Intelligent skiing action evaluation method, system, device and medium

Through the intelligent skiing action evaluation method, real-time monitoring and evaluation of skier movements has been solved, and the problem that traditional skiing teaching cannot be fully evaluated is provided. Detailed skiing action and equipment suggestions are provided, which improves skiing safety and teaching effectiveness.

CN117275092BActive Publication Date: 2024-11-12AOXUE CULTURE COMM BEIJING CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202311301207.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-10-09
Publication Date
2024-11-12
Estimated Expiration
2043-10-09

AI Technical Summary

Technical Problem

Traditional skiing teaching cannot comprehensively evaluate skiers' movement details and space for improvement. The coaching guidance is limited and cannot correct skiing movements in real time, making it difficult to provide full-time skiing exercises.

Method used

The intelligent skiing action evaluation method is adopted to provide detailed evaluation and equipment suggestions by obtaining skier images, intercepting ski videos, tracking and extracting skiing action images, identifying key actions, generating skiing evaluation reports, monitoring skiing actions in real time.

Benefits of technology

It has achieved real-time assessment of the skier's movements and a comprehensive assessment of improved space, improved ski safety and reasonable wear suggestions for ski equipment, and enhanced the effect of ski teaching.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117275092B_ABST
    Figure CN117275092B_ABST
Patent Text Reader

Abstract

A method, system, device and medium for intelligent skiing action evaluation, relating to the field of intelligent skiing. The method comprises the following steps: obtaining a skier image uploaded by a skier; intercepting a skiing video containing the skier in a snow track monitoring video according to the skier image; tracking and selecting the skier in the skiing video according to the skier image, and extracting the skier's skiing action image; determining at least one key action image in the skiing action image; evaluating the key action image, and generating a skiing evaluation report according to the evaluation result. By adopting the technical solution provided by the present application, a reasonable skiing evaluation report is given by monitoring the skier throughout the skiing process, which is conducive to a comprehensive evaluation of the skier's action details and improvement space.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent skiing, and in particular to an intelligent skiing action evaluation method, system, device and medium. Background Art

[0002] Skiing is a winter outdoor activity that has attracted a large number of people to start skiing, and skiing is currently becoming more and more popular.

[0003] When skiing, correct skiing movements are the basis for ensuring skiing safety. When a new skier just starts skiing, he usually needs to learn skiing movements. At present, traditional skiing teaching relies on the direct guidance of the coach, but this guidance is often limited to limited observation time and experience. The coach cannot always guide and correct the skier's skiing movements on the ski resort, and it is difficult to make improvement suggestions for the skier's skiing movements throughout the skiing process. Summary of the invention

[0004] In order to solve the problem that the current traditional skiing teaching cannot comprehensively evaluate the details of the skier's movements and the room for improvement, the present application provides an intelligent skiing movement evaluation method, system, device and medium.

[0005] In a first aspect, the present application provides an intelligent skiing action evaluation method, the method comprising the following steps:

[0006] Get skier images uploaded by skiers;

[0007] intercepting a skiing video including the skier from the ski trail monitoring video according to the skier image;

[0008] Tracking and selecting the skier in the skiing video according to the skier image, and extracting the skiing action image of the skier;

[0009] determining at least one key action image in the skiing action image;

[0010] The key action images are evaluated, and a skiing evaluation report is generated according to the evaluation results.

[0011] By adopting the above technical solution, the skier can be tracked in real time in the snow track monitoring video during the skiing process, so as to determine the skier's skiing action image, analyze and evaluate the skiing action image, and finally evaluate the skier's skiing action. By monitoring the skier throughout the skiing process, a reasonable skiing evaluation report is given, which is conducive to a comprehensive evaluation of the skier's action details and improvement space.

[0012] Optionally, capturing a skiing video containing a skier from a ski trail monitoring video according to the skier image specifically includes:

[0013] extracting skier features based on the skier image;

[0014] Identify all human objects contained in the snow track monitoring video, and extract human features of each identified human object respectively;

[0015] respectively calculating the feature similarities between the skier feature and each of the human body features;

[0016] A video segment of the snow track monitoring video containing a similar human object is intercepted as the skiing video, and the feature similarity between the human features of the similar human object and the features of the skier is greater than a set threshold.

