A scoring method, system, device and storage medium for a hitting type competitive game

Abstract

The present application belongs to a kind of scoring method, for the existing competitive sports game scoring method, there is easy to misjudge, wear inconvenient, equipment is easy to be damaged, and different events cannot use unified scoring system technical problem, provide a kind of hitting type competitive game scoring method, system, equipment and storage medium, collect audio data and video data on the playing field, whether there is possible to occur effective hitting is initially determined according to audio data, then according to video data determine effective hitting information including the competition party that occurs effective hitting, the action of effective hitting adopted and the part in effective hitting, through multidirectional determination, exclude the error that human judgment and scoring can introduce, improve the justice and fairness of competition, can also reduce economic and manpower cost.In addition, for different competition needs, can carry out scoring, can be applied to different competition rules, scoring is simple, need not purchase expensive equipment and equipment maintenance is convenient.

Description

Technical Field

[0001] This invention pertains to a scoring method, specifically relating to a scoring method, system, device, and storage medium for striking-type competitive sports. Background Technology

[0002] Currently, in many individual or multi-person competitive sports events, scoring is often done by live referees who judge the movements visually, or by reviewing real-time video footage from surveillance cameras and then manually scoring. For example, the scoring system of the Chinese Wushu Association's Bingdao Championship uses live referees (one head referee, three assistant referees, and one supervising referee) to judge the movements visually and record the results using an electronic scorer. The final result is then determined by comparing the scores. This scoring method consumes a significant amount of manpower. Furthermore, because observing real-time movements with the naked eye is difficult, and it's challenging to obtain highly accurate results regarding speed, position, whether a strike was made, and whether a foul occurred, misjudgments are prone to occur.

[0003] Some competitive sports, such as karate and taekwondo, use wearable scoring devices. These devices detect impacts and force, and scores are determined based on the interaction between different parts of the body. However, this scoring method is inconvenient to wear, uses extremely expensive materials, and is prone to damage and wear and tear, resulting in inaccurate scoring and potentially affecting the fairness and impartiality of official competitions. These scoring devices and methods are based on different competition rules and require different sensor devices for different competitions, making it impossible to use a unified scoring system across different events. Summary of the Invention

[0004] This invention addresses the technical problems of existing scoring methods for competitive sports, such as susceptibility to misjudgment, inconvenience of wearing protective gear, easy damage to equipment, and the inability to use a unified scoring system for different events. It provides a scoring method, system, device, and storage medium for striking-type competitive sports.

[0005] To achieve the above objectives, the present invention employs the following technical solution:

[0006] In a first aspect, the present invention proposes a scoring method for striking-type competitive sports, comprising the following steps:

[0007] S1 collects audio and video data from the competition venue.

[0008] S2, filter out ambient noise from the audio data to obtain clean audio data;

[0009] S3. Based on the clean audio data, determine whether a valid strike is possible. If so, proceed to step S4. Otherwise, continue collecting audio data until a valid strike is possible.

[0010] S4, extract video data within a preset time before the moment when a valid impact may occur, and record it as valid video data;

[0011] S5, perform spatial and temporal modeling of the actions of both sides in the effective video data, and confirm the effective hitting information based on the modeling results. The effective hitting information includes the side that made the effective hit, the action used in the effective hit, and the part of the body hit by the effective hit.

[0012] S6. Scoring is performed based on the valid strike information.

[0013] Further, step S2 specifically involves using a filtering algorithm to filter out ambient noise from the audio data;

[0014] In step S3, determining whether a valid strike could have occurred based on the audio data specifically involves using a sound analysis function to determine whether a valid strike has occurred.

[0015] Further, in step S1, the audio data on the competition field is specifically obtained by microphones evenly distributed around the competition field and / or on the top of the competition field; the video data on the competition field is specifically obtained by cameras evenly distributed around the competition field and / or on the top of the competition field.

[0016] Step S2 is as follows: If more than half of the microphones collect audio data within a preset time, and it is determined that a valid hit may have occurred, then a valid hit may have occurred; otherwise, it is considered that no valid hit has occurred.

[0017] Further, step S5 specifically involves performing the following steps on the valid video data from each camera:

[0018] The obscured parts of the contestants and weapons in the valid video data are processed using a completion prediction algorithm to obtain complete and valid video data.

