Bidirectional flying saucer data acquisition and analysis device and method

Abstract

The invention discloses a bidirectional flying saucer data acquisition and analysis device and method, belongs to the technical field of shooting and archery movement, computers, big data and image recognition, and solves the technical problems that the prior art depends on an inertial sensor and a camera to acquire data, but does not fully consider the flight path of a flying saucer; therefore, the problem of poor comprehensiveness and accuracy of data acquisition is solved. The system comprises a master high-speed camera for collecting video data of shooting postures of athletes, and a slave high-speed camera for collecting video data of pictures of a whole flying saucer target range, the trigger is connected with the master high-speed camera and used for receiving the trigger signal and sending out a synchronizing signal to trigger the master high-speed camera and the slave high-speed camera to synchronously collect video data, and the high-speed camera and the slave high-speed camera are connected through a synchronizing cable. And the upper computer is used for analyzing and visualizing video data synchronously acquired by the master high-speed camera and the slave high-speed camera. The method is used for bidirectional flying saucer data acquisition.

Description

Technical Field

[0001] A skeet data collection and analysis device and method are used for skeet data collection, belonging to the technical fields of shooting and archery, computers, big data and image recognition. Background Art

[0002] Skeet is a sport whose special nature makes data collection and analysis quite challenging. In this sport, athletes need to accurately hit the flying disc in a very short time, and it is necessary to accurately collect and analyze data such as the athletes' posture and reaction time to help them improve their technical level and competitive performance.

[0003] The flying disc flies at a very fast speed, and the whole process from preparation, aiming, firing to hitting is very short, almost instantaneous. Therefore, the data acquisition equipment needs to have the ability of high-speed continuous acquisition to accurately record the flight trajectory of the flying disc and the movements of the athlete.

[0004] The trajectory of a flying disc in the air requires the use of advanced image processing algorithms to detect and model the trajectory of the flying disc. At the same time, it is also necessary to monitor and analyze the athlete's posture and movements in real time to obtain key technical indicators, such as reaction time and consistency analysis.

[0005] Skeet is a high-speed, high-difficulty sport, and the performance of athletes is affected by many factors, such as environmental factors such as wind speed and light, as well as individual physical fitness and technical level. Therefore, data collection and analysis need to take these factors into account to obtain accurate and reliable data and provide personalized training and guidance for athletes.

[0006] There are many challenges in the current field of skeet data collection. One of the main challenges is the limitations of the data collection method. At present, inertial sensors are mainly used to collect the athlete's posture data, but the most critical problem is that the current data collection method fails to fully consider the flight trajectory of the skeet. Usually, only the athlete's posture data is collected, while the recording of the moment of the skeet launch and other key time nodes is ignored, which leads to incomplete data and affects the comprehensive analysis of the athlete's performance. To fully analyze the athlete's performance, we must pay attention to the flight trajectory of the skeet in order to comprehensively consider the posture data, so as to analyze the shooting preparation time, aiming time and other aspects.

[0007] In addition, the inertial sensors used in current data collection methods need to be installed on athletes, usually in close contact with the skin to ensure accurate data collection. However, this sensor installation method may have a significant impact on the athletes' shooting state, increase their psychological burden, and fail to simulate the usual training environment, resulting in athletes' feelings during the game being inconsistent with their usual training, which in turn affects their performance. Therefore, we need to seek lighter data collection equipment that does not need to be in close contact with the skin to reduce interference with athletes and better simulate actual training and competition environments.

[0008] Therefore, the prior art has the following technical problems:

[0009] 1. Traditional methods mainly rely on inertial sensors and cameras to collect data, but do not fully consider the flight trajectory of the UFO, resulting in poor comprehensiveness and accuracy of data collection.

[0010] 2. The inertial sensors used in traditional methods need to be attached to the athletes, which may affect the athletes' shooting state and feeling, increase their psychological burden, and thus cause poor reliability of data collection. Summary of the invention

[0011] In view of the above-mentioned research problems, the purpose of the present invention is to provide a two-way flying saucer data collection and analysis device and method to solve the problem that the prior art relies on inertial sensors and cameras to collect data but does not fully consider the flight trajectory of the flying saucer, thereby resulting in poor comprehensiveness and accuracy of data collection.

[0012] In order to achieve the above object, the present invention adopts the following technical solution:

[0013] A double-sided skeet data acquisition and analysis device comprises a main high-speed camera for acquiring video data of a player's shooting posture, a slave high-speed camera for acquiring video data of a picture of an entire skeet shooting range, a trigger connected to the main high-speed camera for receiving a trigger signal and sending a synchronization signal to trigger the main high-speed camera and the slave high-speed camera to synchronously acquire video data, the high-speed camera and the slave high-speed camera are connected via a synchronization cable, and a host computer for analyzing and visualizing the video data synchronously acquired by the main high-speed camera and the slave high-speed camera.

[0014] Furthermore, the main high-speed camera uses a high-speed camera with a focal length of 50 mm to collect video data of the athlete's shooting posture, and the slave high-speed camera uses a high-speed camera with a focal length of 18 mm to collect video data of the entire skeet shooting range, including the collected target building status, the trajectory of the skeet, and the target building indicator light signal.

[0015] Further, the specific steps of analyzing and visualizing the video data synchronously collected by the high-speed camera and the slave high-speed camera by the host computer are as follows: Step 1, preprocessing the video data synchronously collected by the main high-speed camera and the slave high-speed camera to obtain enhanced black-and-white video data;

[0016] Step 2: Modeling and processing the athlete's gun shooting posture based on the video data obtained by the main high-speed camera obtained by preprocessing to obtain visualized elbow joint angle change data, wherein the video data obtained by the main high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the athlete's shooting posture;

[0017] Step 3, marking key time points of the video data obtained by preprocessing from the high-speed camera, and obtaining the reaction time and shooting time of the athlete, wherein the video data obtained by preprocessing from the high-speed camera refers to the enhanced black-and-white video data of the entire skeet shooting range;

[0018] Step 4: Based on the pre-processed video data obtained from the high-speed camera, the UFO trajectory is modeled in stages to calculate the UFO flight time, horizontal displacement and vertical displacement, and the stages are divided and labeled;

[0019] Step 5: Based on the results obtained from Step 2 to Step 4 and the recorded hit situations, analyze the influence of joint angle posture on the hit, and analyze the influence of shooting reaction consistency and reaction time at each stage on the hit.