[0017] By adopting the above technical solution, the skiing video of a specific person is obtained, which avoids the influence of the skiing images of other skiers in the ski slope monitoring video on the skier's action evaluation, thereby ensuring the accuracy of the skiing evaluation report.

[0018] Optionally, tracking and selecting the skier in the skiing video according to the skier image to extract the skiing action image of the skier specifically includes:

[0019] Inputting the skier features and the skiing video into a preset tracking model to locate the skier in the skiing video;

[0020] Generate a positioning frame around the skier in real time based on the positioning results;

[0021] An image selected by the positioning frame in each video frame of the skiing video is captured as the skiing action image.

[0022] By adopting the above technical solution, only the skiing action images containing the specific skier are captured, avoiding the influence of other factors such as background on the evaluation of the skiing action, and further ensuring the accuracy of the skiing evaluation report.

[0023] Optionally, the tracking model is a Siamese-RPN target tracking model.

[0024] By adopting the above technical solution, Siamese-RPN (Siamese Region Proposal Network) is a model for target tracking, which combines the methods of twin networks and region proposal networks. The Siamese-RPN target tracking model is used to track specific skiers, ensuring the speed and accuracy of target tracking.

[0025] Optionally, determining at least one key action image in the skiing action image specifically includes:

[0026] Inputting the skiing action image into a preset human body key point extraction model to identify the human body key points of the skier in the skiing action image;

[0027] Constructing a skeleton spatiotemporal graph according to the identified key points of the human body;

[0028] Analyzing the skeleton spatiotemporal graph through a preset action classification model to identify key actions;

[0029] The key action image is determined according to the identified key action.

[0030] Optionally, the key action images are evaluated and a skiing evaluation report is generated according to the evaluation results, specifically including:

[0031] The skiing evaluation report includes skiing action evaluation information, and the key action image is compared with a preset standard key action image;

[0032] The skiing action evaluation information is output according to the comparison result.

[0033] Optionally, the skiing evaluation report also includes skiing equipment evaluation information to obtain the skier's historical skiing data;

[0034] evaluating the skiing experience of the skier according to the historical skiing data to determine the skiing level of the skier;

[0035] Determining matching ski equipment information of the skier at the ski level according to the ski level;

[0036] identifying the skier's current ski equipment information according to the skier image;

[0037] The matching ski equipment information is compared with the current ski equipment information, and the ski equipment evaluation information is determined according to the comparison result.

[0038] By adopting the above technical solution, the ski equipment currently worn by the skier is evaluated according to the skiing experience of the skier, and reasonable equipment wearing suggestions are provided to the skier, thereby further ensuring the safety of the skier when skiing.

[0039] In a second aspect of the present application, an intelligent skiing action evaluation system is provided, the system comprising the following modules:

[0040] A skier image acquisition module, used to acquire skier images uploaded by skiers;

[0041] A skiing video acquisition module, used for capturing a skiing video containing a skier in a ski trail monitoring video according to the skier image;

[0042] A skiing action image acquisition module, used to track and select the skier in the skiing video according to the skier image, and extract the skiing action image of the skier;

[0043] A key action image acquisition module, used to determine at least one key action image in the skiing action image;

[0044] The key action image evaluation module is used to evaluate the key action image and generate a skiing evaluation report according to the evaluation result.

[0045] In a third aspect of the present application, an electronic device is provided;

[0046] The electronic device includes a processor, a memory, a user interface and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs an intelligent skiing action evaluation method.

[0047] In a fourth aspect of the present application, a computer-readable storage medium is provided;

[0048] The computer-readable storage medium stores instructions, and when the instructions are executed, an intelligent skiing maneuver evaluation method is performed.

[0049] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0050] 1. Track the skier in real time in the snow track monitoring video while the skier is skiing, so as to determine the skier's skiing action image, analyze and evaluate the skiing action image, and finally evaluate the skier's skiing action. By monitoring the skier throughout the skiing process, a reasonable skiing evaluation report is given, which is conducive to a comprehensive evaluation of the skier's action details and improvement space.

[0051] 2. Evaluate the ski equipment currently worn by the skier based on the skiing experience of the skier, and provide the skier with reasonable equipment wearing suggestions, so as to further ensure the safety of the skier when skiing. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of an implementation scenario of an intelligent skiing action evaluation method provided in an embodiment of the present application.