[0019] The actions of both sides in the complete and valid video data are modeled in a spatial and temporal manner. The state of the two sides and their weapons is identified through a segmentation algorithm, and then the valid hitting information is obtained.

[0020] Furthermore, step S5 specifically includes:

[0021] Based on the location information, the effective video data from each camera are stitched together into a data matrix;

[0022] Using a 3D reconstruction algorithm, the data corresponding to each time step of all data matrices are transformed into 3D reconstruction models;

[0023] By concatenating all three-dimensional reconstruction models on a time scale, a four-dimensional reconstruction model is obtained.

[0024] Based on the four-dimensional reconstruction model, effective impact information is obtained.

[0025] Furthermore, obtaining the effective strike information specifically refers to:

[0026] The action used in the effective strike is identified by a trained classification network model.

[0027] The participating side that made a valid strike is determined through the following method:

[0028] The weapon and the two players are modeled separately to obtain weapon models and player models. The collision detection results between the weapon model and the two player models are used to determine whether a collision has occurred. Then, the player who made a valid hit is determined by the information of the weapon model and the two player models.

[0029] The effective point of impact is obtained through the following method:

[0030] The weapon and the two players are modeled separately to obtain weapon models and player models. The distance between the weapon model and the two player models is calculated. Combined with the standard human body model, the effective hit points are determined.

[0031] Furthermore, step S6 includes: manually correcting the effective strike information.

[0032] Secondly, this invention proposes a scoring system for striking competitions, including a data acquisition module, a filtering module, an audio recognition module, an extraction module, a video recognition module, and a scoring module;

[0033] The acquisition module is used to acquire audio data and video data from the competition venue.

[0034] The filtering module is used to filter out ambient noise from the audio data to obtain clean audio data;

[0035] The audio recognition module is used to determine whether a valid strike may occur based on the audio data. If so, the extraction module is activated; otherwise, audio data is continuously collected until a valid strike may occur.

[0036] The extraction module is used to extract video data within a preset time period before the moment when an effective impact may occur, and record it as effective video data.

[0037] The video recognition module is used to perform spatial and temporal modeling of the actions of both sides in the valid video data, and to confirm valid hitting information based on the modeling results. The valid hitting information includes the side that made the valid hit, the action used in the valid hit, and the part of the body hit by the valid hit.

[0038] The scoring module is used to score points based on the valid strike information.

[0039] Thirdly, the present invention provides a computer device including a memory and one or more processors, wherein the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, which, when executed by the processor, cause the server to perform the above-described method.

[0040] Fourthly, the present invention provides a computer storage medium comprising computer instructions that, when executed on an electronic device, cause the electronic device to perform the above-described method.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This invention proposes a scoring method for striking competitions. It collects audio and video data from the competition venue. Based on the audio data, it initially determines whether a valid strike could have occurred. Then, based on the video data, it determines the valid strike information, including the party that made the valid strike, the action used, and the body part hit. This multi-faceted assessment eliminates errors that may be introduced by human judgment and scoring, while improving the fairness and impartiality of the competition and reducing economic and labor costs. Furthermore, it can be used to score for different competition needs and can be applied to different competition rules. The scoring method is simple, does not require expensive equipment, and the equipment is easy to maintain. The scoring method of this invention is highly accurate, fast-responding, and highly maintainable.

[0043] 2. In this invention, ambient noise is filtered out from the audio data, which can improve the accuracy of scoring.

[0044] 3. This invention can collect data using a single camera and microphone or multiple cameras and microphones, making it flexible and convenient, with lower requirements for the competition venue, and more suitable for widespread application.

[0045] 4. This invention uses modeling and classification network models for scoring, which can further improve scoring efficiency and accuracy.

[0046] 5. The present invention also proposes a scoring system, computer equipment and computer storage medium for striking competitions, which implements the above-mentioned scoring method in hardware, has the advantages of the above-mentioned scoring method, and can be applied and promoted. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating Example 1;

[0049] Figure 2 This is a connection diagram for an embodiment. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0051] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0052] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper," "lower," "horizontal," or "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, terms such as "first" and "second" are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0053] Furthermore, the use of the term "horizontal" does not imply that the component must be absolutely horizontal, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0054] In the description of the embodiments of the present invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0055] The present invention will be further described in detail below with reference to embodiments and accompanying drawings:

[0056] Example 1

[0057] like Figure 1 The image shows a basic embodiment of a scoring method for striking competitions according to the present invention. The specific steps are as follows:

[0058] S101 collects audio and video data from the competition venue.

