Archery action recognition method and electronic equipment

Abstract

The invention provides an archery action recognition method and electronic equipment, and the method comprises the steps: determining the key point position information of a plurality of key points in each frame of first image according to multiple frames of first images related to the archery action in the archery process of a user, carrying out the recognition processing of the first archery action according to the key point position information of each key point, and carrying out the recognition processing of the second archery action. And performing first archery action recognition processing on the basis of the image classification model according to the multiple frames of first images to obtain a first archery action stage recognition result, performing second archery action recognition processing on the basis of the image classification model according to the multiple frames of first images to obtain a second archery action stage recognition result, and obtaining a target archery action stage recognition result according to the first archery action stage recognition result and the second archery action stage recognition result. Thus, a more accurate target archery action stage recognition result can be obtained according to the first archery action stage recognition result and the second archery action stage recognition result, so that a user can adjust the archery action more accurately according to the archery action stage recognition result, and archery is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to an archery action recognition method and an electronic device. BACKGROUND

[0002] Archery is a sport that uses the elastic force of a bow to shoot arrows to compete for accuracy within a certain distance. Currently, archery is increasingly popular in sports competitions and entertainment projects. In the process of archery, different archery actions are different in different archery action stages. Therefore, it is necessary to recognize the archery action so that the user can analyze the archery action in different archery action stages according to the archery action stage recognition result, find out the problem, and then adjust the archery action to make the archery result more accurate.

[0003] In summary, how to more accurately recognize the archery action stage so that the user can analyze the archery action in different archery action stages in the process of archery is a problem that needs to be solved at present. SUMMARY

[0004] The embodiments of the present application provide an archery action recognition method and an electronic device, which can more accurately recognize the archery action stage in the process of archery, obtain the archery action stage recognition result of the process of archery, so that the user can adjust the archery action in different archery action stages according to the archery action stage recognition result, thereby making the archery result more accurate.

[0005] To solve the above technical problems, in a first aspect, the embodiments of the present application provide an archery action recognition method, which comprises: determining a plurality of first images related to archery action in the process of archery of a user; determining key point position information of a plurality of key points in each first image according to each first image, and performing first archery action recognition processing according to the key point position information of each key point corresponding to each first image, to obtain a first archery action stage recognition result corresponding to the plurality of first images; and performing second archery action recognition processing based on an image classification model according to the plurality of first images, to obtain a second archery action stage recognition result corresponding to the plurality of first images; and obtaining a target archery action stage recognition result corresponding to the plurality of first images according to the first archery action stage recognition result and the second archery action stage recognition result.

[0006] According to the technical solution, the key point position information of the plurality of key points in each frame of the first images is determined according to the plurality of frames of the first images related to the archery action of the user during the archery process, and then the first archery action recognition processing is performed according to the key point position information of each key point, to obtain the first archery action stage recognition result corresponding to the plurality of frames of the first images. The second archery action recognition processing is performed on the archery action of the user according to the plurality of frames of the first images based on the image classification model, to obtain the second archery action stage recognition result corresponding to the plurality of frames of the first images. Then, the target archery action stage recognition result is obtained according to the first archery action stage recognition result and the second archery action stage recognition result. In this way, the archery action stage recognition during the archery process of the user is more accurate, and a more accurate archery action stage recognition result is obtained, so that the user can adjust the archery action according to the archery action stage recognition result, and the archery is more accurate.

[0007] In a possible implementation of the first aspect, after the key point position information of the plurality of key points in each frame of the first images is determined according to each frame of the first images, the method further includes: generating a key point position information time sequence of the key points according to the key point position information of the plurality of key points in each frame of the first images, to perform the first archery action recognition processing according to the key point position information of each key point corresponding to each frame of the first images included in the key point position information time sequence. The key point position information time sequence includes the key point position information of each key point corresponding to each frame of the first images arranged in a forward time sequence.

[0008] In a possible implementation of the first aspect, the first archery action stage recognition result corresponding to the plurality of frames of the first images is obtained by performing the first archery action recognition processing according to the key point position information of each key point corresponding to each frame of the first images, including: determining distance information between each key point corresponding to each frame of the first images according to the key point position information of each key point corresponding to each frame of the first images, and obtaining a distance change trend of the key points corresponding to the plurality of frames of the first images according to the distance information; obtaining a first archery action stage corresponding to a first target frame of the first images in the plurality of frames of the first images and first confidence indication information corresponding to the first archery action stage according to the distance change trend of the key points corresponding to the plurality of frames of the first images; and obtaining the first archery action stage recognition result corresponding to the plurality of frames of the first images according to the first archery action stage corresponding to the first target frame of the first images and the first confidence indication information corresponding to the first archery action stage.

[0009] By using the technical solution, the distance change trend between the key points in the plurality of first images is used to analyze the key points in the archery process of the user, and the archery action stage and the corresponding confidence indication information are obtained. In this way, the archery action is analyzed based on the distance information of the key points, so that the archery action recognition processing is more accurate, and a more accurate archery action stage recognition result is obtained.

[0010] In a possible implementation of the first aspect, the first archery action stage corresponding to the target frame image in the plurality of first images is obtained according to the distance change trend of the key points corresponding to the plurality of first images, including: if it is determined according to the distance change trend of the key points corresponding to the plurality of first images that the distance change of the key points corresponding to the first target frame image in the plurality of first images satisfies a first change condition, the first archery action stage corresponding to the first target frame image is determined as an arrow drawing stage; if it is determined according to the distance change trend of the key points corresponding to the first target frame image in the plurality of first images that the distance change of the key points corresponding to the first target frame image in the plurality of first images satisfies a second change condition, the first archery action stage corresponding to the first target frame image is determined as an arrow storing stage; and if it is determined according to the distance change trend of the key points corresponding to the first target frame image in the plurality of first images that the distance change of the key points corresponding to the first target frame image in the plurality of first images satisfies a third change condition, the first archery action stage corresponding to the first target frame image is determined as an arrow releasing stage.

[0011] By using the technical solution, the distance change trend between the key points in the plurality of first images is used to analyze the key points in the archery process of the user, and the archery action stage and the corresponding confidence indication information are obtained. In this way, the archery action is analyzed based on the distance information of the key points, so that the archery action recognition processing is more accurate, and a more accurate archery action stage recognition result is obtained.

[0012] In a possible implementation of the first aspect, the first change condition is that the distance of the key point corresponding to the first target frame image in the distance change trend is the same between every two frames, and the difference is greater than zero and less than a preset first difference threshold; the second change condition is that the distance of the key point corresponding to the first target frame image in the distance change trend tends to zero between every two frames; and the third change condition is that the distance of the key point corresponding to the first target frame image in the distance change trend is greater than a preset second difference threshold between every two frames.