[0020] further,

[0021] The specific steps of step 1 are:

[0022] Step 1.1: Use a compression algorithm to perform video compression processing in H.264 encoding format on the video data synchronously collected by the main high-speed camera and the slave high-speed camera. The compression algorithm includes an inter-frame prediction method, an intra-frame prediction method, a quantization method and an entropy coding method. The inter-frame prediction method uses the correlation between video data frames to predict the pixel value of the current frame according to the reference frame. The intra-frame prediction method uses the correlation within the video data frame to predict the remaining pixel values ​​according to part of the pixel values ​​of the current frame. The quantization method quantizes the prediction error to reduce the representation accuracy of the data. The entropy coding method further compresses the data according to the statistical characteristics of the data.

[0023] Step 1.2, invert the colors of each frame of the compressed video data, that is, use the three components of the RGB color space to perform weighted average to convert the color value of each pixel in each frame of the image into a corresponding grayscale value, and obtain the grayscale image corresponding to each frame of the image after the conversion. After all the conversions are completed, black and white video data is obtained;

[0024] Step 1.3, using an image enhancement processing method to enhance the black and white video data, wherein the enhancement processing method includes a sharpening processing method and a noise reduction processing method;

[0025] The specific steps of step 2 are:

[0026] Step 2.1: Use the image recognition algorithm to model the preprocessed video data of the shooting posture to obtain a posture model. The specific steps are as follows:

[0027] Step 2.11, using a deep learning model to detect key points of the human body of the preprocessed video data of the shooting posture, including the elbow joint, shoulder, elbow and wrist, wherein the deep learning model is a convolutional neural network;

[0028] Step 2.12: construct a human posture model based on computer graphics technology and human key points, and superimpose it on the original shooting posture video data in the form of point and line markings to form a visible human posture model;

[0029] Step 2.2: Obtain elbow joint angle change data based on the human body posture model and associate and store it with the original shooting posture video data. The specific steps are as follows:

[0030] Step 2.21, based on the visible human posture model, obtain the position information of the elbow joint, and use the triangle calculation method to calculate the angle change data of the elbow joint in the time series, including the size and change trend of the angle;

[0031] Step 2.22, storing the calculated elbow joint angle change data in a database, and associating it with the time point or frame number of the video data of the original shooting posture;

[0032] Step 2.3: Display the change of the elbow joint angle based on the stored elbow joint angle change data and the video data of the original shooting posture. The specific steps are as follows:

[0033] Step 2.31, query and extract the elbow joint angle change data stored in the database;

[0034] Step 2.32, use data visualization tools to display the elbow joint angle change data in the form of charts or graphs;

[0035] The specific steps of step 3 are:

[0036] Step 3.1 performs image processing on the video data obtained from the high-speed camera to obtain the key time points. The specific steps are as follows:

[0037] Step 3.11, by comparing the pixel value differences between adjacent video frames in the video data of the preprocessed picture, use the opencv library function to detect the brightness change of the target building indicator light. When the indicator light changes from bright to dark, it means that the light is off, and the time point when the light is off is recorded as the key time point;

[0038] Step 3.12, according to the change of white pixels in the target building window after the signal light goes out, use the opencv library function to monitor the target building window to detect the appearance of the UFO. When white pixels are detected in the target building window, record the time when the UFO is launched as the key time point;

[0039] Step 3.2, based on the video data obtained from the high-speed camera obtained by preprocessing, other key time points are obtained, including the athlete's action time point, the athlete's firing time point and the flying disc hitting time point. The specific steps are as follows:

[0040] Step 3.21, by detecting and tracking the human body pixel points of the athletes in the video data of the pre-processed picture, if the time point at which the number of human body pixel points of the athletes in the previous and next two frames changes by more than three pixels or the time point at which the average pixel value of the human body pixel points of the athletes in the first three frames and the number of pixel points of the athletes corresponding to the fourth frame change by more than three pixels is taken as the time point of the athlete's action, including the start time of holding the gun and aiming;

[0041] Step 3.22, using the opencv library function to detect the change in the number of pixels in the hand-held muzzle area according to the time point of the athlete's action. When the number of pixels in the hand-held muzzle in the previous and next two frames changes by more than three, it means that the athlete has performed a firing action, and the time point is recorded as the athlete's firing time point;

[0042] Step 3.23, using the opencv library function to detect the pixel points of the target static area in the video data of the pre-processed picture, if the time point when the number of each pixel point in the target static area in the two previous and next frames changes by more than three, it is regarded as the UFO hit time point, wherein the target static area refers to the hit area range;

[0043] Step 3.3, based on the time difference between the athlete's action time, the athlete's firing time and the flying disc hitting time, the athlete's reaction time and shooting time are calculated using the formula:

[0044] Reaction time = athlete's action time - flying disc release time

[0045] Shooting time = athlete’s firing time - athlete’s action time;

[0046] The specific steps of step 4 are:

[0047] Step 4.1: Detect and model the UFO trajectory based on the pre-processed video data. The specific steps are as follows:

[0048] Step 4.11, take the time point when the UFO is launched as the starting point of modeling the UFO trajectory;

[0049] Step 4.12: Automatically model the UFO trajectory based on the starting point and other key time points, and extract parameters to calculate the UFO flight time, horizontal displacement and vertical displacement. The extracted parameters include flight path, initial velocity, gravity acceleration and impact position. That is, different colors are marked for different stages of the UFO operation based on the starting point and other key time points, and the UFO flight time, horizontal displacement and vertical displacement are calculated based on the markings. The calculation formula is:

[0050] UFO flight time t = UFO hit time - UFO launch time

[0051] V 0x =V0*cos(θ)

[0052] V 0y =V0*sin(θ)

[0053] x=V 0x *t

[0054] y=V 0y *t-1 / 2gt 2

[0055] Among them, V0 represents the initial velocity, V 0x and V 0y represents the horizontal and vertical positions of the hit position, x and y represent the horizontal displacement and vertical displacement respectively, θ represents the angle of 45°, and g represents the acceleration due to gravity;

[0056] Step 4.2: Stage division and labeling;

[0057] Step 4.21, the time from the release of the flying disc to the time of the athlete's action is regarded as the reaction stage, and marked;

[0058] Step 4.22, the time from the athlete's action point to the athlete's firing point is regarded as the gun movement stage, and it is marked;

[0059] Step 4.23: The time from the athlete's action point to the flying disc hitting point and thereafter is the firing and subsequent stages, and shall be marked.

[0060] Furthermore, the main high-speed camera and the slave high-speed camera are installed on a two-meter-high platform built ten meters behind the right side of the seventh position of the skeet shooting range through a three-section professional tripod.

