes, including the use of real-time video cameras recording objects (players, ball, puck), and subsequent operational analysis of the received image in the information environment in the central image processing module, which includes statistical, analytical and graphic modules , data from which go to the server of operational applications, characterized in that they register and analyze competitive gaming actions based on the integral parameter - the command area of the game (CS), including generalized dynamic characteristics: the rate of change of the command area of the game (CS _v ); figure of the command area of the game (CS _f ), and particular dynamic characteristics: command area of the attack game (CS _i ); command area of the defense game (CS _p ); overlap area (CS _ip ); Attack game area figure (CS _fi ); figure of the command square of the defense game (CS _fp ); the rate of change of the command area of the attack game (CS _vi ), the rate of change of the command area of the defense game (CS _vp ), the dynamic characteristics are processed in the analytical module and converted into graphic game schemes, game situations, as well as in the interactions and ratios of objects (players, the ball, washers) in the graphics module with subsequent indexing, storage and retrieval of statistical, analytical and graphical information on the operational application server.

Images