[0013] By using the technical solution, the distance change trend between the key points in the plurality of first images is used to analyze the key points in the archery process of the user, and the archery action stage and the corresponding confidence indication information are obtained. In this way, the archery action is analyzed based on the distance information of the key points, so that the archery action recognition processing is more accurate, and a more accurate archery action stage recognition result is obtained.

[0014] In a possible implementation of the first aspect, the second archery action recognition processing is performed on the plurality of first images based on the image classification model to obtain a target archery action stage recognition result corresponding to the target frame image, including: performing the second archery action recognition processing on the plurality of first images based on the image classification model to obtain an archery action category and archery action category confidence indication information of each first image; obtaining a second target archery action stage corresponding to a second target frame image in the plurality of first images and second target archery action stage corresponding confidence indication information based on the archery action category and archery action category confidence indication information of each first image; and obtaining the target archery action stage recognition result corresponding to the plurality of first images based on the second target archery action stage corresponding to the second target frame image and the second target archery action stage corresponding confidence indication information.

[0015] According to the technical solution, the image classification model is used to classify each first image to obtain the archery action category and archery action category confidence indication information of each first image, and then the second target archery action stage corresponding to the second target frame image in the plurality of first images and the second target archery action stage corresponding confidence indication information are obtained. In this way, the second archery action recognition is performed based on the image classification model, and the subsequent fusion is performed, so that a more accurate archery action recognition result is obtained.

[0016] In a possible implementation of the first aspect, the target archery action stage recognition result includes a target archery action stage corresponding to the target frame image, and the target archery action stage recognition result corresponding to the plurality of first images is obtained based on the first archery action stage recognition result and the second archery action stage recognition result, including: in a case where it is determined that the first target frame image and the second target frame image exist in the same frame, and the first archery action stage corresponding to the first target frame image and the second archery action stage corresponding to the second target frame image are the same archery action stage, the first weight coefficient corresponding to the first confidence indication information and the second weight coefficient corresponding to the second confidence indication information are determined based on the second confidence indication information, where if the second confidence indication information is greater than a preset first confidence threshold, the first weight coefficient is less than the second weight coefficient, the target frame image is the second target frame image, if the second confidence indication information is less than a preset second confidence threshold, the first weight coefficient is greater than the second weight coefficient, and the target frame image is the first target frame image; the target confidence indication information corresponding to the target frame image is obtained based on the first confidence indication information, the second confidence indication information, the first weight coefficient, and the second weight coefficient; and in a case where it is determined that the target confidence indication information is greater than a preset third confidence threshold, the first archery action stage or the second archery action stage is determined as the target archery action stage corresponding to the target frame image.

[0017] According to the technical solution, the first arrow shooting action recognition result and the second arrow shooting action recognition result are fused and recognized, so that a more accurate arrow shooting action recognition result can be obtained. Moreover, the weight coefficient is dynamically adjusted based on the second confidence indication information, so that the target confidence indication information can better reflect the accuracy of the target arrow shooting action stage corresponding to the target frame image, thereby making the target arrow shooting action recognition of the target frame image more accurate.

[0018] In a possible implementation of the first aspect, the plurality of first images have corresponding time stamp information, the target arrow shooting action stage recognition result further includes a start time, a duration of the target arrow shooting action stage, and inspection stability information corresponding to the target arrow shooting action stage, and the target arrow shooting action stage recognition result corresponding to the plurality of first images is obtained according to the first arrow shooting action stage recognition result and the second arrow shooting action stage recognition result, including: obtaining the start time and the duration of the target arrow shooting action stage according to the time stamp information corresponding to the target frame image; and determining the detection stability information of the target arrow shooting action stage according to the first confidence indication information and the second confidence indication information corresponding to the target frame image.

[0019] According to the technical solution, the start time, the duration and the detection stability information of the target arrow shooting action stage corresponding to the target frame image are determined, which is more conducive to the analysis of the corresponding arrow shooting action stage by the user.

[0020] In a possible implementation of the first aspect, the plurality of first images related to the arrow shooting action in the user's arrow shooting process are determined, including: obtaining an arrow shooting action video in the user's arrow shooting process, and obtaining the plurality of first images according to the arrow shooting action video; or obtaining a plurality of continuous images related to the arrow shooting action in the user's arrow shooting process, and obtaining the plurality of first images.

[0021] According to the technical solution, the arrow shooting action recognition is performed based on the arrow shooting action video or the plurality of continuous images, so that the arrow shooting action recognition can be more accurate.

[0022] In a possible implementation of the first aspect, the key point position information of the plurality of key points in each first image is determined according to each first image, including: inputting each first image into a key point detection model, so that the key point detection model determines the key point position information of the plurality of key points in each first image according to each first image, and the key points include hand key points.

[0023] According to the technical solution, the key point position information is recognized based on the key point detection model, so that the key point recognition efficiency of the plurality of first images is improved.

[0024] In a possible implementation of the first aspect, the second archery action recognition processing based on the image classification model and the plurality of first images comprises: performing cropping processing on a target region in the plurality of first images to obtain a plurality of second images, inputting the plurality of second images into the image classification model, so that the image classification model performs the second archery action recognition processing based on the plurality of second images, the target region comprising at least one of a hand region, a face region and an archery equipment region; or inputting the plurality of first images into the image classification model, so that the image classification model performs the second archery action recognition processing based on the plurality of first images.

[0025] According to the technical solution, at least one of the hand region, the face region and the archery equipment region in the first image that can better reflect the archery action is cropped to obtain the second image, so that the image classification model performs the second archery action recognition processing based on the plurality of second images, and the input is the image that better reflects the archery action, so that more accurate archery action category and archery action category confidence information can be obtained.

[0026] In a second aspect, the implementation manner of the present application further discloses an electronic device, comprising: a processor and a memory in communication connection with the processor; the memory stores computer execution instructions; and the processor executes the computer execution instructions stored in the memory, so that the electronic device implements the archery action recognition method provided by any one of the implementation manners of the first aspect.

[0027] In a third aspect, the implementation manner of the present application further discloses a computer readable storage medium, which stores a computer program, the computer program can be executed by a computer cluster to implement the inference method based on the neural network model provided by any one of the implementation manners of the first aspect, and / or the archery action recognition method provided by any one of the implementation manners of the second aspect.

[0028] In a fourth aspect, the implementation manner of the present application further discloses a computer program product, comprising a computer program, the computer program is executed by a computer cluster to implement the inference method based on the neural network model provided by any one of the implementation manners of the first aspect, and / or the archery action recognition method provided by any one of the implementation manners of the second aspect.