[0061] A method for collecting and analyzing double-sided flying saucer data comprises the following steps:

[0062] Step S1, the trigger receives a trigger signal and sends a synchronization signal;

[0063] Step S2, the main high-speed camera and the slave high-speed camera synchronously collect video data after receiving a synchronization signal;

[0064] Step S3: The host computer analyzes and visualizes the video data of the athlete's shooting posture collected by the main high-speed camera and the video data of the entire skeet shooting range collected by the high-speed camera.

[0065] Further, the main high-speed camera in step S2 uses a high-speed camera with a focal length of 50 mm to collect video data of the athlete's shooting posture, and the slave high-speed camera uses a high-speed camera with a focal length of 18 mm to collect video data of the entire skeet shooting range, including the collected target building status, the trajectory of the skeet, and the target building indicator light signal.

[0066] Further, the specific steps in step S3 are: step 1, pre-processing the video data synchronously collected by the main high-speed camera and the slave high-speed camera to obtain enhanced black-and-white video data;

[0067] Step 2: Modeling and processing the athlete's gun shooting posture based on the video data obtained by the main high-speed camera obtained by preprocessing to obtain visualized elbow joint angle change data, wherein the video data obtained by the main high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the athlete's shooting posture;

[0068] Step 3, marking key time points of the video data obtained by preprocessing from the high-speed camera, and obtaining the reaction time and shooting time of the athlete, wherein the video data obtained by preprocessing from the high-speed camera refers to the enhanced black-and-white video data of the entire skeet shooting range;

[0069] Step 4: Based on the pre-processed video data obtained from the high-speed camera, the UFO trajectory is modeled in stages to calculate the UFO flight time, horizontal displacement and vertical displacement, and the stages are divided and labeled;

[0070] Step 5: Based on the results obtained from Step 2 to Step 4 and the recorded hit situations, analyze the influence of joint angle posture on the hit, and analyze the influence of shooting reaction consistency and reaction time at each stage on the hit.

[0071] further,

[0072] The specific steps of step 1 are:

[0073] Step 1.1: Use a compression algorithm to perform video compression processing in H.264 encoding format on the video data synchronously collected by the main high-speed camera and the slave high-speed camera. The compression algorithm includes an inter-frame prediction method, an intra-frame prediction method, a quantization method and an entropy coding method. The inter-frame prediction method uses the correlation between video data frames to predict the pixel value of the current frame according to the reference frame. The intra-frame prediction method uses the correlation within the video data frame to predict the remaining pixel values ​​according to part of the pixel values ​​of the current frame. The quantization method quantizes the prediction error to reduce the representation accuracy of the data. The entropy coding method further compresses the data according to the statistical characteristics of the data.

[0074] Step 1.2, invert the colors of each frame of the compressed video data, that is, use the three components of the RGB color space to perform weighted average to convert the color value of each pixel in each frame of the image into a corresponding grayscale value, and obtain the grayscale image corresponding to each frame of the image after the conversion. After all the conversions are completed, black and white video data is obtained;

[0075] Step 1.3, using an image enhancement processing method to enhance the black and white video data, wherein the enhancement processing method includes a sharpening processing method and a noise reduction processing method;

[0076] The specific steps of step 2 are:

[0077] Step 2.1: Use the image recognition algorithm to model the preprocessed video data of the shooting posture to obtain a posture model. The specific steps are as follows:

[0078] Step 2.11, using a deep learning model to detect key points of the human body of the preprocessed video data of the shooting posture, including the elbow joint, shoulder, elbow and wrist, wherein the deep learning model is a convolutional neural network;

[0079] Step 2.12: construct a human posture model based on computer graphics technology and human key points, and superimpose it on the original shooting posture video data in the form of point and line markings to form a visible human posture model;

[0080] Step 2.2: Obtain elbow joint angle change data based on the human body posture model and associate and store it with the original shooting posture video data. The specific steps are as follows:

[0081] Step 2.21, based on the visible human posture model, obtain the position information of the elbow joint, and use the triangle calculation method to calculate the angle change data of the elbow joint in the time series, including the size and change trend of the angle;

[0082] Step 2.22, storing the calculated elbow joint angle change data in a database, and associating it with the time point or frame number of the video data of the original shooting posture;

[0083] Step 2.3: Display the change of the elbow joint angle based on the stored elbow joint angle change data and the video data of the original shooting posture. The specific steps are as follows:

[0084] Step 2.31, query and extract the elbow joint angle change data stored in the database;

[0085] Step 2.32, use data visualization tools to display the elbow joint angle change data in the form of charts or graphs;

[0086] The specific steps of step 3 are:

[0087] Step 3.1 performs image processing on the video data obtained from the high-speed camera to obtain the key time points. The specific steps are as follows:

[0088] Step 3.11, by comparing the pixel value differences between adjacent video frames in the video data of the preprocessed picture, use the opencv library function to detect the brightness change of the target building indicator light. When the indicator light changes from bright to dark, it means that the light is off, and the time point when the light is off is recorded as the key time point;

[0089] Step 3.12, according to the change of white pixels in the target building window after the signal light goes out, use the opencv library function to monitor the target building window to detect the appearance of the UFO. When white pixels are detected in the target building window, record the time when the UFO is launched as the key time point;

[0090] Step 3.2, based on the video data obtained from the high-speed camera obtained by preprocessing, other key time points are obtained, including the athlete's action time point, the athlete's firing time point and the flying disc hitting time point. The specific steps are as follows:

[0091] Step 3.21, by detecting and tracking the human body pixel points of the athletes in the video data of the pre-processed picture, if the time point at which the number of human body pixel points of the athletes in the previous and next two frames changes by more than three pixels or the time point at which the average pixel value of the human body pixel points of the athletes in the first three frames and the number of pixel points of the athletes corresponding to the fourth frame change by more than three pixels is taken as the time point of the athlete's action, including the start time of holding the gun and aiming;

[0092] Step 3.22, using the opencv library function to detect the change in the number of pixels in the hand-held muzzle area according to the time point of the athlete's action. When the number of pixels in the hand-held muzzle in the previous and next two frames changes by more than three, it means that the athlete has performed a firing action, and the time point is recorded as the athlete's firing time point;

[0093] Step 3.23, using the opencv library function to detect the pixel points of the target static area in the video data of the pre-processed picture, if the time point when the number of each pixel point in the target static area in the two previous and next frames changes by more than three, it is regarded as the UFO hit time point, wherein the target static area refers to the hit area range;

[0094] Step 3.3, based on the time difference between the athlete's action time, the athlete's firing time and the flying disc hitting time, the athlete's reaction time and shooting time are calculated using the formula:

[0095] Reaction time = athlete's action time - flying disc release time

[0096] Shooting time = athlete’s firing time - athlete’s action time;

[0097] The specific steps of step 4 are:

[0098] Step 4.1: Detect and model the UFO trajectory based on the pre-processed video data. The specific steps are as follows:

[0099] Step 4.11, take the time point when the UFO is launched as the starting point of modeling the UFO trajectory;

[0100] Step 4.12: Automatically model the UFO trajectory based on the starting point and other key time points, and extract parameters to calculate the UFO flight time, horizontal displacement and vertical displacement. The extracted parameters include flight path, initial velocity, gravity acceleration and impact position. That is, different colors are marked for different stages of the UFO operation based on the starting point and other key time points, and the UFO flight time, horizontal displacement and vertical displacement are calculated based on the markings. The calculation formula is:

[0101] UFO flight time t = UFO hit time - UFO launch time

[0102] V 0x =V0*cos(θ)

[0103] V 0y =V0*sin(θ)

[0104] x=V 0x *t

[0105] y=V 0y *t-1 / 2gt 2

[0106] Among them, V0 represents the initial velocity, V 0x and V 0y represents the horizontal and vertical positions of the hit position, x and y represent the horizontal displacement and vertical displacement respectively, θ represents the angle of 45°, and g represents the acceleration due to gravity;

[0107] Step 4.2: Stage division and labeling;

[0108] Step 4.21, the time from the release of the flying disc to the time of the athlete's action is regarded as the reaction stage, and marked;

[0109] Step 4.22, the time from the athlete's action point to the athlete's firing point is regarded as the gun movement stage, and it is marked;

[0110] Step 4.23: The time from the athlete's action point to the flying disc hitting point and thereafter is the firing and subsequent stages, and shall be marked.

[0111] Furthermore, the main high-speed camera and the slave high-speed camera in step S2 are installed on a two-meter-high platform built ten meters behind the right side of the seventh position of the skeet shooting range through a three-section professional tripod.

[0112] Compared with the prior art, the present invention has the following beneficial effects:

[0113] 1. The device of the present invention constructs a complete skeet information collection system, which can comprehensively capture all aspects of skeet shooting and provide sufficient raw data support for subsequent data analysis. The two high-speed cameras operating in a master-slave synchronization mode ensure the synchronization and accuracy of data collection. The master high-speed camera is equipped with a trigger, set to external synchronization, and communicates normally with the host computer, thereby ensuring the stability and reliability of the data collection process;

[0114] Second, the present invention does not rely on inertial sensors to collect data, but completely relies on image processing algorithms to achieve the division of shooting stages and data collection. Through image processing technology, the athlete's reaction time, shooting time, elbow joint angle data and posture modeling data can be obtained without the need for additional sensor equipment. This method not only simplifies equipment configuration, but also provides more accurate and comprehensive data support, which is conducive to analyzing the athlete's shooting performance;

[0115] 3. In the process of acquiring key data, the present invention uses an image recognition algorithm to detect pixel points, accurately identify the time point of the disc being released, the time point of the athlete's action, the time point of firing, and the time point of hitting, and divides the disc point into different stages based on the above time points and highlights them in the video, so that the time of different shooting stages is obtained for subsequent analysis. In the process of human posture modeling, a deep learning model is used to model and acquire key joint motion data;

[0116] Fourth, the present invention can analyze and visualize historical data, and provide users with personalized training effect evaluation and guidance. The use of this software enables coaches and athletes to understand the training situation more intuitively, adjust the training plan in time, and improve training efficiency. This technical proposal not only has intuitive visualization of historical data, but also analyzes the difficult data collection process during the shooting process, including the analysis and visualization of athletes' reaction time, shooting time, and athlete reaction consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] Figure 1 It is a schematic diagram of the framework of the present invention;

[0118] Figure 2 The monitoring area of ​​the main high-speed camera;

[0119] Figure 3 The monitoring area from the high-speed camera. DETAILED DESCRIPTION

[0120] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods.

[0121] In an embodiment of the present application, a two-way skeet information acquisition system is provided, which includes two high-speed cameras, a host computer, and corresponding tripods, triggers, receivers and other equipment. The two high-speed cameras operate in a master-slave synchronization mode, and internal synchronization is achieved between the cameras. The master high-speed camera is equipped with a trigger, set to external synchronization, and communicates normally with the host computer. When the trigger signal is issued, the receiver receives the synchronization signal, and triggers the master high-speed camera and the slave high-speed camera to perform synchronous data acquisition, and after the trigger signal ends, the host computer can view the collected data. Ten meters behind the right side of the seventh position of the skeet shooting range, a two-meter-high platform is built, and a tripod is installed to place the master high-speed camera and the slave high-speed camera. Among them, the main high-speed camera uses a high-speed camera with a focal length of 50mm to collect video data of the athlete's shooting posture, and the slave high-speed camera uses a camera with a focal length of 18mm to collect video data of the entire skeet shooting range, including the state of the target building (including high and low target buildings), the trajectory of the flying saucer, and the target building indicator light signal. The host computer can also complete the analysis and visualization of historical data by uploading historical hit data.

[0122] Specific technical parameters: The acquisition platform is built at a height of 2m, and a three-section professional tripod is used to fix the high-speed camera. The two-meter platform and the three-section professional tripod are set up just enough to enable the main high-speed camera and the slave high-speed camera to collect all the video data of the entire shooting range and the athletes.

[0123] The master and slave high-speed cameras are set to master-slave synchronization to ensure the accuracy of data collection. The host computer first uses a compression algorithm to pre-process the original video, and then uses an image recognition algorithm to model the human posture to obtain key elbow joint data and identify the following key time points: the time when the signal light goes out, the time when the flying saucer is sent, the time when the athlete moves, the time when the athlete fires, and the time when the flying saucer hits. By calculating these time points, key information such as the athlete's action time and shooting time can be obtained.

[0124] Using image processing technology, the trajectory of the disc can be automatically modeled and divided into stages, and each stage can be labeled. This process is divided into three main stages: Athlete reaction stage: In this stage, the athlete has not yet reacted, the disc has just been launched, and the athlete is preparing to start the action. Athlete gun movement stage: This stage starts from the athlete's start of action until the moment of firing. In this stage, the athlete determines the position and speed of the disc and prepares to shoot. Firing and subsequent stages: After the firing moment, the system will monitor whether the disc is hit. This stage includes not only the firing moment, but also the subsequent flight trajectory of the disc, until it is finally determined whether the disc is successfully hit.