[0059] The goal is simply to capture both audio and video data from the competition venue. Separate audio and video capture devices can be set up, or a device capable of capturing both can be configured. During the capture process, audio and video data from the competition venue are generally captured synchronously. The captured data can be stored or used in real-time.

[0060] S102 filters out ambient noise from the audio data to obtain clean audio data.

[0061] Step S102 is mainly because the collected audio data may contain ambient noise, which can cause interference and hinder subsequent processing. Removing ambient noise will yield pure audio data, which is the audio data from the competition venue that is only related to the competition.

[0062] S103. Based on the clean audio data, determine whether a valid strike is possible. If so, proceed to step S4; otherwise, continue collecting audio data until a valid strike is possible.

[0063] The phrase "may have been valid" is used here because even pure audio data can contain sounds such as shouts from both sides of the match and sounds of hitting the ground. Therefore, based on pure audio data, it can only be determined that a valid hit may have occurred. Further judgment based on video data from the match venue is needed to make an accurate assessment.

[0064] S104, extract video data within a preset time period before the moment when a valid impact may occur, and record it as valid video data.

[0065] After step S103, if it is determined that a valid impact may occur, it is necessary to extract video data from a period of time prior to the possible valid impact. Based on the information in the video data, the valid impact information is confirmed. In practical applications, the preset time here needs to balance the ability to effectively identify valid impact information with the amount of information processing.

[0066] S105, perform spatial-temporal modeling of the actions of both competitors in the valid video data, and confirm valid strike information based on the modeling results. Valid strike information includes the competitor who made the valid strike, the action used in the valid strike, and the body part hit by the valid strike. Valid strike information can be obtained through modeling combined with certain calculation methods.

[0067] S106, Scoring is performed based on the valid striking information. Different striking competitions have different scoring rules; once valid striking information is obtained, scoring can be performed according to the scoring rules.

[0068] Example 2

[0069] The difference between Example 2 and Example 1 is that, after obtaining valid hit information, the valid hit information can be manually corrected to ensure scoring accuracy. For example, an interactive prompt can be made asking whether to manually correct the valid hit information. If manual correction is made, scoring is based on the manually corrected valid hit information, and both the manually corrected and obtained valid hit information are stored in a log. Manual correction can be made at points in time when the scoring results are not retrieved, manually setting the participating player, the action used in the valid hit, and the part of the body hit. In other embodiments of the present invention, the final scoring result and the scoring process can also be corrected, and all correction data can be recorded in a log.

[0070] Example 3

[0071] Example 3 is a preferred embodiment of the scoring method for striking competitions of the present invention, and the specific steps are as follows:

[0072] S301. Microphones and cameras are evenly distributed around the competition venue to collect audio and video data. The specific placement is determined by ensuring equipment safety, minimizing disruption to the competition, preventing interference during data collection, avoiding mutual interference between devices, ensuring high-quality data collection, and minimizing disruption to spectators. Preferably, there are five or more cameras, suspended around the perimeter and overhead of the venue. As another preferred option, the number of microphones equals the number of cameras, and they are positioned close together for synchronized data collection. Both microphones and cameras should point towards the center of the competition venue. Simultaneously, all data collection devices can transmit audio and video data via wired, wireless, or bridging methods.

[0073] In some embodiments of the present invention, the number of microphones and cameras, their installation locations, and data transmission methods can be adjusted as needed.

[0074] In some embodiments of the present invention, before the competition, initialization settings can be performed first. The initialization settings are as follows: (1) Detect whether a new compatible microphone and camera have been added, or whether the existing microphone or camera has been modified. If a new microphone or camera is detected, or the current microphone or camera has changed, the status of the new device can be manually set, or the current status can be automatically confirmed through mutual detection of multiple devices. (2) After the microphone and camera are driven, if information is saved in real time selectively, the selected data items can be stored in real time in common audio or video formats.

