Shooting training method and device combining motion capture and image recognition

Abstract

The invention relates to a shooting training method and device combining motion capture and image recognition. The method comprises the steps that motion capture equipment is used for collecting displacement tracks of different identification points in the shooting process; generating displacement tracks of different limb classifications according to the identification point set and the displacement track of each identification point; acquiring a basketball key frame in a shooting process and acquiring initial shooting data in the shooting process by using image acquisition equipment; performing data screening and classification on the initial shooting data based on the hit basket position; and analyzing the correlation degree of the hit basket position in the shooting data set and the displacement tracks of different limb classifications. According to the method, the action capture equipment and the image acquisition equipment are combined to respectively obtain different limb action tracks and shooting data of the to-be-trained person in the shooting process in real time and carry out association analysis, so that the influence degree of different limbs of the to-be-trained person on a hit result in the shooting process can be accurately analyzed and obtained; and more targeted training and improvement are provided for weak points of limbs.

Description

Technical Field

[0001] The present application relates to the technical field of sports training data processing, and in particular to a shooting training method and device combining motion capture and image recognition. Background Art

[0002] In the field of sports training, accurate technical analysis and personalized guidance are the key to improving athletes' performance. With the rapid development of science and technology, motion capture and image recognition technologies have gradually emerged, bringing new changes to sports training.

[0003] Motion capture technology records the movement trajectory and posture information of objects or human bodies in real time by placing sensors on key parts of objects or human bodies. In recent years, with the continuous advancement of sensor technology, motion capture devices based on inertial measurement units have emerged. Such devices are small in size and light in weight, and can be worn directly on athletes, achieving freer and more convenient sports data collection.

[0004] Image recognition technology uses computer vision algorithms to analyze and understand objects and scenes in images or videos. In the field of sports, image recognition technology can monitor and analyze athletes' movements in real time, such as identifying athletes' postures and the quality of their movements. With the rise of deep learning algorithms, especially the widespread application of convolutional neural networks, the accuracy of image recognition has been greatly improved, making its application in sports training possible.

[0005] Most of the existing shooting training methods rely on a single technology, either focusing on motion capture analysis or image recognition evaluation, and fail to give full play to the advantages of both. For example, although simple motion capture can accurately record the athlete's limb movement data, it does not obtain enough information about the spatial relationship between the basketball and the basket, environmental factors during shooting, etc.; and simple image recognition may have recognition errors under complex lighting conditions, and it is difficult to fully analyze the athlete's body mechanics parameters. Therefore, it is necessary to study a shooting training method that combines motion capture and image recognition to give full play to the advantages of the two technologies, realize the correlation analysis between the athlete's limb movements and the hit results, and then accurately locate the impact of different limbs of the athlete on the hit results, realize the precise positioning and training of the disadvantaged limbs, and comprehensively improve the overall physical fitness of the athlete. Summary of the invention

[0006] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a shooting training method and device combining motion capture and image recognition, which specifically adopts the following technical solutions: A shooting training method combining motion capture and image recognition, the method comprising the following steps: Placing identification points on key parts of the trainee, and marking different identification points according to limb classification to form a set of identification points corresponding to the limb classification; the limb classification includes upper limbs, torso and lower limbs; Use motion capture equipment to collect the displacement trajectory of different marking points during each shooting process of the person to be trained; Generate displacement trajectories of different limb classifications according to the set of identification points and the displacement trajectories of each identification point; Using an image acquisition device to obtain a basketball key frame of each shooting process of the person to be trained, and obtaining initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes a shooting position, a basketball motion trajectory, and a position that hits the basket; wherein the position that hits the basket at least includes a hit effective area, a miss effective area, and a hit basket; The initial shooting data is screened based on the position of the basket being hit, and the valid shooting data of the effective area and the basket being hit are retained; Classify the historical effective shooting data of the person to be trained based on the shooting position, and obtain a set of shooting data at the same shooting position; Analyze the correlation between the position of the basket hit and the displacement trajectory of different limb classifications in the shooting data set, and screen out the high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree; Combine shooting positions and high-correlation limb classifications to conduct shooting training for the trainees.

[0007] Optionally, the step of marking different identification points according to limb classification to form a set of identification points corresponding to the limb classification includes: Get each marker point i , and based on the mapping function, each identification point i The ID is mapped to the corresponding limb category C ,in Represents a single identification point; They represent the classification of limbs into upper limbs, trunk, and lower limbs; Gather the identification points corresponding to the same limb classification to obtain the corresponding limb classification C The set of identification points ;in .