[0125] Finally, by receiving and preprocessing the coaching records, key information such as training date, duration and hit rate is extracted, converted into structured data and stored in the database. Based on historical data, the system displays basic personal information such as personal hit rate, assessment ranking and training ranking, and calculates the hit rate of training on the day, data analysis in different directions and the hit rate of each round of training, providing users with personalized training effect evaluation. In addition, the system realizes the analysis of the hit rate of different training targets, and draws a trend chart of the training assessment hit rate for a certain month, providing comprehensive and visual training effect analysis and guidance. Specifically:

[0126] A double-sided skeet data acquisition and analysis device comprises a main high-speed camera for acquiring video data of a player's shooting posture, a slave high-speed camera for acquiring video data of a picture of an entire skeet shooting range, a trigger connected to the main high-speed camera for receiving a trigger signal and sending a synchronization signal to trigger the main high-speed camera and the slave high-speed camera to synchronously acquire video data, the high-speed camera and the slave high-speed camera are connected via a synchronization cable, and a host computer for analyzing and visualizing the video data synchronously acquired by the main high-speed camera and the slave high-speed camera.

[0127] Configure the trigger for the main high-speed camera to ensure that the trigger can accurately send out the synchronization signal. First, configure the trigger on the main high-speed camera. The trigger is implemented through a hardware circuit with a built-in high-precision clock. The process of configuring the trigger includes connecting the trigger to the synchronization port of the main high-speed camera and inserting it correctly to ensure that it can send out the synchronization signal. The main high-speed camera has a built-in receiver to ensure that the camera can receive external synchronization signals and start the internal synchronization function. Connect the configured trigger to the main high-speed camera. The main high-speed camera is equipped with a receiver for receiving synchronization signals, which is used to receive the synchronization signal sent by the trigger. The process of configuring the receiver includes connecting the trigger interface to ensure that it can correctly receive the synchronization signal, and start the synchronization mechanism inside the camera after receiving the signal. After the synchronization cable connects the slave high-speed camera to the synchronization output port of the main high-speed camera, ensure that the synchronization cable is firmly connected and the signal transmission is smooth, the receiver can trigger the synchronization processing of the main high-speed camera and the slave high-speed camera to ensure that data acquisition can start at the same time.

[0128] The specific steps of the host computer analyzing and visualizing the high-speed camera and the video data synchronously collected from the high-speed camera are:

[0129] Step 1: Preprocess the video data synchronously collected by the main high-speed camera and the slave high-speed camera to obtain enhanced black-and-white video data. The specific steps are as follows:

[0130] Step 1.1 uses a compression algorithm to perform video compression processing in H.264 encoding format on the video data synchronously collected by the main high-speed camera and the slave high-speed camera, wherein the compression algorithm includes an inter-frame prediction method, an intra-frame prediction method, a quantization method and an entropy coding method. The inter-frame prediction method uses the correlation between video data frames to predict the pixel value of the current frame according to the reference frame. The intra-frame prediction method uses the correlation within the video data frame to predict the remaining pixel values ​​according to part of the pixel values ​​of the current frame. The quantization method quantizes the prediction error to reduce the representation accuracy of the data. The entropy coding method further compresses the data according to the statistical characteristics of the data. The combination of the above compression algorithms enables the compressed video file to maintain a relatively high quality while reducing the amount of data.

[0131] Step 1.2, inverting the colors of each frame image in the compressed video data, that is, using the three components of the RGB color space for weighted averaging to convert the color value of each pixel in each frame image into a corresponding grayscale value, and obtaining the grayscale image corresponding to each frame image after the conversion. After all the conversions are completed, black and white video data is obtained to enhance the visual appeal of the video content;

[0132] Step 1.3, using an image enhancement processing method to enhance the black and white video data to further improve the quality and viewing experience of the video, wherein the enhancement processing method includes a sharpening processing method and a noise reduction processing method;

[0133] Step 2: Model the athlete's shooting posture based on the video data obtained by the main high-speed camera obtained by preprocessing to obtain visualized elbow joint angle change data, wherein the video data obtained by the main high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the athlete's shooting posture. The specific steps are:

[0134] Step 2.1: Use the image recognition algorithm to model the preprocessed video data of the shooting posture to obtain a posture model. The specific steps are as follows:

[0135] Step 2.11, using a deep learning model to detect the key points of the human body of the preprocessed shooting posture video data, including the elbow joint, shoulder, elbow and wrist, to provide accurate data support for subsequent posture modeling, wherein the deep learning model is a convolutional neural network;

[0136] Step 2.12: construct a human posture model based on computer graphics technology and human key points, and superimpose it on the original shooting posture video data in the form of point and line markings to form a visible human posture model;

[0137] Step 2.2: Obtain elbow joint angle change data based on the human body posture model and associate and store it with the original shooting posture video data. The specific steps are as follows:

[0138] Step 2.21, based on the visible human posture model, obtain the position information of the elbow joint, and use the triangle calculation method to calculate the angle change data of the elbow joint in the time series, including the size and change trend of the angle;

[0139] Step 2.22, storing the calculated elbow joint angle change data in a database, and at the same time associating it with the time point or frame number of the original shooting posture video data, to ensure that the subsequent analysis can accurately correspond to the specific posture information in the original video;

[0140] Step 2.3: Display the change of the elbow joint angle based on the stored elbow joint angle change data and the video data of the original shooting posture. The specific steps are:

[0141] Step 2.31, query and extract the elbow joint angle change data stored in the database;

[0142] Step 2.32, use data visualization tools to display the elbow joint angle change data in the form of charts or graphs;

[0143] Step 3: marking key time points of the video data obtained from the high-speed camera obtained by preprocessing, and obtaining the reaction time and shooting time of the athlete, wherein the video data obtained from the high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the entire skeet shooting range. The specific steps are:

[0144] Step 3.1 performs image processing on the video data obtained from the high-speed camera to obtain the key time points. The specific steps are as follows:

[0145] Step 3.11, by comparing the pixel value differences between adjacent video frames in the video data of the preprocessed picture, use the opencv library function to detect the brightness change of the target building indicator light. When the indicator light changes from bright to dark, it means that the light is off, and the time point when the light is off is recorded as the key time point;

[0146] Step 3.12, according to the change of white pixels in the target building window after the signal light goes out, use the opencv library function to monitor the target building window to detect the appearance of the UFO. When white pixels are detected in the target building window, record the time when the UFO is launched as the key time point;

[0147] Step 3.2, based on the video data obtained from the high-speed camera obtained by preprocessing, other key time points are obtained, including the athlete's action time point, the athlete's firing time point and the flying disc hitting time point. The specific steps are as follows:

[0148] Step 3.21, by detecting and tracking the human body pixel points of the athletes in the video data of the pre-processed picture, if the time point at which the number of human body pixel points of the athletes in the previous and next two frames changes by more than three pixels or the time point at which the average pixel value of the human body pixel points of the athletes in the first three frames and the number of pixel points of the athletes corresponding to the fourth frame change by more than three pixels is taken as the time point of the athlete's action, including the start time of holding the gun and aiming;

[0149] Step 3.22, using the opencv library function to detect the change in the number of pixels in the hand-held muzzle area according to the time point of the athlete's action. When the number of pixels in the hand-held muzzle in the previous and next two frames changes by more than three, it means that the athlete has performed a firing action, and the time point is recorded as the athlete's firing time point;

[0150] Step 3.23, using the opencv library function to detect the pixel points of the target static area in the video data of the pre-processed picture, if the time point when the number of each pixel point in the target static area in the two previous and next frames changes by more than three, it is regarded as the UFO hit time point, wherein the target static area refers to the hit area range;

[0151] Step 3.3, based on the time difference between the athlete's action time, the athlete's firing time and the flying disc hitting time, the athlete's reaction time and shooting time are calculated using the formula:

[0152] Reaction time = athlete's action time - flying disc release time

[0153] Shooting time = athlete’s firing time - athlete’s action time;

[0154] Step 4: Based on the pre-processed video data obtained from the high-speed camera, the UFO trajectory is modeled in stages to calculate the UFO flight time, horizontal displacement and vertical displacement, and the stages are divided and labeled. The specific steps are as follows:

[0155] Step 4.11, take the time point when the UFO is launched as the starting point of modeling the UFO trajectory;

[0156] Step 4.12: Automatically model the UFO trajectory based on the starting point and other key time points, and extract parameters to calculate the UFO flight time, horizontal displacement and vertical displacement. The extracted parameters include flight path, initial velocity, gravity acceleration and impact position. That is, different colors are marked for different stages of the UFO operation based on the starting point and other key time points, and the UFO flight time, horizontal displacement and vertical displacement are calculated based on the markings. The calculation formula is:

[0157] UFO flight time t = UFO hit time - UFO launch time

[0158] V 0x =V0*cos(θ)

[0159] V 0y =V0*sin(θ)

[0160] x=V 0x *t

[0161] y=V 0y *t-1 / 2gt 2

[0162] Among them, V0 represents the initial velocity, V 0x and V 0y represents the horizontal and vertical positions of the hit position, x and y represent the horizontal displacement and vertical displacement respectively, θ represents the angle of 45°, and g represents the acceleration due to gravity;

[0163] Step 4.2: Stage division and labeling;

[0164] Step 4.21, the time from the release of the flying disc to the time of the athlete's action is regarded as the reaction stage, and marked;

[0165] Step 4.22, the time from the athlete's action point to the athlete's firing point is regarded as the gun movement stage, and it is marked;

[0166] Step 4.23, the time from the athlete's action point to the flying disc hitting point and thereafter is the firing and subsequent stages, and is marked;

[0167] Step 5: Based on the results obtained from Step 2 to Step 4 and the recorded hit situations, analyze the influence of joint angle posture on the hit, and analyze the influence of shooting reaction consistency and reaction time at each stage on the hit.

[0168] like Figure 1 The shooting range is shown, and the main high-speed camera and the slave high-speed camera are placed together, and the recorded video content is different. They are all placed on a 2m high platform. In summary, the human body structure of the athlete's posture is marked to facilitate the coach to see the posture problem; the change of the elbow joint angle should be visualized with a line graph; the consistency of the shooting reaction in each shooting process (reaction time, change of shooting time) is also presented through a line graph; the hit situation of different intervals of the shooting time reaction time is statistically analyzed; the change of reaction time on different days can reflect the athlete's attention situation; in summary, the present invention adopts the master-slave synchronization mode of two high-speed cameras, covering two important data sources of athlete posture and flying saucer trajectory, ensuring the comprehensiveness and accuracy of data collection. The present invention no longer relies on inertial sensors, reduces the burden on athletes, and improves the reliability of data collection. The present invention uses image processing algorithms to realize the automatic modeling and stage division of athlete posture and flying saucer trajectory, obtains the training data of key processes, and provides training data to facilitate the corresponding application analysis software or artificially generate shooting reaction consistency analysis, individual different time different target position different direction hit rate and other data analysis reports, and presents them in a visual form, providing important guidance for athlete training.

[0169] The above are only representative embodiments of the present invention in many specific application scopes, and do not constitute any limitation on the protection scope of the present invention. Any technical solutions formed by transformation or equivalent replacement fall within the protection scope of the present invention.

Claims

1. A double-sided flying saucer data collection and analysis device, characterized in that: The invention comprises a main high-speed camera for collecting video data of the shooting posture of athletes, a slave high-speed camera for collecting video data of the whole flying clay pigeon shooting range, a trigger connected to the main high-speed camera for receiving a trigger signal and sending a synchronization signal to trigger the main high-speed camera and the slave high-speed camera to synchronously collect video data, the high-speed camera and the slave high-speed camera are connected by a synchronization cable, and a host computer for analyzing and visualizing the video data synchronously collected by the main high-speed camera and the slave high-speed camera.

2. A double-sided flying saucer data acquisition and analysis device according to claim 1, characterized in that: The main high-speed camera uses a high-speed camera with a focal length of 50mm to collect video data of the athlete's shooting posture, and the slave high-speed camera uses a high-speed camera with a focal length of 18mm to collect video data of the entire skeet shooting range, including the collected target building status, the trajectory of the skeet, and the target building indicator light signal.

3. A double-sided flying saucer data collection and analysis device according to claim 2, characterized in that: The specific steps of analyzing and visualizing the video data synchronously collected by the high-speed camera and the slave high-speed camera by the host computer are as follows: Step 1, preprocessing the video data synchronously collected by the main high-speed camera and the slave high-speed camera to obtain enhanced black-and-white video data; Step 2: Modeling and processing the athlete's gun shooting posture based on the video data obtained by the main high-speed camera obtained by preprocessing to obtain visualized elbow joint angle change data, wherein the video data obtained by the main high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the athlete's shooting posture; Step 3, marking key time points of the video data obtained by preprocessing from the high-speed camera, and obtaining the reaction time and shooting time of the athlete, wherein the video data obtained by preprocessing from the high-speed camera refers to the enhanced black-and-white video data of the entire skeet shooting range; Step 4: Based on the pre-processed video data obtained from the high-speed camera, the UFO trajectory is modeled in stages to calculate the UFO flight time, horizontal displacement and vertical displacement, and the stages are divided and labeled; Step 5: Based on the results obtained from Step 2 to Step 4 and the recorded hit situations, analyze the influence of joint angle posture on the hit, and analyze the influence of shooting reaction consistency and reaction time at each stage on the hit.

