Sports item result generation system and method based on artificial intelligence and visual technology

Abstract

The invention discloses an athletic event result generation system and method based on artificial intelligence and a visual technology, relates to the technical field of artificial intelligence and computer vision, and solves the problems that in the prior art, the condition that whether an athletic event can be smoothly carried out is ignored, scoring standards of different events are different, and feedback suggestions cannot be given to athletes in time. And the accuracy and the comprehensiveness of result generation are relatively low. A project environment influence coefficient is generated according to a project ID and environment data; an exercise result is generated according to the exercise video corresponding to the item type and the athlete ID, the environment state in the exercise area is monitored in real time, and whether the exercise item can be normally carried out or not is judged in a self-adaptive mode according to different requirements of different exercise items for the environment; according to different scoring standards of sports items, scores in different items are accurately calculated, sports evaluation and personalized suggestions are given to the sports items, and the comprehensiveness of sports item result generation is improved.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence and computer vision technology, and specifically is a system and method for generating sports event results based on artificial intelligence and vision technology. Background Art

[0002] "Physical education" occupies an important position in the five aspects of "moral, intellectual, physical, aesthetic and labor education". It not only refers to physical health, but also emphasizes the promotion of students' all-round development through physical exercise. Strengthening physical education can not only enhance students' physical fitness, but also cultivate students' sound personality and good spiritual qualities, laying a solid foundation for students' lifelong development. In various activities of sports events, whether it is training or examination, results are needed. The accuracy and comprehensiveness of the results are of great significance to the evaluation and guidance of participants.

[0003] The traditional way of generating sports results mainly relies on manual timing, scoring and recording, which has problems such as low efficiency, easy errors and strong subjectivity. With the rapid development of artificial intelligence, existing technologies also use artificial intelligence technology to generate sports results, but ignore the conditions for the smooth progress of sports. At the same time, there are differences in the scoring standards of different projects and the failure to provide timely feedback and suggestions to athletes, which makes the accuracy and comprehensiveness of the results generated low; for example, different sports have different requirements for the environment, and some sports may not be able to proceed smoothly under the same environment; therefore, the sports result generation system still needs further improvement. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art; to this end, the present application proposes a sports event result generation system and method based on artificial intelligence and visual technology, which is used to solve the technical problems that the prior art ignores the conditions for whether the sports event can be carried out smoothly, there are differences in the scoring standards of different events, and fails to provide athletes with timely feedback and suggestions, resulting in low accuracy and comprehensiveness of the result generation.

[0005] To achieve the above-mentioned purpose, the first aspect of the present application provides a sports event result generation system based on artificial intelligence and visual technology, including: a data acquisition module, a data analysis module, an early warning module and a database;

[0006] The data acquisition module: acquires athlete data, sports event data and environmental data through data acquisition equipment; the sports event data includes the event ID and its corresponding event detection video and event type; the athlete data includes the athlete ID and its corresponding physical condition;

[0007] The data analysis module generates a project environmental impact coefficient according to the project ID and environmental data; generates an alarm signal according to the project environmental impact coefficient; generates a sports video after identifying the athlete ID according to the project detection video; generates a sports result according to the sports video corresponding to the project type and the athlete ID;

[0008] The early warning module: makes prompts according to the alarm signal and contacts the management personnel.

[0009] Through the above steps, this application monitors the environmental status of the sports area in real time, and based on the environmental requirements of different sports events, adaptively determines whether the events can be carried out normally, and provides accurate data support for the sports events; at the same time, according to the scoring standards of different events, a multi-branch network is constructed to calculate the scores of athletes in each event in real time, and provide sports evaluation and personalized suggestions, thereby improving the comprehensiveness of the generation of sports event results.

[0010] Furthermore, generating the project environmental impact coefficient according to the project ID and environmental data includes:

[0011] Obtaining a project ID and environmental data; the environmental data includes several environmental parameters;

[0012] By formula Calculate the project environmental impact coefficient XHYX i ; Among them, α ij is the jth weight coefficient corresponding to the i-th project ID; the weight coefficient is generated according to the project ID; HC j Expressed as several environmental parameters; DHC j Expressed as unit environmental parameters corresponding to several environmental parameters; ZHC ij It is expressed as the optimal environmental parameter value corresponding to several environmental parameters of each sport.

[0013] Furthermore, the weight coefficient is generated according to the project ID, including:

[0014] Get the project ID and its corresponding environmental impact preference HYP ij ; The environmental impact preference is assessed by experts based on the impact of the project ID on several environmental parameters;

[0015] By formula Calculate the weight coefficient α ij .