[0008] Optional: The motion capture device is used to collect the displacement trajectory of different identification points during each shooting process of the person to be trained: Use motion capture equipment to synchronously collect different identification points during each shot of the trainee i location data; According to the time series, each mark point i At different collection times TThe obtained position data are arranged to obtain the corresponding identification points of the person to be trained in the current shooting process i A collection of location points ;in , t is the total duration of the collection time; A collection of location points for a single marker Perform trajectory fitting to obtain corresponding identification points i The displacement trajectory .

[0009] Optionally, the step of generating displacement trajectories of different limb classifications according to the set of identification points and the displacement trajectories of each identification point includes: Get limb classification C The set of identification points and the different identification points in the corresponding identification point set i The displacement trajectory ; Limb classification is obtained by weighted average C The aggregate displacement trajectory of: ; in Classify the limbs C The aggregate displacement trajectory of For marking point i The weight of .

[0010] Optional: When using an image acquisition device to obtain basketball key frames of each shooting process of the person to be trained, the basketball key frames acquired by the image acquisition device include at least a first basketball key frame when the basketball is out of the hand and a second basketball key frame when the basketball first touches the backboard.

[0011] Optionally, the step of analyzing the correlation between the position of hitting the basket and the displacement trajectories of different limb classifications in the shooting data set includes: Combined with limb classification based on the collection time of the first basketball key frame C The aggregate displacement trajectory is used to obtain the limb classification when the basketball is released C location data; Based on the classification of limbs when basketball is released C The position data of the limb category C at the current moment is obtained; Obtaining the straight-line distance between the basketball hitting position and the center of the basket based on the second basketball key frame; Calculate the limb classification for the shooting data of different limb classifications in the shooting data set C The Pearson correlation coefficient of the instantaneous height and straight-line distance was calculated and a significance test was performed; Obtaining limb classification in the shooting data set based on Pearson correlation coefficient results C The degree of relevance to the location of the basket hit.

[0012] Optional: When limbs are classified C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 1, it indicates that the limb classification C The displacement trajectory is strongly positively correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 0, it indicates that the limb classification C The displacement trajectory is weakly correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to -1, it indicates that the limb classification C The displacement trajectory is strongly negatively correlated with the position of the basket.

[0013] Optional: Computational limb classification C The formula for the Pearson correlation coefficient of the instantaneous height and straight-line distance is: ; in Classify the limbs C Pearson correlation coefficient between instantaneous height and straight-line distance; N is the number of shooting data in the shooting data set; is the first j The corresponding limb classification in the shooting data C The instantaneous height of is the first j The straight-line distance between the basketball's hitting position and the center of the basket in each shot data; Classify limbs in the shooting data set C The instantaneous height mean of It is the mean straight-line distance between the basketball hitting position and the center of the basket in the shooting data set.

[0014] Optional: When dividing the hit valid area and the miss valid area, the backboard area wireframe is used as the boundary, and the backboard area wireframe and the inner area are divided into the hit valid area; the area outside the backboard area wireframe is divided into the miss valid area.

[0015] In addition, the present application also discloses a shooting training device combining motion capture and image recognition, the device comprising: A marking point setting module is used to place marking points on key parts of the trainee and mark different marking points according to limb classification to form a marking point set corresponding to the limb classification; the limb classification includes upper limbs, torso and lower limbs; A marking point acquisition module is used to use a motion capture device to collect the displacement trajectory of different marking points during each shooting process of the trainee; A limb trajectory acquisition module is used to generate displacement trajectories of different limb classifications based on a set of identification points and the displacement trajectories of each identification point; A shooting data acquisition module is used to use an image acquisition device to obtain a basketball key frame of each shooting process of the trainee, and obtain initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes a shooting position, a basketball motion trajectory, and a position that hits the basket; wherein the position that hits the basket at least includes a hit effective area, a miss effective area, and a hit basket; A data screening module, used for screening the initial shooting data based on the position of the basket hit, and retaining the valid shooting data of the effective area hit and the basket hit; A data classification module, used to classify the historical effective shooting data of the trainee based on the shooting position, and obtain a set of shooting data of the same shooting position; A correlation analysis module is used to analyze the correlation between the position of the basket hit and the displacement trajectory of different limb classifications in the shooting data set, and screen out the high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree; The result output module is used to combine the shooting position and the high-correlation limb classification to conduct shooting training for the trainees.