[0053] Figure 2It is a flow chart of an intelligent skiing action evaluation method provided in an embodiment of the present application.

[0054] Figure 3 It is a structural schematic diagram of an intelligent skiing action evaluation system disclosed in an embodiment of the present application.

[0055] Figure 4 It is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present application.

[0056] Explanation of the reference numerals: 301, skier image acquisition module; 302, skiing video acquisition module; 303, skiing action image acquisition module; 304, key action image acquisition module; 305, key action image evaluation module; 400, electronic device; 401, processor; 402, communication bus; 403, user interface; 404, network interface; 405, memory. DETAILED DESCRIPTION

[0057] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0058] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0059] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0060] Reference Figure 1The present application provides an intelligent skiing action evaluation method for use in a skiing sports scenario. A plurality of ski field monitoring devices are arranged on a ski slope. By setting the shooting angle of each ski field monitoring device, the motion video of all skiers skiing on the ski slope can be obtained, which is the ski slope monitoring video.

[0061] When a skier starts skiing, he can enter the start instruction on the smart terminal he carries with him. In response to the start instruction, the smart terminal will access the back-end server of the ski monitoring system set up in the ski resort and obtain relevant data from the back-end server. The ski slope monitoring video obtained by the ski slope monitoring equipment will also be transmitted to the back-end server for processing.

[0062] In a feasible embodiment of the present application, a front-end application for front-end interaction is provided to skiers on a smart terminal. The front-end application can provide relevant services to skiers in the form of a smart skiing APP. These services include the skiing action evaluation service and skiing equipment evaluation service mentioned in subsequent embodiments. At the same time, skiers can also be provided with GPS skiing positioning services, skiing community services, rescue services and other smart services related to skiing.

[0063] Reference Figure 2 The present application provides an intelligent skiing action evaluation method, which specifically includes the following steps:

[0064] S1: Obtain skier images uploaded by skiers;

[0065] Specifically, before the skier starts skiing, the skier's image is photographed through the smart terminal, and the photographed skier image is uploaded to the back-end server, and the back-end server acquires the uploaded skier image.

[0066] Optionally, the skier may also take a photo of the skier image using a skier image taking device provided at the ski resort.

[0067] The skier image specifically includes the skier's front image and the skier's back image. When the skier takes the skier image by himself, the validity of the skier image needs to be verified to ensure that the skier image contains the skier's full-body image, thereby providing effective data support for subsequent steps.

[0068] S2: intercepting a skiing video containing a skier from the ski trail monitoring video according to the skier image;

[0069] Specifically, there are multiple monitoring videos of the ski slope, and each monitoring video of the ski slope records the skiing videos of all skiers on the ski slope. Therefore, it is necessary to select the skiing video of the skier that contains the currently uploaded skier image from all the monitoring videos of the ski slope.

[0070] Extract skier features based on the acquired skier image. The skier features will describe and characterize a specific person. Each specific skier has a unique skier feature corresponding to it. Identify all human objects contained in the ski trail monitoring video, and extract human features of each human object. The identification of human objects can be achieved through the ROI region of interest algorithm. In a feasible embodiment of the present application, the skier features are image features corresponding to the skier image, and the human features are image features corresponding to the human objects.

[0071] The feature similarities between the skier's features and the features of each human body are calculated respectively, so as to determine the similarity between the skier and a large number of human objects in the ski slope monitoring video. The feature similarity is determined by the cosine vector similarity.

[0072] A video clip of the snow track surveillance video containing a similar human object is intercepted as a skiing video, and the feature similarity between the human features of the similar human object and the features of the skier is greater than a set threshold. In a feasible embodiment of the present application, the set threshold is 90%.

[0073] S3: Tracking and selecting the skier in the skiing video according to the skier image, and extracting the skiing action image of the skier;

[0074] Specifically, the skiing video includes a skier corresponding to the skier image and other skiers. Therefore, it is necessary to track the skier corresponding to the skier image and capture its corresponding skiing action image. The skiing action image only includes the skier corresponding to the skier image.