[0029] The related beneficial effects of the above-mentioned second aspect to fourth aspect can be referred to the related description in the first aspect or the second aspect, and will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the present application, the drawings used in the implementation manner description will be briefly introduced as follows.

[0031] Figure 1A flowchart of a shooting action recognition method provided by an embodiment of the present application is shown in FIG. 1.

[0032] Figure 2 A flowchart of determining a first shooting action phase recognition result provided by an embodiment of the present application is shown in FIG. 2.

[0033] Figure 3 A flowchart of determining a second shooting action phase recognition result provided by an embodiment of the present application is shown in FIG. 3.

[0034] Figure 4 A flowchart of determining a target shooting action phase recognition result provided by an embodiment of the present application is shown in FIG. 4.

[0035] Figure 5 A structural diagram of an electronic device provided by an embodiment of the present application is shown in FIG. 5. DETAILED DESCRIPTION

[0036] As mentioned above, in the shooting process, the shooting actions are different in different shooting phases. In the shooting simulation or the shooting training process, the shooting actions need to be recognized to identify the different shooting action phases of the user in the shooting process, and then the shooting actions in different shooting action phases are analyzed to adjust the shooting actions, so that the shooting actions in each shooting action phase are more standard, and then the shooting result is more accurate.

[0037] The existing shooting action recognition is mostly to set simple sensors such as accelerometers, gyroscopes, magnetometers, etc. on the wrist or bow of the user, to recognize the changes of the hand movement or the changes of the bow movement of the user based on the acquired motion information according to the traditional body sensing recognition system, so as to realize the recognition of the shooting action of the user, or to recognize the shooting action of the user in the shooting process through a pre-set rule matching algorithm. The two recognition methods are not accurate enough for the recognition of complex human actions, lack naturalness and interactivity, and the traditional body sensing recognition system has problems such as high delay, low accuracy, poor generalization in motion decomposition and action intention judgment.

[0038] Based on this, the application provides an archery action recognition method, which can be applied to real-time recognition in the archery process and can also be used for process analysis after the archery is completed. The method comprises the following steps: acquiring a plurality of first images related to archery actions in the archery process of a user, determining key point position information of a plurality of key points in each first image, performing first archery action recognition processing based on the key point position information, obtaining a first archery action stage recognition result, further performing second archery action recognition processing based on an image classification model and the plurality of first images, obtaining a second archery action stage recognition result, and finally obtaining a final archery action stage recognition result according to the first archery action stage recognition result and the second archery action stage recognition result. In this way, the current archery action stage can be recognized in real time in the archery process, different archery action stages in the entire archery process can also be recognized, and double recognition makes the archery action stage recognition in the archery process of the user more accurate, so that a more accurate archery action stage recognition result is obtained, so that the user can adjust the archery action according to the archery action stage recognition result, and the archery is more accurate.

[0039] Referring to Figure 1 The archery action recognition method provided by the implementation manner of the application specifically comprises the following steps.

[0040] S100, a plurality of first images related to archery actions in the archery process of a user are determined.

[0041] S200, key point position information of a plurality of key points in each first image is determined according to each first image, and first archery action recognition processing is performed according to the key point position information of each key point corresponding to each first image, so as to obtain a first archery action stage recognition result corresponding to the plurality of first images.

[0042] S300, second archery action recognition processing is performed based on an image classification model and the plurality of first images, so as to obtain a second archery action stage recognition result corresponding to the plurality of first images.

[0043] S400, a target archery action stage recognition result corresponding to the plurality of first images is obtained according to the first archery action stage recognition result and the second archery action stage recognition result.

[0044] The arrow shooting action recognition method provided by the implementation manner of the present application determines the key point position information of a plurality of key points in each first image according to a plurality of first images related to arrow shooting actions in the user's arrow shooting process, and then performs first arrow shooting action recognition processing according to the key point position information of each key point to obtain first arrow shooting action stage recognition results corresponding to the plurality of first images, and performs second arrow shooting action recognition processing on the plurality of first images based on an image classification model to obtain second arrow shooting action stage recognition results corresponding to the plurality of first images, so as to obtain target arrow shooting action stage recognition results according to the first arrow shooting action stage recognition results and the second arrow shooting action stage recognition results. In this way, the first arrow shooting action stage recognition results are obtained by performing first arrow shooting action recognition processing on the key point position information in the plurality of first images in the user's arrow shooting process, and the second arrow shooting action stage recognition results are obtained by performing second arrow shooting action recognition processing on the user's arrow shooting actions based on the image classification model according to the plurality of first images, and then the target arrow shooting action stage recognition results are obtained according to the first arrow shooting action stage recognition results and the second arrow shooting action stage recognition results, so that the arrow shooting action stage recognition in the user's arrow shooting process is more accurate, and more accurate arrow shooting action stage recognition results are obtained, so that the user can adjust the arrow shooting action according to the arrow shooting action stage recognition results, and the arrow shooting is more accurate.

[0045] Firstly, step S100 is performed to determine a plurality of first images related to arrow shooting actions in the user's arrow shooting process.

[0046] In an implementation manner of the present application, the plurality of first images related to arrow shooting actions in the user's arrow shooting process are determined, including: obtaining an arrow shooting action video related to arrow shooting actions in the user's arrow shooting process, and obtaining the plurality of first images according to the arrow shooting action video.

[0047] For example, an image acquisition device is arranged near the user's arrow shooting area, and an arrow shooting action video containing complete arrow shooting actions is acquired based on the front or side camera of the image acquisition device, and a plurality of first images are obtained by performing a multi-frame extraction operation on the arrow shooting action video. Each image includes key information such as a user's hand region, a face region, and a bow and arrow device.

[0048] Further, the arrow shooting action video is an arrow shooting action video of the entire arrow shooting process or an arrow shooting action video of a part of the arrow shooting process (for example, an arrow shooting action video corresponding to the current arrow shooting process). Therefore, the plurality of first images are arrow shooting action images of the entire arrow shooting process or arrow shooting action images of a part of the arrow shooting process.

[0049] In another implementation manner of the present application, the plurality of first images related to arrow shooting actions in the user's arrow shooting process are determined, including: obtaining a plurality of continuous images related to arrow shooting actions in the user's arrow shooting process to obtain the plurality of first images.

[0050] For example, the front or side camera of the image acquisition device acquires a plurality of continuous archery action images containing the complete archery action process to obtain a plurality of first images.