[0016] Beneficial Effects The technical solution of this application has the following beneficial effects: (1) The shooting training method of the present application uses a motion capture device and an image acquisition device to respectively obtain in real time the different limb movement trajectories and shooting data of the trainee during the shooting process, and performs correlation analysis on the limb movement trajectories and the shooting data. This method can accurately analyze and obtain the degree of influence of different limbs of the trainee on the hit result during the shooting process, and then obtain the degree of participation of each limb in the shooting stage. This method can help athletes understand the accuracy of their own shooting movements, and obtain their own strong and weak limbs in the shooting stage, and then provide more targeted training and improvement for the weak points of the limbs. This solves the problem that traditional shooting training often relies on the coach's naked eye observation and judgment, is easily affected by subjective factors, and is difficult to accurately evaluate the accuracy and weak points of the user's shooting movements.

[0017] (2) The shooting training method of the present application can accurately obtain the details of limb movement by placing identification points on the key parts of the trainees and combining motion capture equipment to collect displacement trajectories, which can be subdivided into upper limbs, torso and lower limbs, so as to fully understand the movement of each part during the shooting process; at the same time, the image acquisition equipment is used to obtain basketball key frames and shooting data, and data is collected from both limb movements and basketball movement trajectories, making the analysis more comprehensive and accurate, and providing a solid data foundation for subsequent training.

[0018] (3) The shooting training method of the present application screens the initial shooting data based on the position of the basket hit, retains the valid shooting data, removes invalid interference information, and focuses on truly valuable data; classifies the historical valid shooting data based on the shooting position, analyzes the correlation between the position of the basket hit and the limb displacement trajectory, and screens out high and low correlation limb classifications, so as to combine the shooting position and the high correlation limb classification for training, so as to achieve efficient training and greatly improve the training efficiency and pertinence.

[0019] (4) The shooting training method of the present application takes into account the performance of different trainees at different shooting positions and the differences in the influence of various limb parts on the hit results. It can tailor personalized training plans for them based on their individual shooting data to meet the specific needs of different trainees. Compared with traditional general training methods, it can better tap the potential of each trainee, achieve targeted training for different trainees, and improve training effects and training adaptability.

[0020] (5) The shooting training method of the present application can collect and analyze data in real time during the training process, and provide timely feedback to the trainers, so that they can understand their own movements and shooting effects, and adjust the training strategies and movements in time according to the feedback, so as to achieve continuous optimization of the training effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Flowchart of the shooting training method in the embodiment of the present application.

[0022] Figure 2 2 is a structural diagram of a basketball shooting training device in an embodiment of the present application.

[0023] Figure 3 This is a structural diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The present application is further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and cannot be used to limit the scope of protection of the present application. It should be noted that the following detailed descriptions are all exemplary and are intended to provide further explanation of the present application.

[0025] With the popularity of basketball and the continuous improvement of competitive level, shooting, as the most critical scoring method in basketball games, has attracted much attention for its scientific and effective training methods. Traditional shooting training mainly relies on the coach's naked eye observation and experience judgment, which has many limitations. For example, the coach's subjective factors will have a great impact on the evaluation results, making it difficult to accurately judge the subtle deviations of the athlete's shooting movements, and it is also impossible to accurately analyze the specific degree of influence of different limbs on the hit results during the shooting process. This makes it difficult for athletes to improve their weak links in a targeted manner during training, and the training effect is greatly reduced. In order to solve the above problems, the embodiments of the present application combine the motion capture device and the image acquisition device to respectively obtain the different limb movement trajectories and shooting data of the trainee during the shooting process in real time, and perform correlation analysis on the limb movement trajectories and the shooting data, so as to accurately analyze and obtain the degree of influence of different limbs of the trainee on the hit result during the shooting process, so as to help athletes understand the accuracy of their own shooting movements, and obtain their own strong and weak limbs in the shooting stage, and then provide more targeted training and improvement for the weak points of the limbs, thereby solving the problem that traditional shooting training often relies on the coach's naked eye observation and judgment, which is easily affected by subjective factors and difficult to accurately evaluate the user's shooting movement accuracy and weaknesses.

[0026] Combination Figure 1 As shown, the embodiment of the present application discloses a shooting training method combining motion capture and image recognition, and the method comprises the following steps: Step S1: Place identification points on key parts of the trainee, and mark different identification points according to limb classification to form a set of identification points corresponding to the limb classification; the limb classification in this embodiment includes upper limbs, torso and lower limbs.