[0016] Furthermore, the generating of an alarm signal according to the project environmental impact coefficient includes:

[0017] Obtain environmental impact coefficients for several projects;

[0018] Determine whether the project environmental impact coefficient is greater than its corresponding project environmental threshold;

[0019] Yes, generate a sports event stop alarm signal;

[0020] If no, determine whether the project environmental impact coefficient is greater than its corresponding D times project environmental threshold; if yes, generate a warning signal for a poor sports project environment; if no, do nothing; wherein D is the proportional coefficient, D∈(0,1).

[0021] Furthermore, the step of identifying the athlete ID according to the event detection video and generating a sports video therefrom includes:

[0022] Get project inspection video;

[0023] Inputting the event detection video into a target detection model to obtain the athlete ID; the target detection model is constructed by an artificial intelligence model;

[0024] Use target tracking algorithm to track the detected athlete ID and record its movement trajectory;

[0025] The project detection video is processed into frames to obtain several frames of detection images;

[0026] Obtain several frames of detection images corresponding to each athlete ID, and splice them in chronological order to obtain the sports video corresponding to the athlete ID.

[0027] Through the above steps, this application first identifies the athlete IDs participating in the sports event, and extracts the sports video corresponding to each athlete ID from the event detection video based on the athlete ID, providing accurate data support for the subsequent calculation of the sports event results corresponding to each athlete ID.

[0028] Furthermore, the target detection model is constructed by an artificial intelligence model, including:

[0029] Obtain several historical event detection videos and their corresponding athlete IDs;

[0030] Divide several historical event detection videos and their corresponding several athlete IDs into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain a training set, a verification set and a test set;

[0031] Choose artificial intelligence as the base model;

[0032] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model;

[0033] By verifying the pre-trained model on the test set, we finally get a target detection model that inputs the event detection video and outputs the athlete ID.

[0034] Furthermore, the generating of the sports result according to the sports video corresponding to the event type and the athlete ID includes:

[0035] Obtain sports videos and physical conditions corresponding to the project type and athlete ID; the project type includes result type, process type and dual type;

[0036] Inputting the athlete ID, sports video and event type into a score prediction model to obtain a final score; the score prediction model is obtained by constructing a multi-task branch;

[0037] The athlete ID, physical condition and sports video are input into the evaluation suggestion generation model to obtain sports evaluation and sports suggestions; the evaluation suggestion generation model is constructed through a deep learning model.

[0038] Furthermore, the rating prediction model is obtained by constructing a multi-task branch, including:

[0039] Get several historical athlete IDs, historical sports videos and corresponding event types and historical scores;

[0040] Divide several historical athlete IDs, historical sports videos, corresponding event types and historical scores into several subtasks according to event types; divide the data in several subtasks into several training data, verification data and test data; perform data preprocessing on several training data, verification data and test data to obtain several training sets, verification sets and test sets;

[0041] Select the multi-task learning model as the base model;

[0042] Train the basic model through several training sets, and adjust the learning rate and hyperparameters on several validation sets to obtain the pre-trained model;

[0043] By verifying the pre-trained model on several test sets, we finally obtain a score estimation model that inputs motion videos and outputs the final score.

[0044] Furthermore, the evaluation suggestion generation model is constructed through a deep learning model, including:

[0045] Obtain several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions;

[0046] Divide several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain a training set, a verification set and a test set;

[0047] Select a deep learning model as the base model;

[0048] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model;

[0049] By verifying the pre-trained model on the test set, we finally obtained an evaluation and suggestion generation model that inputs athlete ID, physical condition and sports video, and outputs the sports evaluation and sports suggestion corresponding to the athlete ID.

[0050] Another aspect of the present invention provides a method for generating sports event results based on artificial intelligence and visual technology, comprising:

[0051] S0: Acquire athlete data, sports event data and environmental data; the sports event data includes the event ID and its corresponding event detection video, event standard specification and event type; the athlete data includes the athlete ID and its corresponding physical condition;

[0052] S1: Generate a project environmental impact coefficient based on the project ID and environmental data; generate an alarm signal based on the project environmental impact coefficient;

[0053] S2: after identifying the athlete ID according to the event detection video, generate a sports video; generate a sports result according to the sports video corresponding to the event type and the athlete ID;

[0054] S3: Prompt according to the alarm signal and contact the management.