[0051] The plurality of continuous images are archery action images of the entire archery process or archery action images of a part of the archery process (for example, archery action images corresponding to the current archery process). Therefore, the plurality of first images are archery action images of the entire archery process or archery action images of a part of the archery process.

[0052] It should be noted that each first image carries corresponding timestamp information.

[0053] Next, step S200 is performed to determine key point position information of a plurality of key points in each first image according to each first image, and to perform first archery action recognition processing according to the key point position information of each key point corresponding to each first image to obtain first archery action stage recognition results corresponding to the plurality of first images.

[0054] In an implementation manner of the present application, the key points include at least one of a hand key point, an archery key point, and a face key point.

[0055] The hand key point includes at least one joint point of a wrist point of a hand holding a bow wall and five fingers (thumb, index finger, middle finger, ring finger, and little finger) of a hand holding a bow string.

[0056] The archery key point includes at least one bow arm key point and at least one bow string key point.

[0057] The face key point includes at least one upper eyelid key point and at least one lower eyelid key point.

[0058] Preferably, the key points include the hand key point, and the hand key point includes a wrist point of a hand holding a bow wall and five fingers (index finger, middle finger, ring finger, and little finger) of a hand holding a bow string, four joint points (TIP, DIP, PIP, and MCP) of each finger, and three joint points of the thumb, totaling 20 key points.

[0059] Of course, in another implementation manner of the present application, in addition to the above 20 key points, one more joint point can be arranged between the palm joint of the thumb of the hand holding the bow string and the wrist joint to obtain a total of 21 key points.

[0060] Preferably, the key points include at least one key point in the face key point and the bow string key point.

[0061] Further, the first plurality of images are subjected to image detection processing to obtain key point position information of each key point in each first image, and a time sequence of the position information of each key point (e.g., a, b, …, k) is generated based on the frame order (e.g., 1, 2, …, n) of the first images, to obtain a time sequence of the position information of the key points, which is used for subsequent arrow shooting action stage analysis.

[0062] The time sequence of the position information of the key points is in the following form:

[0063]

[0064] The position information of the key point a in the n-th first image is: The position information of the key point b in the n-th first image is: The position information of the key point k in the n-th first image is:

[0065] Of course, the time sequence of the position information of the key points can also be in the following form:

[0066]

[0067] The position information of the key points a, b, …, k in the first first image is: The position information of the key points a, b, …, k in the second first image is: The position information of the key points a, b, …, k in the n-th first image is:

[0068] In the implementation of the present application, the position information of the key points can be three-dimensional coordinate information of the key points.

[0069] Next, taking the key points including hand key points as an example, the specific process of obtaining the key point position information of each key point in each first image is described in detail.

[0070] In the implementation of the present application, the key point position information of the plurality of key points in each first image is determined according to each first image, including: inputting each first image into a key point detection model, so that the key point detection model determines the key point position information of the plurality of key points in each first image according to each first image.

[0071] For example, the key point detection model is customized and trained in advance based on MediaPipe to obtain the key point detection model, such as a hand key point detection model.

[0072] ​​MediaPipe is an open-source multimedia machine learning model application framework, and the MediaPipe includes a hand landmark detection open-source framework for generating a hand landmark detection model.

[0073] The hand landmark detection model is composed of a palm detection model and a hand landmark detection model that cooperate with each other. Each frame of the first image is input, and the first frame of the first image is detected first to generate key point position information corresponding to the first frame of the image. Then, the second frame of the first image is detected to generate key point position information corresponding to the second frame of the image. In this way, the detection of all frames of the first image is completed, and the time sequence of the key point position information of the key points corresponding to the multiple frames of the first image is obtained. The time sequence of the key point position information includes the key point position information of each key point corresponding to each frame of the image arranged in the forward time sequence of each frame of the image.

[0074] It should be noted that, because the first image is an image containing a complete archery action, the number and type of key points detected in each frame of the first image are the same, and the position information of the same key points may be the same or different in different frames of the first image.

[0075] Further, according to the key point position information of each key point corresponding to each frame of the first image included in the time sequence of the key point position information, a first archery action recognition process is performed to obtain a first archery action stage recognition result corresponding to the multiple frames of the first image.

[0076] In the implementation manner of the present application, as shown in Figure 2 According to the key point position information of each key point corresponding to each frame of the first image, a first archery action recognition process is performed to obtain a first archery action stage recognition result corresponding to the multiple frames of the first image, including the following steps.

[0077] S210, according to the key point position information of each key point corresponding to each frame of the first image, the distance information between each key point corresponding to each frame of the first image is determined, and the distance change trend of the key points corresponding to the multiple frames of the first image is obtained according to the distance information.

[0078] For example, the hand landmark detection model determines the distance information between each two key points corresponding to each frame of the first image according to the key point position information of each two key points included in each frame of the first image included in the time sequence of the key point position information, obtains multiple distance information between each two key points corresponding to the multiple frames of the first image, and obtains the distance change trend of each key point corresponding to the multiple frames of the first image according to the multiple distance information of each key point.

[0079] For example, taking the index finger tip (Index_Tip, as an example of a hand key point) and the wrist point (as another example of a hand key point) as examples, the Euclidean distance between the two points is calculated:

[0080]

[0081] wherein, is the distance information of the key point i and the key point w corresponding to the first image of the t-th frame, the position information of the key point i in the first image of the t-th frame is (xi, yi), and the position information of the key point w in the first image of the t-th frame is (xw, yw). .

[0082] In this way, the distance information of the key point i (for example, the index finger tip) and the key point w (for example, the wrist point) in multiple frames of the first image is obtained, to obtain multiple distance information of the key point i and the key point w corresponding to multiple frames of the first image, so as to analyze the distance change trend of the two key points in the time dimension on multiple frames of the first image, and obtain the distance change trend of the two key points.

[0083] In this way, the distance change trend between all key points is obtained.

[0084] The distance change trend of each key point can be a time (i.e., frame)-distance change graph between each key point.

[0085] S220, according to the distance change trend of the key points corresponding to the multiple frames of the first image, obtaining the first archery action stage corresponding to the first target frame image in the multiple frames of the first image and the first confidence indication information corresponding to the first archery action stage.

[0086] For example, the hand key point detection model divides the multiple frames of the first image according to the distance change trend of the key points corresponding to the multiple frames of the first image, to obtain the first archery action stage corresponding to the first target frame image and the first confidence indication information corresponding to the first archery action stage.