[0075] S302 uses a filtering algorithm to filter ambient noise in real time on the audio data to obtain the purest possible audio data. At this time, the filtered pure audio data is used as the input of the sound analysis function, and the output of the sound analysis function is whether a valid hit may have occurred.

[0076] The sound analysis function combines machine learning and deep learning techniques. By building and training a deep learning network, an effective deep learning network and its weights for recognizing impact sounds can be obtained. For example, LSTM and Transformer networks can be built and trained. The sound analysis function processes audio data by calling this network, and can identify and return whether a valid impact may have occurred.

[0077] If multiple microphones are used, multiple clean audio data will be generated. The input of each microphone will be processed in parallel using the same method. The final results will be voted on and tallied. If, within a predetermined time, more than half of the microphone signals, after passing through the above filtering and sound analysis functions, output that a valid hit may have occurred, then a valid hit is considered to have occurred in the competition area. Otherwise, a valid hit is considered not to have occurred. If a valid hit may have occurred, then step S303 is executed.

[0078] S303, extract all video data within three seconds before the possible valid strike moment. In other embodiments of the present invention, the extraction time of the video data can also be adjusted. The movement states of both sides on the field will be captured by one or more depth-of-field or non-depth-of-field cameras. The generated temporal data can be used to perform digital modeling of the human body shape, position, and state in a temporal manner using spatial modeling algorithms. For example, truncated signed distance function TSDF with time information recording can be used to confirm valid strike information. Valid strike information includes the side that made the valid strike, the action used in the valid strike, and the part of the body hit by the valid strike. If multiple cameras are present, the information from multiple cameras will be used in combination. The model obtained from the modeling (with spatial and temporal sequence modeling) will be used to locate the side that made the valid strike, the action used in the valid strike, and the part of the body hit by the valid strike. This information will then be used as input data for subsequent scoring functions.

[0079] There are various algorithms for obtaining valid impact information, such as the following two-dimensional and three-dimensional methods:

[0080] (1) Two-dimensional method

[0081] Each camera's video feed is processed individually. For a given video dataset, a model is first created. The system then uses a segmentation algorithm to identify the states of the competitors and weapons within the video data, and further identifies the relationship between the weapons and the competitors, i.e., valid firing information. The segmentation algorithm employs a deep learning-based segmentation network. By building and effectively training a deep learning network, such as UNet or UNet3D networks, regions of interest can be segmented from the camera video data. Since occlusions may exist in the video data, a completion prediction algorithm can be used to address occluded competitors and weapons during processing.

[0082] In addition, if there are multiple cameras, a voting mechanism can be used. During the voting process, the video data from each camera may have different results in the algorithm recognition. Therefore, the results of each camera will be counted. If the voting result of an option is greater than or equal to half of the total number of cameras, this result will be taken as the final result. If there is no final result or there are two results with a 50% vote rate, a specific error message will be fed back.

[0083] (2) Three-dimensional method

[0084] Data from each camera is stitched together into a complete data matrix based on location information. For images from the same time period within the video data, a 3D reconstruction algorithm is used to transform the complete data matrix into a 3D reconstruction model. All generated 3D reconstruction models are then concatenated over time to create a 4D reconstruction model incorporating temporal information. This 4D reconstruction model is then used to locate the player who made the strike, determine their striking motion, and identify the body part struck.

[0085] In other embodiments of the present invention, other methods may also be used for identification. Taking the above-described two-dimensional and three-dimensional methods as examples, the specific methods for identifying valid impact information are as follows:

[0086] The action used for an effective strike can be identified using deep learning networks:

[0087] Deep learning networks are a machine learning method that simulates the connections between neurons in the human brain to perform complex pattern recognition and decision-making tasks. Deep learning networks mimic the function, characteristics, and structure of biological neurons. By specifying parameters or methods such as loss functions, the parameters of neurons can be adjusted during backpropagation, thereby obtaining the desired results during forward propagation. Common deep learning network backbones include convolutional neural networks, Transformer networks, and RNN networks.