[0055] Compared with the prior art, the beneficial effects of this application are:

[0056] 1. The present application generates a project environmental impact coefficient according to the project ID and environmental data; generates an alarm signal according to the project environmental impact coefficient; generates a sports video based on the athlete ID after identifying the athlete ID according to the project detection video; generates a sports result according to the sports video corresponding to the project type and the athlete ID, monitors the environmental status in the sports area in real time, and adaptively determines whether the sports project can be carried out normally according to the different requirements of different sports for the environment, and provides accurate data support for the progress of the sports project; at the same time, according to the different scoring standards of the sports projects, constructs a multi-branch network to calculate the score of each athlete in different projects in real time, and gives them sports evaluation and provides personalized suggestions according to their own status, thereby improving the comprehensiveness of the generation of sports project results.

[0057] 2. This application calculates the impact coefficient of the environment on all sports events according to the event ID, and takes into account that different sports events are affected differently by environmental parameters, and adopts different weight coefficients for several environmental parameters in different events, so as to meet the connection between sports events and environmental parameters, provide accurate data support for the conduct of sports events, and improve the comprehensiveness of the entire sports event result generation system.

[0058] 3. This application obtains the final score corresponding to the athlete ID according to the sports video corresponding to the project type and the athlete ID, and gives the corresponding sports evaluation and sports suggestions for the sports video of the athlete ID; when calculating the final score, it takes into account that the final score of the sports event is affected by the project type corresponding to the sports event, and constructs a score estimation model for multi-branch tasks to accurately obtain the final score, thereby improving the scoring efficiency and accuracy; at the same time, it provides personalized evaluation and suggestions for athletes to help their subsequent development. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 This is a schematic diagram of the principle of the sports event result generation system based on artificial intelligence and visual technology of this application;

[0061] Figure 2 A flow chart for generating an alarm signal for this application;

[0062] Figure 3 This is a flow chart of the method for generating sports event results based on artificial intelligence and visual technology in this application. DETAILED DESCRIPTION

[0063] The technical solution of the present application will be described clearly and completely in conjunction with the embodiments below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0064] See also Figure 1 , the first aspect of the present application provides a sports event result generation system based on artificial intelligence and visual technology, including: a data acquisition module, a data analysis module, an early warning module and a database;

[0065] Data acquisition module: acquire athlete data, sports event data and environmental data through data acquisition equipment; sports event data includes event ID and its corresponding event detection video and event type; athlete data includes athlete ID and its corresponding physical condition; data acquisition equipment includes high-precision cameras and various sensors, etc.; event detection video refers to the video of each sport during the competition or examination;

[0066] Data analysis module: Generate the project environmental impact coefficient according to the project ID and environmental data. The project environmental impact coefficient refers to the degree of impact of environmental data on sports events; generate an alarm signal according to the project environmental impact coefficient; generate a sports video after identifying the athlete ID according to the project detection video, and each athlete ID has its corresponding sports video; generate sports results according to the sports video corresponding to the project type and athlete ID; the sports results include final scores, sports evaluations and sports suggestions;

[0067] Early warning module: Prompt according to the alarm signal and contact the management personnel; the alarm signal includes the alarm signal for stopping the sports project and the alarm signal for the poor sports environment.

[0068] In this embodiment, the project environmental impact coefficient is generated according to the project ID and environmental data, including:

[0069] Get project ID and environmental data; environmental data includes several environmental parameters; environmental parameters include temperature, humidity, wind speed and rainfall, etc.;

[0070] By formula Calculate the project environmental impact coefficient XHYX i ; Among them, α ij is the jth weight coefficient corresponding to the i-th project ID; the weight coefficient is generated according to the project ID; HC j Expressed as several environmental parameters; DHC j Expressed as unit environmental parameters corresponding to several environmental parameters; ZHC ij It is expressed as the optimal environmental parameter value corresponding to several environmental parameters of each sport; the weight coefficient of each sport is different. In a certain sport, the weight coefficient of several environmental parameters is fixed. Once the environmental parameter value deviates from its corresponding optimal environmental parameter value, the environmental data will have a greater impact on the sport; therefore, the project environmental impact coefficient will increase accordingly.

[0071] The weight coefficient in this embodiment is generated according to the project ID, including:

[0072] Get the project ID and its corresponding environmental impact preference HYP ij; Environmental impact preference is assessed by experts based on the impact of the project ID on several environmental parameters; the greater the impact of the environmental parameters in the environmental data on the sport, the greater the corresponding environmental impact preference;

[0073] By formula Calculate the weight coefficient α ij The weight coefficient of several environmental parameters in any sport is equal to the ratio of its environmental impact preference to the sum of the environmental impact preferences of all environmental parameters corresponding to the sport.