[0075] The skier features extracted from the skier image and the acquired skiing video are input into a preset tracking model to track and locate the skier corresponding to the skier image in the skiing video. In a preferred embodiment of the present application, the tracking model uses a Siamese-RPN target tracking model.

[0076] The Siamese-RPN (Siamese Region Proposal Network) target tracking model is a model for target tracking. It combines the methods of the twin network and the region proposal network to generate candidate boxes near the tracked object, thereby tracking the tracked target in the input video sequence. The specific target tracking process of the target tracking model is as follows:

[0077] Step 1: Select the target object in the initial frame as a template and extract the feature representation of the target object. This can be done by using a pre-trained convolutional neural network (e.g., a model based on ImageNet) to extract features, or by using a custom network structure.

[0078] Step 2: Use the skier image as the search image and obtain the skier features corresponding to the skier image.

[0079] Step 3: Use a similarity calculation method (e.g., cosine similarity or Euclidean distance) to compare the similarity between the template features and the search features. Typically, two branches of the Siamese network can be used to extract feature representations of the template and search image and calculate the similarity score between them.

[0080] Step 4: Based on the similarity score, use the region proposal network (RPN) to generate candidate boxes or candidate regions. RPN provides a series of candidate boxes that may contain the target based on the similarity score and location information.

[0081] Step 5: Perform target classification and position regression on the generated candidate box. The target classifier is used to determine whether the candidate box contains the target object, and the position regressor is used to accurately locate the position of the target. These classification and regression modules are usually based on deep learning network structures, such as convolutional neural networks (CNNs).

[0082] Step 6: According to the results of classification and regression, determine the location and bounding box of the target. Input the obtained location information as a new template into the twin network, and continue to process the template image and the search image and calculate the similarity. In this way, the target can be continuously tracked and updated, so that the target can be stably tracked in the video sequence.

[0083] After being processed by the target tracking model, multiple positioning boxes are added to the skiing video, and the object selected by each positioning box is the skier corresponding to the skier image. The skiing video is split into video frames, and the area selected by the positioning box in each video frame is intercepted to obtain the skiing action image.

[0084] S4: determining at least one key action image in the skiing action image;

[0085] Specifically, the captured multiple skiing action images are input into a preset human key point extraction model to extract the human key points of the skier in the skiing action images. In a preferred embodiment preset in the present application, the human key points are recognized by OpenPose.

[0086] OpenPose is an open source real-time system developed by Carnegie Mellon University in the United States based on convolutional neural networks and supervised learning algorithms. It is used to detect key points of the human body (face, limbs, etc.) in 2D images. It uses bottom-up human key point information recognition, that is, finding various parts of the human body and then connecting them through partial affinity domains.

[0087] A skeleton space-time graph of the skier is constructed based on the identified key points of the human body. Specifically, the skeleton space-time graph abstracts the skier's skiing movements, simplifies the skier into an abstract body structure of torso and limbs, and focuses on describing the skier's skiing movements.

[0088] The skeleton space-time graph is analyzed by a preset action classification model to identify key actions. Specifically, the action classification model is a multi-classifier. For the input skeleton space-time graph, the action classification model will automatically identify the skeleton space-time graph and classify the skeleton space-time graph into preset action categories. In a feasible embodiment of the present application, the action categories specifically include flat ground sliding, parallel straight descent, plow straight descent, plow braking, half-plow turn, plow turn (deep arc), parallel turn (middle turn), change of direction horizontal descent, small bend sliding, large bend carving, mushroom snow track sliding and meaningless actions, a total of 12 types, and the action classification model can adopt the STGCN model.

[0089] The skiing action images are divided into multiple action categories through the action classification of the action classification model, and the key action images are selected from the skiing action images according to the preset key action selection criteria. It should be noted that the key action selection criteria can be set by the skier. For example, if the skier currently wants to learn and correct the two actions of point-and-stick small turn sliding and large turn carving, then the key action selection criteria can be set to point-and-stick small turn sliding and large turn carving.

[0090] S5: Evaluate the key action images and generate a skiing evaluation report based on the evaluation results.

[0091] Specifically, a standard key action image corresponding to the key action image is obtained, the key action image is compared with the standard key action image, and skiing action evaluation information of the current skier is obtained according to the comparison result.