[0027] The selection of the above key parts is based on in-depth research and analysis of shooting movements, covering the upper limbs, torso and lower limbs, which play an important role in the shooting process. For example, in the upper limbs, markers will be placed at the wrists, elbows, shoulders and other joints, because the movements of these parts directly affect the shooting angle, strength and stability; in the torso, markers will be set at the chest, waist and other positions, and the rotation and stability of the torso play a key supporting role in the shooting force; in the lower limbs, markers will be placed at the knees, ankles and other parts, and the push-off action of the lower limbs is an important source of shooting power.

[0028] After placing the identification points in the above step S1, different identification points are marked according to the limb classification, so as to form a set of identification points corresponding to the limb classification. Specifically, in this embodiment, the steps of marking different identification points according to the limb classification to form a set of identification points corresponding to the limb classification include the following: Get each marker point i, and based on the mapping function, each identification point i The ID is mapped to the corresponding limb category C ,in Represents a single identification point; Respectively represent the limb classification into upper limbs, trunk and lower limbs; In this embodiment, it is assumed that the ID of each identification point is a unique number in a series of sequences, and the limb classification , where a represents the upper limbs, b represents the trunk, and c represents the lower limbs. Through the pre-set mapping rules, each identification point is accurately classified into the corresponding limb classification.

[0029] Then the identification points corresponding to the same limb classification are collected to obtain the corresponding limb classification C The set of identification points ;in For example, in this embodiment, for the upper limb classification, all the identification points mapped to a are grouped together to form an upper limb identification point set, which is recorded as ;in Indicates the various identification points belonging to the upper limb classification. The identification points corresponding to different limb classifications are different and unique.

[0030] Step S2: Use a motion capture device to collect the displacement trajectory of different identification points during each shooting process of the person to be trained.

[0031] Specifically, in step S2, the motion capture device is used to synchronously collect different identification points of the trainee during each shooting process. i In this embodiment, the motion capture device can use advanced optical or inertial measurement technology to track the movement of the marker points with high precision. During the collection process, the motion capture device is used to synchronously collect the different marker points of the trainee during each shot. i The motion capture device collects data at a very high frequency to ensure that every subtle change in the movement of the marker during the shooting process can be captured. Assuming the collection frequency is 100 times per second, this means that a large amount of marker position information can be obtained during a shooting process.

[0032] Then, each mark point is divided into i At different collection times T The obtained position data are arranged to obtain the corresponding identification points of the person to be trained in the current shooting process i A collection of location points ;in , tis the total duration of the collection time; for example, in a shooting process, from the start of the shooting action to the release of the basketball, it lasts for 1 second, and the collection frequency is 100 times per second, so there will be 100 position data at different times. Arrange these data in chronological order, and then obtain the corresponding identification points of the trainee in the current shooting process. i A collection of location points .

[0033] Finally, the location point set of a single identification point Perform trajectory fitting to obtain corresponding identification points i The displacement trajectory In this embodiment, the trajectory fitting can adopt a professional line fitting algorithm, such as a spline curve fitting algorithm. Through the line fitting algorithm, a smooth and continuous curve can be generated according to the discrete position point data. This curve is the corresponding mark point i By collecting and generating the displacement trajectory of each identification point, the movement trajectory of each key part of the person to be trained during the shooting process can be understood in detail.

[0034] Step S3: Generate displacement trajectories of different limb classifications based on the set of identification points and the displacement trajectories of each identification point. The specific operation process is as follows: First get the limb classification C The set of identification points and the different identification points in the corresponding identification point set i The displacement trajectory ; Taking the limb classification as upper limb as an example, obtain the upper limb identification point set And collection The displacement trajectory of each marked point in .

[0035] Then the limb classification is obtained by weighted average C The aggregate displacement trajectory of: ; in Classify the limbs C The aggregate displacement trajectory, expressed in time t The overall displacement of limb classification C; For marking point i The weights are set based on the evaluation of the importance of different landmarks in the shooting action. For example, the landmarks on the wrist play a key role in controlling the shooting angle, and their weights may be relatively high; while the landmarks on other auxiliary parts of the arm have relatively low weights. Represents the displacement data of the marker point i at time t. Through this weighted average method, the displacement of different marker points in the same limb classification can be comprehensively considered to generate an aggregated displacement trajectory that better represents the overall motion characteristics of the limb classification.