[0074] Through the above steps, this embodiment calculates the influence coefficient of the environment on all sports events, and sets differentiated weight coefficients for key environmental parameters in each event based on the differences in sensitivity of different events to environmental parameters, thereby accurately reflecting the correlation between sports events and environmental parameters, providing reliable data support for the development of sports events, and improving the comprehensiveness of the sports event result generation system.

[0075] See also Figure 2 In this embodiment, generating an alarm signal according to the project environmental impact coefficient includes:

[0076] Obtain environmental impact coefficients for several projects;

[0077] Determine whether the project environmental impact coefficient is greater than its corresponding project environmental threshold; the project environmental threshold is set based on experience;

[0078] Yes, generate a sports event stop alarm signal;

[0079] No, determine whether the project environmental impact coefficient is greater than its corresponding D times project environmental threshold; Yes, generate a warning signal for poor sports project environment; No, do nothing; wherein D is the proportional coefficient, D∈(0,1), the specific value is set according to experience, and in this embodiment, D is set to 0.7.

[0080] In this embodiment, the sports video is generated by identifying the athlete ID according to the event detection video, including:

[0081] Get project inspection video;

[0082] The event detection video is input into the target detection model to obtain the athlete ID; the target detection model is constructed through the artificial intelligence model;

[0083] A target tracking algorithm is used to track the detected athlete ID and record its movement trajectory; the target tracking algorithm includes the deepsort algorithm, etc.;

[0084] The project detection video is frame-processed to obtain a plurality of frames of detection images; in this embodiment, a 1-second project detection video can be divided into 90 frames, and in other embodiments, it can be divided into 60 frames or 120 frames, etc.;

[0085] Obtain several frames of detection images corresponding to each athlete ID, and splice them in chronological order to obtain the sports video corresponding to the athlete ID.

[0086] Through the above steps, this embodiment first identifies the ID of each athlete in the sports event, and extracts the independent sports video corresponding to each athlete from the event detection video based on the identification result, providing a reliable data basis for the subsequent accurate calculation of the event results of each athlete.

[0087] The target detection model in this embodiment is constructed by an artificial intelligence model, including:

[0088] Obtain several historical event detection videos and their corresponding athlete IDs;

[0089] Divide several historical event detection videos and their corresponding several athlete IDs into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain training set, verification set and test set; the ratio between the training set, test set and verification set is 7:2:1;

[0090] Select artificial intelligence as the basic model; artificial intelligence models include Yolo model, etc.

[0091] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model;

[0092] By verifying the pre-trained model on the test set, we finally get a target detection model that inputs the event detection video and outputs the athlete ID.

[0093] In this embodiment, the sports results are generated according to the sports video corresponding to the event type and the athlete ID, including:

[0094] Get the sports video and physical status corresponding to the project type and athlete ID; project types include result type, process type and dual type; the result type means that the final score of the sports project only focuses on the result, not the process, such as shooting and archery; the process type means that the final score of the sports project only focuses on the process, not the result, such as aerobics and cheerleading; the dual type means that both have an impact on the final score, such as diving;

[0095] The athlete ID, sports video and event type are input into the score estimation model to obtain the final score; the score estimation model is obtained by constructing a multi-task branch;

[0096] The athlete ID, physical condition and sports video are input into the evaluation suggestion generation model to obtain sports evaluation and sports suggestions; the evaluation suggestion generation model is constructed through a deep learning model; the physical condition is taken into consideration in the process of generating evaluation suggestions in order to make an objective evaluation based on the physical condition and to give targeted suggestions.

[0097] Through the above steps, this embodiment outputs the final score of the athlete ID and generates sports evaluation and suggestions based on its sports video; in the score calculation, a multi-branch task score estimation model is constructed according to the impact of different project types on the score, which effectively improves the scoring efficiency and accuracy; at the same time, by providing personalized evaluation and suggestions, professional guidance is provided for the subsequent development of athletes.