[0087] Wherein, the first target frame image is at least one group, each group of first target frame image includes at least one frame of first image, and the first confidence indication information corresponding to each group of first target frame image is the average value of the first confidence indication information of each frame of first image included by the first target frame image. The first confidence indication information of each key point in each frame of first image is the average value of the first confidence indication information of each key point in each frame of first image. That is, the first confidence indication information of each key point in each frame of first image can be the score of the archery action in each frame of first image included by the first target frame image as the first archery action stage, or can be the probability information of the archery action in each frame of first image included by the first target frame image as the first archery action stage. ​

[0088] Further, in the implementation of the present application, if it is determined according to the distance change trend of the key points corresponding to the first target frame image in the plurality of first images that the distance change of the key points corresponding to the first target frame image satisfies a first change condition, it is determined that the first arrow action stage corresponding to the first target frame image is an arrow drawing stage.

[0089] For example, the first change condition is that the distance of the key points corresponding to the first target frame image in the distance change trend is the same between every two frames and the difference is greater than zero and less than a preset first difference threshold. That is, the first change condition is that the distance corresponding to the target frame in the distance change trend is steadily increasing frame by frame, and the steady increase of the distance between frames indicates that the bowstring is being steadily drawn, indicating that the bowstring is being drawn at present.

[0090] Therefore, the first image in the plurality of first images whose distance is steadily increasing frame by frame (i.e., satisfying the first change condition) is determined as the first target frame image, and the first arrow action stage corresponding to the first target frame image is the arrow drawing stage.

[0091] If it is determined according to the key point change trend of the key points corresponding to the first target frame image in the plurality of first images that the distance change of the key points corresponding to the first target frame image satisfies a second change condition, it is determined that the first arrow action stage corresponding to the first target frame image is a force storage stage.

[0092] For example, the second change condition is that the distance of the key points corresponding to the first target frame image in the distance change trend tends to zero between every two frames. That is, the second change condition is that the distance corresponding to the target frame in the distance change trend is relatively stable and does not change, and the distance difference tends to zero, and the relatively stable distance between frames indicates that the bowstring is also in a tight state at present, indicating that the force is being stored at present.

[0093] Therefore, the first image in the plurality of first images whose distance is relatively stable and does not change (i.e., satisfying the second change condition) is determined as the first target frame image, and the first arrow action stage corresponding to the first target frame image is the force storage stage.

[0094] If it is determined according to the key point change trend of the key points corresponding to the first target frame image in the plurality of first images that the distance change of the key points corresponding to the first target frame image satisfies a third change condition, it is determined that the first arrow action stage corresponding to the first target frame image is a release stage.

[0095] Exemplarily, the third change condition is that a difference between distances of the key point in the distance change trend corresponding to the first target frame image in each two frames is greater than a preset second difference threshold. That is, the third change condition is that the distance corresponding to the target frame in the distance change trend appears a sharp jump (for example, a distance slope corresponding to multiple frames increases). The sharp jump indicates that the speed of the finger tip between frames increases, and it indicates that the finger has released the bowstring, indicating that the current is releasing the bow and arrow.

[0096] Therefore, the first image in which the distance sharply jumps (that is, the second change condition is met) in the multiple frames of first images is determined as the first target frame image, and the first archery action stage corresponding to the first target frame image is the release stage.

[0097] Therefore, based on the distance change trend of each key point, the first change condition, the second change condition, and the third change condition, the first archery action stage corresponding to each first target frame image in the multiple frames of first images can be identified.

[0098] Further, the hand key point detection model also scores or calculates the probability of each frame of first image included in the first target frame image being in the first archery action stage to obtain the first confidence indication information of the first archery action of each frame of first image included in the first target frame image being in the first archery action stage.

[0099] S230, according to the first archery action stage corresponding to the first target frame image and the first confidence indication information corresponding to the first archery action stage, obtaining the first archery action stage recognition result corresponding to the multiple frames of first images.

[0100] Exemplarily, if the multiple frames of first images only include a group of first target frame images, the first archery action stage corresponding to the group of first target frame images and the first confidence indication information corresponding to the first archery action stage are taken as the first archery action stage recognition result corresponding to the multiple frames of first images.

[0101] If the multiple frames of first images include multiple groups of first target frame images, the first archery action stage corresponding to each group of first target frame images and the first confidence indication information corresponding to the first archery action stage are taken as the first archery action stage recognition result corresponding to the multiple frames of first images.

[0102] Next, step S300 is performed, and based on the image classification model, the second archery action recognition processing is performed according to the multiple frames of first images to obtain the second archery action stage recognition result corresponding to the multiple frames of first images.

[0103] Exemplarily, to enhance the fault tolerance capability for the abnormal state, an image classification model (MobileNetV) is introduced, and the second archery action recognition processing is performed on the multiple frames of first images based on the image classification model to obtain the second archery action stage recognition result corresponding to the multiple frames of first images.

[0104] In the implementation of the present application, the second archery action recognition processing is performed on the multiple frames of first images based on the image classification model to obtain the second archery action stage recognition result corresponding to the multiple frames of first images, which includes: performing cropping processing on the target region in the multiple frames of first images to obtain multiple frames of second images, and inputting the multiple frames of second images into the image classification model, so that the image classification model performs the second archery action recognition processing according to the multiple frames of second images, and the target region includes at least one of a hand region, a face region and an archery appliance region.

[0105] Exemplarily, at least one of the hand region, the face region and the archery appliance region in the first image is cropped to obtain the target region, so as to obtain the multiple frames of second images according to the target region of the multiple frames of first images.

[0106] For example, each frame of second image includes a hand+face region or a hand+archery appliance region.

[0107] Further, in another implementation of the present application, the multiple frames of first images are directly input into the image classification model, so that the image classification model performs the second archery action recognition processing according to the multiple frames of first images.

[0108] The image classification model learns the static image features of typical actions through a training set, and can accurately distinguish still frames.

[0109] Further, as shown in the figure, Figure 3 In the implementation of the present application, the second archery action recognition processing is performed on the multiple frames of first images based on the image classification model to obtain the second archery action stage recognition result corresponding to the multiple frames of first images, which includes the following steps.

[0110] S310, the second archery action recognition processing is performed on the multiple frames of first images based on the image classification model to obtain the archery action category and the archery action category confidence indication information of each frame of first image.

[0111] Exemplarily, the multiple frames of first images are input into the image classification model, and the second archery action recognition processing is performed by the image classification model to output the archery action category and the corresponding archery action category confidence indication information of each frame of image.

[0112] The archery action category includes categories such as drawing a bow, accumulating force, releasing, idling and abnormality, one frame of image corresponds to one or more categories, and each category corresponds to archery action category confidence indication information.

[0113] The shooting action class confidence indication information corresponding to each frame of image is obtained based on a softmax function, and is used for subsequent fusion judgment.