[0036] Step S4: Use an image acquisition device to obtain a basketball key frame of each shooting process of the trainee, and obtain the initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes the shooting position, the basketball movement trajectory, and the position of hitting the basket. The position of hitting the basket at least includes hitting the effective area, not hitting the effective area, and hitting the basket. In this embodiment, the image acquisition device preferably uses a high-definition camera and is installed in a position that can fully capture the shooting process, such as the side and front of the basketball stand.

[0037] It should be noted that in step S4, when the image acquisition device is used to acquire the basketball key frames of each shooting process of the trainee, the basketball key frames acquired by the image acquisition device include at least the first basketball key frame when the basketball is released and the second basketball key frame when the basketball first touches the backboard. These two key frames are crucial for analyzing the shooting process. The key frame when the basketball is released can obtain the initial state information of the shot, such as the shooting position, shooting speed, shooting angle, etc.; the key frame when the basketball first touches the backboard can reflect the trajectory changes of the basketball during the flight process and the relative position relationship with the basket.

[0038] It should be noted that the initial shooting data obtained according to the basketball key frame in this embodiment at least includes the shooting position, the basketball movement trajectory and the position of hitting the basket. The shooting position can be determined by image recognition technology combined with the coordinate system of the venue. For example, a coordinate system is established in the horizontal and vertical directions with the basketball stand as the origin. The specific position coordinates of the shot are calculated according to the position of the basketball in the image when it is released and the parameters of the camera. The basketball movement trajectory is obtained by tracking and analyzing the position of the basketball in multiple basketball key frames. Using the image processing algorithm, the movement curve of the basketball in space can be fitted. The position of hitting the basket at least includes a hit effective area, a miss effective area and a hit basket. When dividing the hit effective area and the miss effective area in this embodiment, the backboard hitting area wireframe is used as the boundary, and the backboard hitting area wireframe and the internal area are divided into the hit effective area; the area outside the backboard hitting area wireframe is divided into the miss effective area. This division method is based on the actual scoring situation and shooting skills in basketball games. The backboard area is an area often used when shooting. Hitting this area usually means that the shot has high accuracy and rationality. Of course, the effective hit area can be re-divided according to actual needs.

[0039] Step S5: Screen the initial shooting data based on the position of the basket hit, and retain the valid shooting data that hits the effective area and the basket hit. In a large amount of shooting data, there are some invalid data, such as the situation where the shot misses and deviates far from the basket. These data are not very meaningful for analyzing the effectiveness of the shooting action. Through screening, these invalid interference information can be removed, focusing on truly valuable data, and improving the accuracy and efficiency of subsequent analysis.

[0040] Step S6: Classify the historical valid shooting data of the trainee based on the shooting position to obtain a set of shooting data at the same shooting position. For example, all valid shooting data at the same position outside the three-point line are classified into one category, and the valid shooting data at the free throw line position are classified into another category. The purpose of such classification is to analyze the differences in the shooting actions and hit situations of the trainee at different shooting positions, so as to provide more targeted training guidance.

[0041] Step S7: Analyze the correlation between the position of hitting the basket and the displacement trajectories of different limb classifications in the shooting data set, and screen out high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree.

[0042] Specifically, the step of analyzing the correlation between the position of hitting the basket and the displacement trajectories of different limb classifications in the shooting data set in step S7 includes the following steps: First, according to the collection time of the first basketball key frame, combined with the limb classification C The aggregate displacement trajectory is used to obtain the limb classification when the basketball is released C For example, at the moment t when the basketball is released, by querying the corresponding limb classification C The aggregated displacement trajectory data can be used to obtain the specific position information of the upper limbs, trunk and lower limbs at time t.

[0043] Then based on the limb classification when the basketball is released C The instantaneous height of the limb category C at the current moment is obtained by using the position data of the body. Taking the body as an example, the vertical coordinate data of the body position data at time t can be obtained through the specific position information of the body obtained in the previous step, and then the height of the body when the basketball is released can be calculated.

[0044] Then, based on the second basketball key frame, the straight-line distance between the basketball's hitting position and the center of the basket is obtained; through image processing technology, in the key frame where the basketball first touches the backboard, the straight-line distance between the basketball and the center of the basket is calculated. This distance can reflect the accuracy of the basketball hitting the basket.

[0045] Secondly, the limb classification is calculated for the shooting data of different limb classifications in the shooting data set. CThe Pearson correlation coefficient between the instantaneous height and the straight-line distance of limb category C was calculated and a significance test was performed. By calculating the Pearson correlation coefficient, the correlation between the instantaneous height of limb category C and the straight-line distance between the basketball hitting position and the center of the basket can be quantified. A significance test was also performed to determine whether this correlation was statistically significant.