4. A double-sided flying saucer data collection and analysis device according to claim 3, characterized in that: The specific steps of step 1 are: Step 1.1 uses a compression algorithm to perform video compression processing in H.264 encoding format on the video data synchronously collected by the main high-speed camera and the slave high-speed camera. The compression algorithm includes an inter-frame prediction method, an intra-frame prediction method, a quantization method and an entropy coding method. The inter-frame prediction method uses the correlation between video data frames to predict the pixel value of the current frame based on the reference frame. The intra-frame prediction method uses the correlation within the video data frame to predict the remaining pixel values ​​based on part of the pixel values ​​of the current frame. The quantization method quantizes the prediction error to reduce the representation accuracy of the data. The entropy coding method further compresses the data based on the statistical characteristics of the data: Step 1.2, invert the colors of each frame of the compressed video data, that is, use the three components of the RGB color space to perform weighted average to convert the color value of each pixel in each frame of the image into a corresponding grayscale value, and obtain the grayscale image corresponding to each frame of the image after the conversion. After all the conversions are completed, black and white video data is obtained; Step 1.3, using an image enhancement processing method to enhance the black and white video data, wherein the enhancement processing method includes a sharpening processing method and a noise reduction processing method; The specific steps of step 2 are: Step 2.1: Use the image recognition algorithm to model the preprocessed video data of the shooting posture to obtain a posture model. The specific steps are as follows: Step 2.11, using a deep learning model to detect key points of the human body of the preprocessed video data of the shooting posture, including the elbow joint, shoulder, elbow and wrist, wherein the deep learning model is a convolutional neural network; Step 2.12: construct a human posture model based on computer graphics technology and human key points, and superimpose it on the original shooting posture video data in the form of point and line markings to form a visible human posture model; Step 2.2: Obtain elbow joint angle change data based on the human body posture model and associate and store it with the original shooting posture video data. The specific steps are as follows: Step 2.21, based on the visible human posture model, obtain the position information of the elbow joint, and use the triangle calculation method to calculate the angle change data of the elbow joint in the time series, including the size and change trend of the angle; Step 2.22, storing the calculated elbow joint angle change data in a database, and associating it with the time point or frame number of the video data of the original shooting posture; Step 2.3: Display the change of the elbow joint angle based on the stored elbow joint angle change data and the video data of the original shooting posture. The specific steps are as follows: Step 2.31, query and extract the elbow joint angle change data stored in the database; Step 2.32, use data visualization tools to display the elbow joint angle change data in the form of charts or graphs; The specific steps of step 3 are: Step 3.1 performs image processing on the video data obtained from the high-speed camera to obtain the key time points. The specific steps are as follows: Step 3.11, by comparing the pixel value differences between adjacent video frames in the video data of the preprocessed picture, use the opencv library function to detect the brightness change of the target building indicator light. When the indicator light changes from bright to dark, it means that the light is off, and the time point when the light is off is recorded as the key time point; Step 3.12, according to the change of white pixels in the target building window after the signal light goes out, use the opencv library function to monitor the target building window to detect the appearance of the UFO. When white pixels are detected in the target building window, record the time when the UFO is launched as the key time point; Step 3.2, based on the video data obtained from the high-speed camera obtained by preprocessing, other key time points are obtained, including the athlete's action time point, the athlete's firing time point and the flying disc hitting time point. The specific steps are as follows: Step 3.21, by detecting and tracking the human body pixel points of the athletes in the video data of the pre-processed picture, if the time point at which the number of human body pixel points of the athletes in the previous and next two frames changes by more than three pixels or the time point at which the average pixel value of the human body pixel points of the athletes in the first three frames and the number of pixel points of the athletes corresponding to the fourth frame change by more than three pixels is taken as the time point of the athlete's action, including the start time of holding the gun and aiming; Step 3.22, using the opencv library function to detect the change in the number of pixels in the hand-held muzzle area according to the time point of the athlete's action. When the number of pixels in the hand-held muzzle in the previous and next two frames changes by more than three, it means that the athlete has performed a firing action, and the time point is recorded as the athlete's firing time point; Step 3.23, using the opencv library function to detect the pixel points of the target static area in the video data of the pre-processed picture, if the time point when the number of each pixel point in the target static area in the two previous and next frames changes by more than three, it is regarded as the UFO hit time point, wherein the target static area refers to the hit area range; Step 3.3, based on the time difference between the athlete's action time, the athlete's firing time and the flying disc hitting time, the athlete's reaction time and shooting time are calculated using the formula: Reaction time = athlete's action time - flying disc release time Shooting time = athlete’s firing time - athlete’s action time; The specific steps of step 4 are: Step 4.1: Detect and model the UFO trajectory based on the pre-processed video data. The specific steps are as follows: Step 4.11, take the time point when the UFO is launched as the starting point of modeling the UFO trajectory; Step 4.12: Automatically model the UFO trajectory based on the starting point and other key time points, and extract parameters to calculate the UFO flight time, horizontal displacement and vertical displacement. The extracted parameters include flight path, initial velocity, gravity acceleration and impact position. That is, different colors are marked for different stages of the UFO operation based on the starting point and other key time points, and the UFO flight time, horizontal displacement and vertical displacement are calculated based on the markings. The calculation formula is: UFO flight time t = UFO hit time - UFO launch time V 0x =V0*cos(θ) V 0y =V0*sin(θ) x=V 0x *t y=V 0y *t-1 / 2gt 2 Among them, V0 represents the initial velocity, V 0x and V 0y represents the horizontal and vertical positions of the hit position, x and y represent the horizontal displacement and vertical displacement respectively, θ represents the angle of 45°, and g represents the acceleration due to gravity; Step 4.2: Stage division and labeling; Step 4.21, the time from the release of the flying disc to the time of the athlete's action is regarded as the reaction stage, and marked; Step 4.22, the time from the athlete's action point to the athlete's firing point is regarded as the gun movement stage, and it is marked; Step 4.23: The time from the athlete's action point to the flying disc hitting point and thereafter is the firing and subsequent stages, and shall be marked.

5. The double-sided flying saucer data acquisition and analysis device according to claim 1, characterized in that: The main high-speed camera and the slave high-speed camera are installed on a two-meter-high platform built ten meters behind the right side of the seventh position of the skeet shooting range through three-section professional tripods.