[0114] The shooting action class confidence indication information corresponding to each frame of image can be probability information corresponding to each shooting action class (S320) ), or score information of each shooting action class.

[0115] S320, according to the shooting action class and the shooting action class confidence indication information of each frame of first image, the second shooting action stage corresponding to the second target frame image in the plurality of frames of first image and the second confidence indication information corresponding to the second shooting action stage are obtained.

[0116] Exemplarily, the continuous plurality of frames of first image of the same shooting action class are taken as a group of second target frame image, the second shooting action stage corresponding to each group of second target frame image is obtained, and the average value of the shooting action class confidence indication information corresponding to the continuous plurality of frames of first image of the same shooting action class is taken as the second confidence indication information of the second shooting action corresponding to the group of second target frame image.

[0117] S330, according to the second shooting action stage corresponding to the second target frame image and the second confidence indication information corresponding to the second shooting action stage, the second shooting action stage recognition result corresponding to the plurality of frames of first image is obtained.

[0118] Exemplarily, if the plurality of frames of first image only includes a group of second target frame image, the second shooting action stage corresponding to the group of second target frame image and the second confidence indication information corresponding to the second shooting action stage are taken as the second shooting action stage recognition result corresponding to the plurality of frames of first image.

[0119] If the plurality of frames of first image includes a plurality of groups of second target frame image, the second shooting action stage corresponding to each group of second target frame image and the second confidence indication information corresponding to the second shooting action stage are taken as the second shooting action stage recognition result corresponding to the plurality of frames of first image.

[0120] It should be noted that the second shooting action recognition manner based on the plurality of frames of second image is the same as the second shooting action recognition manner based on the plurality of frames of first image. The second shooting action recognition processing is performed based on the plurality of frames of second image, so that each frame of second image only retains the key region, so that the recognition is more accurate.

[0121] Next, step S400 is performed, and according to the first shooting action stage recognition result and the second shooting action stage recognition result, the target shooting action stage recognition result corresponding to the plurality of frames of first image is obtained.

[0122] In the implementation of the present application, the target archery action stage recognition result includes a target archery action stage, a start time of the target archery action stage, a duration of the target archery action stage, and check stability information corresponding to the target archery action stage.

[0123] In the implementation of the present application, as shown in Figure 4 According to the first archery action stage recognition result and the second archery action stage recognition result, the target archery action stage recognition result corresponding to the plurality of first images is obtained, including the following steps.

[0124] S410, in the case where it is determined that the first target frame image and the second target frame image exist the same frame, and the first archery action stage corresponding to the first target frame image and the second archery action stage corresponding to the second target frame image are the same archery action stage, the first weight coefficient corresponding to the first confidence indication information and the second weight coefficient corresponding to the second confidence indication information are determined according to the second confidence indication information.

[0125] For example, the first target frame image and the second target frame image in each group of the first archery action stage recognition result and each group of the second target frame image are compared. If it is determined that the first target frame image and the second target frame image exist the same frame, it indicates that the recognition of the same first image in the plurality of first images is performed. If it is determined that the first archery action stage corresponding to the first target frame image and the second archery action stage corresponding to the second target frame image are the same archery action stage, it indicates that the first archery action stage corresponding to the same first image and the second archery action stage are the same.

[0126] If the second confidence indication information is greater than a preset first confidence threshold, the target frame image is the second target frame image, the first weight coefficient is less than the second weight coefficient, and the second confidence indication information is less than a preset second confidence threshold, the target frame image is the first target frame image, and the first weight coefficient is greater than the second weight coefficient. .

[0127] In the implementation of the present application, if the second confidence indication information corresponding to the second archery action stage recognized by the image classification model is greater than a preset first confidence threshold (for example, 0.85), it indicates that the image classification confidence is high, and the image classification is mainly used, so that .

[0128] If the second confidence indication information is less than a preset second confidence threshold (for example, 0.4) or the image classification model output is null, the key point recognition result (the first archery action recognition result) is mainly used, so that .

[0129] ​​​The preset second confidence threshold is less than the preset first confidence threshold.

[0130] Further, if the second confidence indication information is greater than or equal to the preset second confidence threshold and less than or equal to the preset first confidence threshold, the target frame image is the first target frame image if the first confidence indication information is greater than or equal to the second confidence indication information, and the target frame image is the second target frame image if the first confidence indication information is less than the second confidence indication information.

[0131] It should be noted that, in the implementation manner of the present application, the first confidence indication information and the second confidence indication information can be obtained by .

[0132] Further, the first confidence indication information and the second confidence indication information can also be obtained by .

[0133] Further, in the implementation manner of the present application, the first weight coefficient and the second weight coefficient can also be determined according to the first confidence indication information, wherein if the first confidence indication information is greater than the preset first confidence threshold, the target frame image is the first target frame image, the first weight coefficient is greater than the second weight coefficient, and if the first confidence indication information is less than the preset second confidence threshold, the target frame image is the second target frame image, the first weight coefficient is less than the second weight coefficient.

[0134] Further, the first weight coefficient and the second weight coefficient can also be determined according to the first confidence indication information and the second confidence indication information, wherein if the first confidence indication information is greater than the second confidence indication information, the target frame image is the first target frame image, the first weight coefficient is greater than the second weight coefficient, and if the first confidence indication information is less than the second confidence indication information, the target frame image is the second target frame image, the first weight coefficient is less than the second weight coefficient.

[0135] S420, obtaining target confidence indication information corresponding to the target frame image according to the first confidence indication information, the second confidence indication information, the first weight coefficient and the second weight coefficient.

[0136] For example, for each group of target frame images, target confidence indication information corresponding to the target frame image is obtained according to the first confidence indication information, the second confidence indication information, the first weight coefficient and the second weight coefficient.

[0137] Taking the first confidence indication information and the second confidence indication information as scores as an example, the target confidence indication information of the target frame image is obtained by the following way:

[0138] ​​​​

[0139] wherein, is target confidence indication information of the target frame image, is a second weight coefficient, is a first judgment score (as an example of the second confidence indication information), is a first weight coefficient, is a second judgment score (as an example of the first confidence indication information), , is a [0, 1] adjustable fusion coefficient.

[0140] When , the target confidence indication information of the target frame image is obtained by the following way:

[0141]

[0142] In this way, the target confidence indication information of the target frame image is obtained.

[0143] S430, in the case that the target confidence indication information is greater than a preset third confidence threshold, determining that the first shooting action stage or the second shooting action stage is the target shooting action stage.