[0046] It should be noted that in this embodiment, the limb classification is calculated. C The formula for the Pearson correlation coefficient of the instantaneous height and straight-line distance is: ; in Classify the limbs C Pearson correlation coefficient between instantaneous height and straight-line distance; N is the number of shooting data in the shooting data set; is the first j The corresponding limb classification in the shooting data C The instantaneous height of is the first j The straight-line distance between the basketball's hitting position and the center of the basket in each shot data; Classify limbs in the shooting data set C The instantaneous height mean of It is the mean straight-line distance between the basketball hitting position and the center of the basket in the shooting data set.

[0047] Obtaining limb classification in the shooting data set based on Pearson correlation coefficient results C The degree of relevance to the location of the basket. C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 1, it indicates that the limb classification C The displacement trajectory is strongly positively correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 0, it indicates that the limb classification C The displacement trajectory is weakly correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to -1, it indicates that the limb classification C The displacement trajectory is strongly negatively correlated with the position of the basket. Through this correlation analysis, we can screen out the high-correlation limb classifications that have a greater impact on the hit results and the low-correlation limb classifications that have a smaller impact.

[0048] Step S8: Combine the shooting position and the highly correlated limb classification to conduct shooting training for the trainee. For example, at a certain shooting position, the analysis finds that the displacement trajectory of the upper limb is strongly positively correlated with the position of hitting the basket. Then, during training, the upper limb movements can be focused on. Special upper limb strength training can be arranged, such as using dumbbells to perform arm flexion and extension exercises to enhance the muscle strength of the arm to improve the force stability during shooting; detailed training of shooting movements can also be performed, such as adjusting the shooting angle, the wrist flicking movement, etc. Through repeated practice, the correct movements of the highly correlated limb classifications can be strengthened to improve the accuracy and hit rate of shooting.

[0049] Through the specific embodiments of the above-mentioned shooting training method, it can be seen that the present application has significant advantages. First, it uses motion capture equipment and image acquisition equipment to respectively obtain different limb motion trajectories and shooting data of the trainee in real time during the shooting process, and associates the limb motion trajectories with the shooting data, which can accurately analyze and obtain the degree of influence of different limbs of the trainee on the hit result during the shooting process, and then obtain the degree of participation of each limb in the shooting stage. This helps athletes understand the accuracy of their own shooting movements, clarify their own strong and weak limbs in the shooting stage, and thus provide more targeted training and improvement for limb weaknesses.

[0050] Secondly, by placing identification points on key parts of the trainees and collecting displacement trajectories with motion capture equipment, we can accurately obtain the details of limb movement, subdivided into upper limbs, torso and lower limbs, and fully understand the movement of each part during the shooting process; at the same time, we use image acquisition equipment to obtain basketball key frames and shooting data, and collect data from both limb movements and basketball movement trajectories, making the analysis more comprehensive and accurate, and providing a solid data foundation for subsequent training.

[0051] Furthermore, by screening the initial shooting data based on the position of the basket hit and retaining the valid shooting data, invalid interference information can be removed and the focus can be placed on truly valuable data. By classifying the historical valid shooting data based on the shooting position, the correlation between the position of the basket hit and the limb displacement trajectory is analyzed, and high and low correlation limb classifications are screened out. In this way, training can be carried out in combination with the shooting position and the high correlation limb classification, so as to achieve efficient training and greatly improve the efficiency and pertinence of training.

[0052] In addition, the present application takes into account the performance of different trainees in different shooting positions and the differences in the influence of various limb parts on the hit results. It can tailor personalized training plans for them based on their individual shooting data to meet the specific needs of different trainees. Compared with traditional general training methods, it can better tap the potential of each trainee, achieve targeted training for different trainees, and improve training effects and training adaptability.

[0053] Finally, data can be collected and analyzed in real time during the training process, and feedback can be provided to trainers in a timely manner so that they can understand their movements and shooting effects, and adjust training strategies and movements in a timely manner based on the feedback to achieve continuous optimization of training effects.

[0054] To sum up, the shooting training method and device combining motion capture and image recognition of the present application have important application value and promotion significance, can provide a scientific, efficient and personalized solution for basketball players' shooting training, and help improve basketball players' shooting technique level and competitive ability.