[0092] When performing the comparison, an image similarity comparison method may be used, but preferably, a skeleton space-time graph corresponding to the key action image and a skeleton space-time graph corresponding to the standard key action image may be obtained, and the skeleton space-time graphs may be compared to further improve the accuracy of the action assessment.

[0093] In addition, the skiing evaluation report also includes skiing equipment evaluation information, which is descriptive information on whether the skier's current equipment is suitable and correct.

[0094] The skier's historical skiing data is obtained, which specifically includes skiing time, skiing action standard values, and skiing mileage. The skier's skiing level is evaluated based on the historical skiing data, and then the matching ski equipment information corresponding to the skier's current skiing level is obtained. It can be understood that the matching ski equipment information is the most suitable ski equipment for the skier based on the skier's skiing experience. The specific matching method between the matching ski equipment information and the skiing level can be obtained through expert experience or based on big data analysis.

[0095] The skier's current equipment information is identified based on the skier image obtained in the aforementioned step, the current ski equipment information is compared with the matching ski equipment information, and ski equipment evaluation information is obtained based on the comparison result, thereby providing reference opinions for the skier's ski equipment selection.

[0096] Reference Figure 3 The present application also provides an intelligent skiing action evaluation system, which specifically includes the following modules:

[0097] A skier image acquisition module 301 is used to acquire a skier image uploaded by a skier;

[0098] A skiing video acquisition module 302 is used to capture a skiing video containing a skier in a ski trail monitoring video according to the skier image;

[0099] A skiing action image acquisition module 303 is used to track and select the skier in the skiing video according to the skier image, and extract the skiing action image of the skier;

[0100] A key action image acquisition module 304, used to determine at least one key action image in the skiing action image;

[0101] The key action image evaluation module 305 is used to evaluate the key action images and generate a skiing evaluation report according to the evaluation results.

[0102] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0103] The present application also discloses an electronic device 400. Figure 4 , Figure 44 is a schematic diagram of the structure of an electronic device 400 disclosed in an embodiment of the present application. The electronic device 400 may include: at least one processor 401 , at least one network interface 404 , a user interface 403 , a memory 405 , and at least one communication bus 402 .

[0104] The communication bus 402 is used to realize the connection and communication between these components.

[0105] The user interface 403 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 403 may also include a standard wired interface and a wireless interface.

[0106] The network interface 404 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0107] Among them, the processor 401 may include one or more processing cores. The processor 401 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 405, and calling data stored in the memory 405. Optionally, the processor 401 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 401 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 401, and it can be implemented separately through a chip.

[0108] Among them, the memory 405 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 405 includes a non-transitory computer-readable storage medium. The memory 405 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 405 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 405 may optionally also be at least one storage device located away from the aforementioned processor 401. Refer to Figure 4 , the memory 405 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of an intelligent skiing action evaluation method.

[0109] exist Figure 4 In the electronic device 400 shown, the user interface 403 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 401 can be used to call the application program storing an intelligent skiing action evaluation method in the memory 405, and when executed by one or more processors 401, the electronic device 400 executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, some steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0110] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0112] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0114] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory 405. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory 405 and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned memory 405 includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk.

[0115] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0116] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. An intelligent skiing action evaluation method, characterized in that: The method comprises the following steps: Get skier images uploaded by skiers; intercepting a skiing video containing the skier from the ski trail monitoring video according to the skier image; Inputting skier features and the skiing video into a preset tracking model, wherein the skier features are image features corresponding to the skier image, and the tracking model is a Siamese-RPN target tracking model, to locate the skier in the skiing video; The target object in the initial frame is selected as a template, and the feature representation of the target object is extracted; the skier image is used as a search image, and the skier features corresponding to the skier image are obtained; the similarity between the template features and the search features is compared using a similarity calculation method, and a similarity score is calculated, and based on the similarity score, a candidate box or a candidate region is generated using a region proposal network; a series of candidate boxes containing the target are provided according to the similarity score and the position information; target classification and position regression are performed on the generated candidate boxes; the position and positioning box of the target are determined according to the classification and regression results; intercepting an image selected by the positioning frame in each video frame of the skiing video as the skiing action image; determining at least one key action image in the skiing action image; evaluating the key action images and generating a skiing evaluation report according to the evaluation results; The skiing evaluation report includes skiing equipment evaluation information and obtains the skier's historical skiing data; evaluating the skiing experience of the skier according to the historical skiing data to determine the skiing level of the skier; Determining matching ski equipment information of the skier at the ski level according to the ski level; identifying the skier's current ski equipment information according to the skier image; The matching ski equipment information is compared with the current ski equipment information, and the ski equipment evaluation information is determined according to the comparison result.