[0144] For example, if the target confidence indication information is greater than a preset third confidence threshold (for example, 0.8), the output target shooting action stage is the first shooting action stage or the second shooting action stage.

[0145] S440, obtaining the start time and the duration of the target shooting action stage according to the timestamp information corresponding to the second target frame image.

[0146] For example, the timestamp information of the target frame image is recorded at the same time, the timestamp information with the smallest time in the multiple frame images included in the target frame image is taken as the start time of the target shooting action stage, and the duration is obtained according to the timestamp information with the largest time and the timestamp information with the smallest time in the multiple frame images included in the target frame image. In this way, the target shooting action stage-time mapping table is obtained.

[0147] S450, determining the detection stability information of the target shooting action stage according to the first confidence indication information and the second confidence indication information corresponding to the target frame image.

[0148] For example, the detection stability information corresponding to the target frame image is obtained by weighted average of the first confidence indication information corresponding to the first target frame image and the second confidence indication information corresponding to the second target frame image, that is, the detection stability information of the target shooting action stage. For example, the detection stability score.

[0149] Further, in a case where it is determined that the first target frame image and the second target frame image exist the same frame, and the first shooting motion phase corresponding to the first target frame image and the second shooting motion phase corresponding to the second target frame image are non-identical shooting motion phases, it is determined that the target shooting motion phase corresponding to the target frame image is the other phase.

[0150] In a case where it is determined that the target confidence degree indication information is less than the third preset confidence degree threshold, second abnormal information is obtained.

[0151] If the first abnormal information and the second abnormal information are obtained, it is determined that the target shooting motion phase corresponding to the target frame image is the other phase.

[0152] The recognized shooting motion phase (for example, drawing a bow / accumulating force / releasing / other), the shooting motion phase switching time point, the shooting motion phase duration, and the detection stability score are output.

[0153] In this way, one or more target shooting motion phases in the plurality of frames of first images, and the switching time (that is, the start time), the duration, and the detection stability information of the target shooting motion phase are obtained.

[0154] The shooting motion recognition method provided by the implementation manner of the present application, that is, the somatosensory shooting motion recognition method based on the key point time sequence change of MediaPipe detection and the self-training AI classification model (that is, the intelligent image classification model), for example, is a shooting motion phase recognition method based on the combination of hand key point time sequence change and image classification, which can accurately recognize various key motions of a user in a shooting process, and can provide real-time feedback on shooting motion information of each shooting motion phase, thereby improving the shooting interactive experience. In addition, the user can adjust the shooting motion of each shooting motion phase according to the shooting motion phase recognition result, so that the shooting result is more accurate.

[0155] The shooting motion recognition method provided by the implementation manner of the present application can be applied to an electronic device.

[0156] Figure 5 As shown in the figure, the electronic device provided by the embodiment of the present application includes a transceiver 121, a processor 122, and a memory 123. Figure 5 As shown in the figure, the electronic device can include a transceiver 121, a processor 122, and a memory 123.

[0157] The processor 122 executes computer-executed instructions stored in the memory to cause the processor 122 to perform the technical solutions of the arrow shooting action recognition method in the above embodiments. The processor 122 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; and can also be a digital processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0158] The memory 123 is connected with the processor 122 through a system bus and completes mutual communication, and the memory 123 is used for storing computer program instructions.

[0159] By way of example, and without limitation, the memory 123 can include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. Where appropriate, the memory 123 can include removable or non-removable (or fixed) media. Where appropriate, the memory 123 can be internal or external to the integrated gateway device. In particular embodiments, the memory 123 is non-volatile solid-state memory. In particular embodiments, the memory 123 includes read-only memory (ROM). Where appropriate, this ROM can be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), electrically alterable ROM (EAROM) or flash memory, or a combination of two or more of these. The transceiver 121 can be used to obtain a to-be-executed task and configuration information of the to-be-executed task.

[0160] The system bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus. The transceiver is used to realize the communication between the database access device and other computers (such as clients, read-write libraries and read-only libraries). The memory can include random access memory (RAM) and can also include non-volatile memory.

[0161] Further, the electronic device may, for example, be a computer, a mobile phone, a server, or the like.

[0162] The embodiments of the present application also provide a chip for running instructions, which is used to execute the technical solutions of the archery action recognition method in the above embodiments.

[0163] The embodiments of the present application also provide a computer readable storage medium, which stores computer instructions, and when the computer instructions run on a processor of a computer device, the processor of the computer device executes the technical solutions of the archery action recognition method in the above embodiments.

[0164] In some possible implementation manners, various aspects of the method provided by the present application can also be implemented in the form of a program product, which includes program codes for causing the processor of the computer device to execute the steps in the method according to various exemplary implementation manners of the present application described above in the specification when the program product runs on the processor of the computer device, for example, the computer device can execute the archery action recognition method described in the embodiments of the present application.

[0165] The program product can adopt any combination of one or more readable media. The readable medium can be a readable data medium or a readable storage medium. The readable storage medium may, for example, be but is not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CDROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0166] The implementation of the present application also provides a computer program product, which comprises a computer program stored in a computer readable storage medium, at least one processor can read the computer program from the computer readable storage medium, and the at least one processor can implement the technical solutions of the arrow shooting action recognition method in the above embodiments when executing the computer program.

[0167] It should be noted that in addition to the implementation of the present application described in the above specific embodiments, other advantages and effects of the present application can be easily understood by those skilled in the art from the content disclosed in the present application. Although the description of the present application is introduced in combination with the preferred embodiments, it does not mean that the features of the present application are limited to the implementation. On the contrary, the purpose of introducing the present application in combination with the implementation is to cover other options or modifications that can be extended from the present application. In order to provide a deep understanding of the present application, many specific details are included in the above description, and the present application can also be implemented without using these details. In addition, in order to avoid confusion or obscure the focus of the present application, some specific details will be omitted in the description. It should be noted that the embodiments and features in the embodiments in the present application can be combined with each other without conflict.

[0168] It should be noted that in the present specification, similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0169] It should be noted that the terms "first", "second" and the like are only used to distinguish description, and cannot be understood as indicating or implying relative importance.

[0170] It should be noted that in the drawings, some structural or method features can be shown in a specific arrangement and / or order. However, it should be understood that such specific arrangement and / or order can not be required. Rather, in some embodiments, the features can be arranged in a manner different from that shown in the illustrative figures. Additionally, the inclusion of a structural or method feature in a particular figure does not imply that the feature is required in all embodiments, and in some embodiments, the feature can not be included or can be combined with other features in a manner not shown in a particular figure.