[0098] The score estimation model in this embodiment is obtained by constructing a multi-task branch, including:

[0099] Get several historical athlete IDs, historical sports videos and corresponding event types and historical scores;

[0100] Divide several historical athlete IDs, historical sports videos, corresponding event types and historical scores into several subtasks according to event types; divide the data in several subtasks into several training data, verification data and test data; perform data preprocessing on several training data, verification data and test data to obtain several training sets, verification sets and test sets; the ratio between training set, test set and verification set is 7:2:1;

[0101] Select a multi-task learning model as the basic model; the multi-task learning model includes a multi-task learning model based on an attention mechanism, such as the MTAN model and the PLE model;

[0102] Train the basic model through several training sets, and adjust the learning rate and hyperparameters on several validation sets to obtain the pre-trained model;

[0103] By verifying the pre-trained model on several test sets, we finally obtain a score estimation model that inputs motion videos and outputs the final score.

[0104] The evaluation suggestion generation model in this embodiment is constructed through a deep learning model, including:

[0105] Obtain several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions;

[0106] Several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions are divided into training data, verification data and test data; data preprocessing is performed on the training data, verification data and test data to obtain training sets, verification sets and test sets; the ratio between the training set, the test set and the verification set is 7:2:1;

[0107] Select a deep learning model as the basic model; deep learning models include large language models, such as the T5 model;

[0108] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model;

[0109] By verifying the pre-trained model on the test set, we finally obtained an evaluation and suggestion generation model that inputs athlete ID, physical condition and sports video, and outputs the sports evaluation and sports suggestion corresponding to the athlete ID.

[0110] See also Figure 3 Another aspect of the present application provides a method for generating sports event results based on artificial intelligence and visual technology, including:

[0111] S0: Acquire athlete data, sports event data and environmental data; sports event data includes event ID and its corresponding event detection video, event standard specification and event type; athlete data includes athlete ID and its corresponding physical condition;

[0112] S1: Generate a project environmental impact coefficient based on the project ID and environmental data; generate an alarm signal based on the project environmental impact coefficient;

[0113] S2: after identifying the athlete ID according to the event detection video, generate a sports video; generate a sports result according to the sports video corresponding to the event type and the athlete ID;

[0114] S3: Prompt according to the alarm signal and contact the management.

[0115] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0116] The working principle of the present application is as follows: by generating a project environmental impact coefficient according to the project ID and environmental data; generating an alarm signal according to the project environmental impact coefficient; generating a sports video after identifying the athlete ID according to the project detection video; generating a sports result according to the sports video corresponding to the project type and the athlete ID, monitoring the environmental status in the sports area in real time, and adaptively judging whether the sports project can be carried out normally according to the different requirements of different sports for the environment, and providing accurate data support for the progress of the sports project; at the same time, according to the different scoring standards of the sports projects, constructing a multi-branch network to calculate the score of each athlete in different projects in real time, and giving them a sports evaluation and providing personalized suggestions according to their own status, thereby improving the comprehensiveness of the generation of sports project results, and avoiding the problem that the prior art ignores the conditions for the smooth progress of the sports project, the differences in the scoring standards of different projects, and the failure to provide timely feedback and suggestions to athletes, resulting in low accuracy and comprehensiveness of the result generation.

[0117] The above embodiments are only used to illustrate the technical method of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical method of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present application.

Claims

1. A sports event result generation system based on artificial intelligence and visual technology, characterized in that: include: Data collection module, data analysis module, early warning module and database; The data acquisition module: acquires athlete data, sports event data and environmental data through data acquisition equipment; the sports event data includes the event ID and its corresponding event detection video and event type; the athlete data includes the athlete ID and its corresponding physical condition; The data analysis module generates a project environmental impact coefficient according to the project ID and environmental data; generates an alarm signal according to the project environmental impact coefficient; generates a sports video after identifying the athlete ID according to the project detection video; generates a sports result according to the sports video corresponding to the project type and the athlete ID; The early warning module: makes prompts according to the alarm signal and contacts the management personnel.

2. The sports event result generation system based on artificial intelligence and visual technology according to claim 1, characterized in that: The generating of the project environmental impact coefficient according to the project ID and environmental data includes: Obtaining a project ID and environmental data; the environmental data includes several environmental parameters; By formula Calculate the project environmental impact coefficient XHYX i ; Among them, α ij is the jth weight coefficient corresponding to the i-th project ID; the weight coefficient is generated according to the project ID; HC j Expressed as several environmental parameters; DHC j Expressed as unit environmental parameters corresponding to several environmental parameters; ZHC ij It is expressed as the optimal environmental parameter value corresponding to several environmental parameters of each sport.