[0055] In addition, combined Figure 2 As shown, the present application also discloses a shooting training device combining motion capture and image recognition, the device comprising: A marking point setting module is used to place marking points on key parts of the trainee and mark different marking points according to limb classification to form a marking point set corresponding to the limb classification; the limb classification includes upper limbs, torso and lower limbs; A marking point acquisition module is used to use a motion capture device to collect the displacement trajectory of different marking points during each shooting process of the trainee; A limb trajectory acquisition module is used to generate displacement trajectories of different limb classifications based on a set of identification points and the displacement trajectories of each identification point; A shooting data acquisition module is used to use an image acquisition device to obtain a basketball key frame of each shooting process of the trainee, and obtain initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes a shooting position, a basketball motion trajectory, and a position that hits the basket; wherein the position that hits the basket at least includes a hit effective area, a miss effective area, and a hit basket; A data screening module, used for screening the initial shooting data based on the position of the basket hit, and retaining the valid shooting data of the effective area hit and the basket hit; A data classification module, used to classify the historical effective shooting data of the trainee based on the shooting position, and obtain a set of shooting data of the same shooting position; A correlation analysis module is used to analyze the correlation between the position of the basket hit and the displacement trajectory of different limb classifications in the shooting data set, and screen out the high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree; The result output module is used to combine the shooting position and the high-correlation limb classification to conduct shooting training for the trainees.

[0056] The device provided in the embodiment of the present application can achieve Figure 1 To avoid repetition, the various processes implemented by the method embodiment will not be described again here.

[0057] like Figure 3 As shown, the embodiment of the present application also provides an electronic device, including a processor and a memory, a program or instruction stored in the memory and executable on the processor, and when the program or instruction is executed by the processor, the following is achieved: Figure 1 The various processes of the method embodiment shown in the figure can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0058] The embodiment of the present application also provides a readable storage medium on which a program or instruction is stored. When the program or instruction is executed by a processor, the above Figure 1 The various processes of the method embodiments described above can achieve the same technical effect, and will not be described again here to avoid repetition.

[0059] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above Figure 1 The various processes of the method embodiments described above can achieve the same technical effect, and will not be described again here to avoid repetition.

[0060] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned sequence numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0061] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0062] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another device, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0063] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0064] In addition, all functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be a separate unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0065] A person skilled in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, etc., various media that can store program codes.

[0066] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application can essentially or in other words, the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a device (which can be a terminal or a platform, etc.) to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0067] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A shooting training method combining motion capture and image recognition, characterized in that: The method comprises the following steps: Placing identification points on key parts of the trainee, and marking different identification points according to limb classification to form a set of identification points corresponding to the limb classification; the limb classification includes upper limbs, torso and lower limbs; Use motion capture equipment to collect the displacement trajectory of different marking points during each shooting process of the person to be trained; Generate displacement trajectories of different limb classifications according to the set of identification points and the displacement trajectories of each identification point; Using an image acquisition device to obtain a basketball key frame of each shooting process of the person to be trained, and obtaining initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes a shooting position, a basketball motion trajectory, and a position that hits the basket; wherein the position that hits the basket at least includes a hit effective area, a miss effective area, and a hit basket; The initial shooting data is screened based on the position of the basket being hit, and the valid shooting data of the effective area and the basket being hit are retained; Classify the historical effective shooting data of the person to be trained based on the shooting position, and obtain a set of shooting data at the same shooting position; Analyze the correlation between the position of the basket hit and the displacement trajectory of different limb classifications in the shooting data set, and screen out the high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree; Combine shooting positions and high-correlation limb classifications to conduct shooting training for the trainees.

2. The basketball shooting training method according to claim 1, characterized in that: The step of marking different identification points according to the limb classification to form a set of identification points corresponding to the limb classification comprises: Get each marker point i , and based on the mapping function, each identification point i The ID is mapped to the corresponding limb category C ,in Represents a single identification point; They represent the classification of limbs into upper limbs, trunk, and lower limbs; Gather the identification points corresponding to the same limb classification to obtain the corresponding limb classification C The set of identification points ;in .

3. The shooting training method according to claim 1, characterized in that: The step of using a motion capture device to collect the displacement trajectory of different identification points during each shooting process of the person to be trained includes: Use motion capture equipment to synchronously collect different identification points during each shot of the trainee i location data; According to the time series, each mark point i At different collection times T The obtained position data are arranged to obtain the corresponding identification points of the person to be trained in the current shooting process i A collection of location points ;in , t is the total duration of the collection time; A collection of location points for a single marker Perform trajectory fitting to obtain corresponding identification points i The displacement trajectory .