[0171] Although the present application has been illustrated and described with reference to certain preferred implementations, it should be understood by those skilled in the art that the above is a further detailed description of the present application in combination with specific implementations, and the specific implementation of the present application cannot be limited to these descriptions. Those skilled in the art can make various changes in form and details, including making several simple deductions or substitutions, without departing from the spirit and scope of the present application.

Claims

1. A method for recognizing archery actions, characterized in that, The method includes: Determine the first images of multiple frames related to the archery action during the user's archery process; Based on each frame of the first image, key point location information of multiple key points in each frame of the first image is determined. Then, based on the key point location information of each key point corresponding to each frame of the first image, a first archery action recognition process is performed to obtain the first archery action stage recognition result corresponding to the multiple frames of the first image; and Based on the image classification model, the second archery action recognition process is performed on the multiple frames of the first image to obtain the recognition result of the second archery action stage corresponding to the multiple frames of the first image; Based on the first archery action phase recognition result and the second archery action phase recognition result, the target archery action phase recognition result corresponding to the multi-frame first image is obtained.

2. The archery action recognition method according to claim 1, characterized in that, Based on the key point position information of each key point corresponding to each frame of the first image, a first archery action recognition process is performed to obtain the first archery action stage recognition result corresponding to the multiple frames of the first image, including: Based on the key point location information of each key point corresponding to each frame of the first image, the distance information between each key point corresponding to each frame of the first image is determined, and the distance change trend of the key points corresponding to the multiple frames of the first image is obtained based on the distance information. Based on the distance change trend of the key points corresponding to the multi-frame first images, the first archery action stage corresponding to the first target frame image in the multi-frame first images and the first confidence indication information corresponding to the first archery action stage are obtained. Based on the first archery action stage corresponding to the first target frame image and the first confidence indication information corresponding to the first archery action stage, the recognition result of the first archery action stage corresponding to the multiple first frames is obtained.

3. The archery action recognition method according to claim 2, characterized in that, Based on the distance change trend of the key points corresponding to the multiple first images, the first archery action stage corresponding to the first target frame image in the multiple first images is obtained, including: If, based on the distance change trend of the key points corresponding to the multi-frame first images, it is determined that the distance change of the key points corresponding to the first target frame image in the multi-frame first images satisfies the first change condition, then the first archery action stage corresponding to the first target frame image is determined to be the bow-drawing stage. If, based on the key point change trend of the key point corresponding to the first target frame image in the multi-frame first images, it is determined that the distance change of the key point in the first target frame image in the multi-frame first images satisfies the second change condition, then the first archery action stage corresponding to the first target frame image is determined to be the charging stage. If, based on the key point change trend of the key point corresponding to the first target frame image in the multi-frame first images, it is determined that the distance change of the key point in the first target frame image in the multi-frame first images satisfies the third change condition, then the first archery action stage corresponding to the first target frame image is determined to be the release stage.

4. The archery action recognition method according to claim 3, characterized in that, Based on the image classification model, the second archery action recognition process is performed on the multiple frames of the first image to obtain the recognition result of the second archery action stage corresponding to the multiple frames of the first image, including: Based on the image classification model, the second archery action recognition process is performed on the multiple frames of the first image to obtain the archery action category and archery action category confidence indication information for each frame of the first image; Based on the archery action category and archery action category confidence indication information of the first image in each frame, the second archery action stage corresponding to the second target frame image in the multi-frame first image and the second confidence indication information corresponding to the second archery action stage are obtained. Based on the second archery action stage corresponding to the second target frame image and the second confidence level indication information corresponding to the second archery action stage, the recognition result of the second archery action stage corresponding to the multiple first images is obtained.

5. The archery action recognition method according to claim 4, characterized in that, The target archery action phase recognition result includes the target archery action phase corresponding to the target frame image. Based on the first archery action phase recognition result and the second archery action phase recognition result, the target archery action phase recognition result corresponding to the multiple first frames of images is obtained, including: If it is determined that the first target frame image and the second target frame image have the same frame, and the first archery action phase corresponding to the first target frame image and the second archery action phase corresponding to the second target frame image are the same archery action phase, then according to the second confidence level indication information, the first weight coefficient corresponding to the first confidence level indication information and the second weight coefficient corresponding to the second confidence level indication information are determined. If the second confidence level indication information is greater than the preset first confidence level threshold, then the first weight coefficient is less than the second weight coefficient, and the target frame image is the second target frame image. If the second confidence level indication information is less than the preset second confidence level threshold, then the first weight coefficient is greater than the second weight coefficient, and the target frame image is the first target frame image. Based on the first confidence level indication information, the second confidence level indication information, the first weight coefficient, and the second weight coefficient, the target confidence level indication information corresponding to the target frame image is obtained; If the target confidence indication information is determined to be greater than a preset third confidence threshold, the first archery action phase or the second archery action phase is determined to be the target archery action phase corresponding to the target frame image.

6. The archery action recognition method according to claim 5, characterized in that, The multiple first images have corresponding timestamp information. The target archery action phase recognition result also includes the start time, duration, and stability information corresponding to the target archery action phase. Based on the first and second archery action phase recognition results, the target archery action phase recognition result corresponding to the multiple first images is obtained, including: The start time and duration of the target archery action phase are obtained based on the timestamp information corresponding to the target frame image; The detection stability information of the target archery action phase is determined based on the first confidence indication information and the second confidence indication information corresponding to the target frame image.

7. The archery action recognition method according to claim 6, characterized in that, Determine multiple first images related to the archery action during the user's archery process, including: Acquire video of the archery actions related to the user's archery movements during the archery process, and obtain the multi-frame first image based on the archery action video; or Multiple consecutive images related to the archery action during the user's archery process are acquired to obtain the first multi-frame image.

8. The archery action recognition method according to claim 7, characterized in that, Based on each frame of the first image, determine the key point location information of multiple key points in each frame of the first image, including: Each frame of the first image is input into the keypoint detection model so that the keypoint detection model determines the keypoint location information of multiple keypoints in each frame of the first image, including hand keypoints.

9. The archery action recognition method according to any one of claims 1-8, characterized in that, Based on an image classification model, the second archery action recognition process is performed according to the multiple frames of the first image, including: The target region in the multiple first images is cropped to obtain multiple second images. These multiple second images are then input into the image classification model, enabling the model to perform a second archery action recognition based on the multiple second images. The target region includes at least one of a hand region, a face region, and a bow and arrow equipment region; or The first multi-frame image is input into the image classification model so that the image classification model can perform a second archery action recognition process based on the first multi-frame image.

10. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes the computer execution instructions stored in the memory to perform the archery action recognition method as described in any one of claims 1-9.

Images