2. The intelligent skiing action evaluation method according to claim 1, characterized in that: In intercepting a skiing video containing a skier from a ski trail monitoring video according to the skier image, the method specifically includes: extracting skier features based on the skier image; Identify all human objects contained in the snow track monitoring video, and extract human features of each identified human object respectively; respectively calculating the feature similarities between the skier feature and each of the human features; A video segment of the snow track monitoring video containing a similar human object is intercepted as the skiing video, and the feature similarity between the human features of the similar human object and the features of the skier is greater than a set threshold.

3. The intelligent skiing action evaluation method according to claim 1, characterized in that: Determining at least one key action image in the skiing action image specifically includes: Inputting the skiing action image into a preset human body key point extraction model to identify the human body key points of the skier in the skiing action image; Constructing a skeleton spatiotemporal graph according to the identified key points of the human body; Analyzing the skeleton spatiotemporal graph through a preset action classification model to identify key actions; The key action image is determined according to the identified key action.

4. The intelligent skiing action evaluation method according to claim 1, characterized in that: The key action images are evaluated and a skiing evaluation report is generated according to the evaluation results, specifically including: The skiing evaluation report includes skiing action evaluation information, and the key action image is compared with a preset standard key action image; The skiing action evaluation information is output according to the comparison result.

5. An intelligent skiing action evaluation system, characterized in that: The system comprises: A skier image acquisition module (301), used to acquire a skier image uploaded by a skier; A skiing video acquisition module (302), used for capturing a skiing video containing a skier in a ski trail monitoring video according to the skier image; The skiing action image acquisition module (303) is used to input the skier feature and the skiing video into a preset tracking model, wherein the skier feature is the image feature corresponding to the skier image, and the tracking model is a Siamese-RPN target tracking model, and locate the skier in the skiing video; select the target object in the initial frame as a template and extract the feature representation of the target object; use the skier image as a search image and obtain the skier feature corresponding to the skier image; use a similarity calculation method to compare the similarity between the template feature and the search feature, and calculate the similarity score, and use a region proposal network to generate a candidate frame or a candidate region based on the similarity score; provide a series of candidate frames containing the target according to the similarity score and position information; perform target classification and position regression on the generated candidate frames; determine the position and positioning frame of the target according to the classification and regression results, and intercept the image selected by the positioning frame in each video frame of the skiing video as the skiing action image; A key action image acquisition module (304), used to determine at least one key action image in the skiing action image; A key action image evaluation module (305) is used to evaluate the key action image and generate a skiing evaluation report based on the evaluation result; the skiing evaluation report includes skiing equipment evaluation information and obtains the skier's historical skiing data; the skiing experience of the skier is evaluated based on the historical skiing data to determine the skier's skiing level; the matching skiing equipment information of the skier at the skiing level is determined based on the skiing level; the current skiing equipment information of the skier is identified based on the skier image; the matching skiing equipment information is compared with the current skiing equipment information, and the skiing equipment evaluation information is determined based on the comparison result.

6. An electronic device, characterized in that: The electronic device (400) comprises a processor (401), a memory (405), a user interface (403) and a network interface (404), wherein the memory (405) is used to store instructions, the user interface (403) and the network interface (404) are used to communicate with other devices, and the processor (401) is used to execute the instructions stored in the memory (405) so that the electronic device (400) executes the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method steps according to any one of claims 1 to 4 are performed.

Citation Information

Patent Citations

  • EVALUATING POSE-BASED movement

    CN112819852A

  • Skier three-dimensional attitude estimation method based on self-supervision technology

    CN114611600A

  • Snow motion information monitoring method and related device

    CN115346273A

  • Motion scoring method, terminal and computer readable storage medium

    CN115439925A

  • Learning Management System for Task-Based Objectives

    US20190311649A1