[0088] First, a large amount of data on different movements is collected. All collected data is then filtered to remove harmful data that could negatively impact the training of the deep learning network. Simultaneously, the data is manually labeled with tags that effectively score points during the movement, and additional tags that do not score points can be added to improve the deep learning network's recognition capabilities. After training and optimization, the deep learning network can then be used to identify different effective striking movements in practical applications.

[0089] Regarding the side that made a valid strike:

[0090] Weapons and the two players are modeled separately to obtain weapon models and player models. Collision detection between weapon models and player models can be performed using collision detection algorithms such as bounding volume collision detection, separation axis theorem, and ray casting. If a collision occurs between weapon models and player models, the effective player can be determined based on the information from the weapon models and player models.

[0091] Regarding the areas that were effectively hit:

[0092] Modeling is performed separately for the weapon and the two players, resulting in weapon models and player models. When an attack occurs, the positional parameters of the player model hit by the weapon are recorded. This hit player is then registered onto a standard human model, and the positional parameters of the hit body part are simultaneously registered to the standard human model, classifying the specific body part. This allows the determination of the effective hit area. In other words, the distance between the weapon model and the two player models is calculated separately, and combined with the standard human model, the effective hit area is determined.

[0093] S304, based on the obtained valid hitting information and combined with the preset scoring rules, completes the scoring through the scoring function.

[0094] Example 4

[0095] like Figure 2 As shown, a scoring system for striking competitions includes a data acquisition module, a filtering module, an audio recognition module, an extraction module, a video recognition module, and a scoring module.

[0096] The acquisition module is used to acquire audio data and video data from the competition venue.

[0097] The filtering module is used to filter out ambient noise from the audio data to obtain clean audio data;

[0098] The audio recognition module is used to determine whether a valid strike may occur based on the audio data. If so, the extraction module is activated; otherwise, audio data is continuously collected until a valid strike may occur.

[0099] The extraction module is used to extract video data within a preset time period before the moment when an effective impact may occur, and record it as effective video data.

[0100] The video recognition module is used to perform spatial and temporal modeling of the actions of both sides in the valid video data, and to confirm valid hitting information based on the modeling results. The valid hitting information includes the side that made the valid hit, the action used in the valid hit, and the part of the body hit by the valid hit.

[0101] The scoring module is used to score based on the valid hit information. During initial setup, the models and functions involved in each module can be pre-stored in the corresponding module to improve the overall system efficiency.

[0102] In other embodiments of the scoring system for striking competitions of the present invention, each module may also be configured according to the scoring methods for striking competitions in Embodiments 2 and 3, as a preferred embodiment of the scoring system for striking competitions.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of each module is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple modules may be combined or integrated into another device, or some features may be ignored or not executed.

[0104] The modules described as separate components may or may not be physically separate. A component shown as a module can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0105] Furthermore, the modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0106] This invention also provides a computer device. The computer device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0107] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0108] The computer device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device may include, but is not limited to, a processor and memory.

[0109] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0110] The memory can be used to store the computer program and / or module, and the processor implements various functions of the computer device by running or executing the computer program and / or module stored in the memory, and by calling the data stored in the memory.

[0111] If the modules / units integrated into the computer device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0112] The electronic device, computer storage medium, computer program product or chip provided in this embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0113] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A scoring method for striking-type competitive sports, characterized in that, Includes the following steps: S1 collects audio and video data from the competition venue. S2, filter out ambient noise from the audio data to obtain clean audio data; the specific method is as follows: Use filtering algorithms to remove ambient noise from audio data; S3. Based on the clean audio data, determine whether a valid strike is possible. If so, proceed to step S4; otherwise, continue collecting audio data until a valid strike is possible. Specifically, determining whether a valid strike is possible based on the clean audio data involves using a sound analysis function to determine whether a valid strike has occurred. S4, extract video data within a preset time before the moment when a valid impact may occur, and record it as valid video data; S5, perform spatial-temporal modeling of the actions of both competitors in the valid video data, and confirm valid strike information based on the modeling results. The valid strike information includes the competitor who made the valid strike, the action used in the valid strike, and the body part hit by the valid strike. Based on location information, the effective video data from each camera are stitched together into a data matrix. A 3D reconstruction algorithm is used to transform the data corresponding to each moment in all data matrices into 3D reconstruction models. All 3D reconstruction models are concatenated on a time scale to obtain a 4D reconstruction model. Effective impact information is obtained based on the 4D reconstruction model. Specifically, the effective impact information is obtained by identifying the action used in the effective impact through a trained classification network model. The contestant who committed the effective impact is determined by: modeling the weapon and the contestants separately to obtain weapon models and contestant character models; determining whether a collision occurred based on the collision detection results between the weapon model and the two character models; and identifying the contestant who committed the effective impact based on the information from the weapon model and the two character models. The part of the body hit by the effective impact is determined by: modeling the weapon and the contestants separately to obtain weapon models and contestant character models; calculating the distance between the weapon model and the two character models; and combining this distance with a standard human body model to determine the part of the body hit by the effective impact. S6. Scoring is performed based on the valid strike information.