4. The shooting training method according to claim 1, characterized in that: The step of generating displacement trajectories of different limb classifications according to the set of identification points and the displacement trajectories of each identification point comprises: Get limb classification C The set of identification points and the different identification points in the corresponding identification point set i The displacement trajectory ; Limb classification is obtained by weighted average C The aggregate displacement trajectory of: ； in Classify the limbs C The aggregate displacement trajectory of For marking point i The weight of .

5. The basketball shooting training method according to claim 1, characterized in that: When the image acquisition device is used to obtain the basketball key frames of each shooting process of the person to be trained, the basketball key frames acquired by the image acquisition device include at least a first basketball key frame when the basketball is out of the hand and a second basketball key frame when the basketball first touches the backboard.

6. The basketball shooting training method according to claim 5, characterized in that: The step of analyzing the correlation between the position of hitting the basket and the displacement trajectories of different limb classifications in the shooting data set includes: Combined with limb classification based on the collection time of the first basketball key frame C The aggregate displacement trajectory is used to obtain the limb classification when the basketball is released C location data; Based on the classification of limbs when basketball is released C The position data of the limb category C at the current moment is obtained; Obtaining the straight-line distance between the basketball hitting position and the center of the basket based on the second basketball key frame; Calculate the limb classification for the shooting data of different limb classifications in the shooting data set C The Pearson correlation coefficient of the instantaneous height and straight-line distance was calculated and a significance test was performed; Obtaining limb classification in the shooting data set based on Pearson correlation coefficient results C The degree of relevance to the location of the basket hit.

7. The basketball shooting training method according to claim 6, characterized in that: When limbs are classified C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 1, it indicates that the limb classification C The displacement trajectory is strongly positively correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to 0, it indicates that the limb classification C The displacement trajectory is weakly correlated with the position of the basket; when the limb classification C When the Pearson correlation coefficient of the instantaneous height and the straight-line distance is close to -1, it indicates that the limb classification C The displacement trajectory is strongly negatively correlated with the position of the basket.

8. The basketball shooting training method according to claim 6, characterized in that: The computational limb classification C The formula for the Pearson correlation coefficient of the instantaneous height and straight-line distance is: ; in Classify the limbs C Pearson correlation coefficient between instantaneous height and straight-line distance; N is the number of shooting data in the shooting data set; is the first j The corresponding limb classification in the shooting data C The instantaneous height of is the first j The straight-line distance between the basketball's hitting position and the center of the basket in each shot data; Classify limbs in the shooting data set C The instantaneous height mean of It is the mean straight-line distance between the basketball hitting position and the center of the basket in the shooting data set.

9. The basketball shooting training method according to claim 1, characterized in that: When dividing the effective hit area and the effective miss area, the backboard area wireframe is used as the boundary, and the backboard area wireframe and the inner area are divided into the effective hit area; the area outside the backboard area wireframe is divided into the effective miss area.

10. A shooting training device combining motion capture and image recognition, characterized in that: The device comprises: A marking point setting module is used to place marking points on key parts of the trainee and mark different marking points according to limb classification to form a marking point set corresponding to the limb classification; the limb classification includes upper limbs, torso and lower limbs; A marking point acquisition module is used to use a motion capture device to collect the displacement trajectory of different marking points during each shooting process of the trainee; A limb trajectory acquisition module is used to generate displacement trajectories of different limb classifications based on a set of identification points and the displacement trajectories of each identification point; A shooting data acquisition module is used to use an image acquisition device to obtain a basketball key frame of each shooting process of the trainee, and obtain initial shooting data of the shooting process according to the basketball key frame, wherein the shooting data at least includes a shooting position, a basketball motion trajectory, and a position that hits the basket; wherein the position that hits the basket at least includes a hit effective area, a miss effective area, and a hit basket; A data screening module, used for screening the initial shooting data based on the position of the basket hit, and retaining the valid shooting data of the effective area hit and the basket hit; A data classification module, used to classify the historical effective shooting data of the trainee based on the shooting position, and obtain a set of shooting data of the same shooting position; A correlation analysis module is used to analyze the correlation between the position of the basket hit and the displacement trajectory of different limb classifications in the shooting data set, and screen out the high-correlation limb classifications and low-correlation limb classifications that affect the hit results based on the correlation degree; The result output module is used to combine the shooting position and the high-correlation limb classification to conduct shooting training for the trainees.

Images