3. The sports event result generation system based on artificial intelligence and visual technology according to claim 2, characterized in that: The weight coefficient is generated according to the project ID, including: Get the project ID and its corresponding environmental impact preference HYP ij ; The environmental impact preference is assessed by experts based on the impact of the project ID on several environmental parameters; By formula Calculate the weight coefficient α ij .

4. The sports event result generation system based on artificial intelligence and visual technology according to claim 1, characterized in that: The generating of an alarm signal according to the project environmental impact coefficient comprises: Obtain environmental impact coefficients for several projects; Determine whether the project environmental impact coefficient is greater than its corresponding project environmental threshold; Yes, generate a sports event stop alarm signal; If no, determine whether the project environmental impact coefficient is greater than its corresponding D times project environmental threshold; if yes, generate a warning signal for a poor sports project environment; if no, do nothing; wherein D is the proportional coefficient, D∈(0,1).

5. The sports event result generation system based on artificial intelligence and visual technology according to claim 1, characterized in that: The step of identifying the athlete ID according to the event detection video and generating a sports video therefrom includes: Get project inspection video; Inputting the event detection video into a target detection model to obtain the athlete ID; the target detection model is constructed by an artificial intelligence model; Use target tracking algorithm to track the detected athlete ID and record its movement trajectory; The project detection video is processed into frames to obtain several frames of detection images; Obtain several frames of detection images corresponding to each athlete ID, and splice them in chronological order to obtain the sports video corresponding to the athlete ID.

6. The sports event result generation system based on artificial intelligence and visual technology according to claim 5, characterized in that: The target detection model is constructed through an artificial intelligence model, including: Obtain several historical event detection videos and their corresponding athlete IDs; Divide several historical event detection videos and their corresponding several athlete IDs into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain a training set, a verification set and a test set; Choose artificial intelligence as the base model; Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model; By verifying the pre-trained model on the test set, we finally get a target detection model that inputs the event detection video and outputs the athlete ID.

7. The sports event result generation system based on artificial intelligence and visual technology according to claim 1, characterized in that: The generating of the sports result according to the sports video corresponding to the event type and the athlete ID includes: Obtain sports videos and physical conditions corresponding to the project type and athlete ID; the project type includes result type, process type and dual type; Inputting the athlete ID, sports video and event type into a score prediction model to obtain a final score; the score prediction model is obtained by constructing a multi-task branch; The athlete ID, physical condition and sports video are input into the evaluation suggestion generation model to obtain sports evaluation and sports suggestions; the evaluation suggestion generation model is constructed through a deep learning model.

8. The sports event result generation system based on artificial intelligence and visual technology according to claim 7, characterized in that: The rating estimation model is obtained by constructing a multi-task branch, including: Get several historical athlete IDs, historical sports videos and corresponding event types and historical scores; Divide several historical athlete IDs, historical sports videos, corresponding event types and historical scores into several subtasks according to event types; divide the data in several subtasks into several training data, verification data and test data; perform data preprocessing on several training data, verification data and test data to obtain several training sets, verification sets and test sets; Select the multi-task learning model as the base model; Train the basic model through several training sets, and adjust the learning rate and hyperparameters on several validation sets to obtain the pre-trained model; By verifying the pre-trained model on several test sets, we finally obtain a score estimation model that inputs motion videos and outputs the final score.

9. The sports event result generation system based on artificial intelligence and visual technology according to claim 7, characterized in that: The evaluation suggestion generation model is constructed through a deep learning model, including: Obtain several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions; Divide several historical athlete IDs, physical conditions, sports videos and their corresponding sports evaluations and sports suggestions into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain a training set, a verification set and a test set; Select a deep learning model as the base model; Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model; By verifying the pre-trained model on the test set, we finally obtained an evaluation and suggestion generation model that inputs athlete ID, physical condition and sports video, and outputs the sports evaluation and sports suggestion corresponding to the athlete ID.

10. A method for generating sports event results based on artificial intelligence and visual technology, applied to a system for generating sports event results based on artificial intelligence and visual technology as claimed in any one of claims 1 to 9, characterized in that: include: S0: Acquire athlete data, sports event data and environmental data; the sports event data includes the event ID and its corresponding event detection video, event standard specification and event type; the athlete data includes the athlete ID and its corresponding physical condition; S1: Generate a project environmental impact coefficient based on the project ID and environmental data; generate an alarm signal based on the project environmental impact coefficient; S2: after identifying the athlete ID according to the event detection video, generate a sports video; generate a sports result according to the sports video corresponding to the event type and the athlete ID; S3: Prompt according to the alarm signal and contact the management.

Images