2. The scoring method for striking competitions according to claim 1, characterized in that: In step S1, the audio data on the competition field is specifically collected by microphones evenly distributed around the competition field and / or on the top of the competition field; the video data on the competition field is specifically collected by cameras evenly distributed around the competition field and / or on the top of the competition field. Step S2 is as follows: If more than half of the microphones collect audio data within a preset time, and it is determined that a valid hit may have occurred, then a valid hit may have occurred; otherwise, it is considered that no valid hit has occurred.

3. The scoring method for striking competitions according to claim 2, characterized in that, Step S5 specifically involves performing the following steps on the valid video data from each camera: The obscured parts of the contestants and weapons in the valid video data are processed using a completion prediction algorithm to obtain complete and valid video data. The actions of both sides in the complete and valid video data are modeled in a spatial and temporal manner. The state of the two sides and their weapons is identified through a segmentation algorithm, and then the valid hitting information is obtained.

4. The scoring method for striking competitions according to claim 3, characterized in that, Step S6 is followed by: manually correcting the valid strike information.

5. A scoring system for striking-type competitive games, characterized in that: It includes a data acquisition module, a filtering module, an audio recognition module, an extraction module, a video recognition module, and a scoring module; The acquisition module is used to acquire audio data and video data from the competition venue. The filtering module is used to filter out ambient noise from the audio data to obtain clean audio data; specifically, it uses a filtering algorithm to filter out ambient noise from the audio data. The audio recognition module is used to determine whether a valid strike may occur based on the audio data. If so, the extraction module is activated; otherwise, audio data is continuously collected until a valid strike may occur. Specifically, determining whether a valid strike may occur based on the audio data involves using a sound analysis function to determine whether a valid strike has occurred. The extraction module is used to extract video data within a preset time period before the moment when an effective impact may occur, and record it as effective video data. The video recognition module is used to perform spatial and temporal modeling of the actions of both competitors in the valid video data, and to confirm valid strike information based on the modeling results. The valid strike information includes the competitor who made the valid strike, the action used in the valid strike, and the body part hit by the valid strike. Based on location information, the effective video data from each camera are stitched together into a data matrix. A 3D reconstruction algorithm is used to transform the data corresponding to each moment in all data matrices into 3D reconstruction models. All 3D reconstruction models are concatenated on a time scale to obtain a 4D reconstruction model. Effective impact information is obtained based on the 4D reconstruction model. Specifically, the effective impact information is obtained by identifying the action used in the effective impact through a trained classification network model. The contestant who committed the effective impact is determined by: modeling the weapon and the contestants separately to obtain weapon models and contestant character models; determining whether a collision occurred based on the collision detection results between the weapon model and the two character models; and identifying the contestant who committed the effective impact based on the information from the weapon model and the two character models. The part of the body hit by the effective impact is determined by: modeling the weapon and the contestants separately to obtain weapon models and contestant character models; calculating the distance between the weapon model and the two character models; and combining this distance with a standard human body model to determine the part of the body hit by the effective impact. The scoring module is used to score points based on the valid strike information.

6. A computer device, characterized in that, The system includes a memory and one or more processors, the memory being coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, which, when executed by the processor, cause the server to perform as claimed in claim 1. The method described in any one of the following 4.

7. A computer storage medium, characterized in that, The computer storage medium includes computer instructions that, when executed on an electronic device, cause the electronic device to perform the method as described in any one of claims 1-4.

Images