6. A method for collecting and analyzing double-sided flying saucer data, characterized in that: The steps include: Step S1, the trigger receives a trigger signal and sends a synchronization signal; Step S2, the main high-speed camera and the slave high-speed camera synchronously collect video data after receiving a synchronization signal; Step S3: The host computer analyzes and visualizes the video data of the athlete's shooting posture collected by the main high-speed camera and the video data of the entire skeet shooting range collected by the high-speed camera.

7. A method for collecting and analyzing double-sided flying saucer data according to claim 6, characterized in that: The main high-speed camera in step S2 uses a high-speed camera with a focal length of 50 mm to collect video data of the athlete's shooting posture, and the slave high-speed camera uses a high-speed camera with a focal length of 18 mm to collect video data of the entire skeet shooting range, including the collected target building status, the trajectory of the skeet, and the target building indicator light signal.

8. A method for collecting and analyzing double-sided flying saucer data according to claim 7, characterized in that: The specific steps in step S3 are: step 1, pre-processing the video data synchronously collected by the main high-speed camera and the slave high-speed camera to obtain enhanced black-and-white video data; Step 2: Modeling and processing the athlete's gun shooting posture based on the video data obtained by the main high-speed camera obtained by preprocessing to obtain visualized elbow joint angle change data, wherein the video data obtained by the main high-speed camera obtained by preprocessing refers to the enhanced black-and-white video data of the athlete's shooting posture; Step 3, marking key time points of the video data obtained by preprocessing from the high-speed camera, and obtaining the reaction time and shooting time of the athlete, wherein the video data obtained by preprocessing from the high-speed camera refers to the enhanced black-and-white video data of the entire skeet shooting range; Step 4: Based on the pre-processed video data obtained from the high-speed camera, the UFO trajectory is modeled in stages to calculate the UFO flight time, horizontal displacement and vertical displacement, and the stages are divided and labeled; Step 5: Based on the results obtained from Step 2 to Step 4 and the recorded hit situations, analyze the influence of joint angle posture on the hit, and analyze the influence of shooting reaction consistency and reaction time at each stage on the hit.

9. A double-sided flying saucer data collection and analysis device according to claim 8, characterized in that: The specific steps of step 1 are: Step 1.1: Use a compression algorithm to perform video compression processing in H.264 encoding format on the video data synchronously collected by the main high-speed camera and the slave high-speed camera. The compression algorithm includes an inter-frame prediction method, an intra-frame prediction method, a quantization method and an entropy coding method. The inter-frame prediction method uses the correlation between video data frames to predict the pixel value of the current frame according to the reference frame. The intra-frame prediction method uses the correlation within the video data frame to predict the remaining pixel values ​​according to part of the pixel values ​​of the current frame. The quantization method quantizes the prediction error to reduce the representation accuracy of the data. The entropy coding method further compresses the data according to the statistical characteristics of the data. Step 1.2, invert the colors of each frame of the compressed video data, that is, use the three components of the RGB color space to perform weighted average to convert the color value of each pixel in each frame of the image into a corresponding grayscale value, and obtain the grayscale image corresponding to each frame of the image after the conversion. After all the conversions are completed, black and white video data is obtained; Step 1.3, using an image enhancement processing method to enhance the black and white video data, wherein the enhancement processing method includes a sharpening processing method and a noise reduction processing method; The specific steps of step 2 are: Step 2.1: Use the image recognition algorithm to model the preprocessed video data of the shooting posture to obtain a posture model. The specific steps are as follows: Step 2.11, using a deep learning model to detect key points of the human body of the preprocessed video data of the shooting posture, including the elbow joint, shoulder, elbow and wrist, wherein the deep learning model is a convolutional neural network; Step 2.12: construct a human posture model based on computer graphics technology and human key points, and superimpose it on the original shooting posture video data in the form of point and line markings to form a visible human posture model; Step 2.2: Obtain elbow joint angle change data based on the human body posture model and associate and store it with the original shooting posture video data. The specific steps are as follows: Step 2.21, based on the visible human posture model, obtain the position information of the elbow joint, and use the triangle calculation method to calculate the angle change data of the elbow joint in the time series, including the size and change trend of the angle; Step 2.22, storing the calculated elbow joint angle change data in a database, and associating it with the time point or frame number of the video data of the original shooting posture; Step 2.3: Display the change of the elbow joint angle based on the stored elbow joint angle change data and the video data of the original shooting posture. The specific steps are as follows: Step 2.31, query and extract the elbow joint angle change data stored in the database; Step 2.32, use data visualization tools to display the elbow joint angle change data in the form of charts or graphs; The specific steps of step 3 are: Step 3.1 performs image processing on the video data obtained from the high-speed camera to obtain the key time points. The specific steps are as follows: Step 3.11, by comparing the pixel value differences between adjacent video frames in the video data of the preprocessed picture, use the opencv library function to detect the brightness change of the target building indicator light. When the indicator light changes from bright to dark, it means that the light is off, and the time point when the light is off is recorded as the key time point; Step 3.12, according to the change of white pixels in the target building window after the signal light goes out, use the opencv library function to monitor the target building window to detect the appearance of the UFO. When white pixels are detected in the target building window, record the time when the UFO is launched as the key time point; Step 3.2, based on the video data obtained from the high-speed camera obtained by preprocessing, other key time points are obtained, including the athlete's action time point, the athlete's firing time point and the flying disc hitting time point. The specific steps are as follows: Step 3.21, by detecting and tracking the human body pixel points of the athletes in the video data of the pre-processed picture, if the time point at which the number of human body pixel points of the athletes in the previous and next two frames changes by more than three pixels or the time point at which the average pixel value of the human body pixel points of the athletes in the first three frames and the number of pixel points of the athletes corresponding to the fourth frame change by more than three pixels is taken as the time point of the athlete's action, including the start time of holding the gun and aiming; Step 3.22, using the opencv library function to detect the change in the number of pixels in the hand-held muzzle area according to the time point of the athlete's action. When the number of pixels in the hand-held muzzle in the previous and next two frames changes by more than three, it means that the athlete has performed a firing action, and the time point is recorded as the athlete's firing time point; Step 3.23, using the opencv library function to detect the pixel points of the target static area in the video data of the pre-processed picture, if the time point when the number of each pixel point in the target static area in the two previous and next frames changes by more than three, it is regarded as the UFO hit time point, wherein the target static area refers to the hit area range; Step 3.3, based on the time difference between the athlete's action time, the athlete's firing time and the flying disc hitting time, the athlete's reaction time and shooting time are calculated using the formula: Reaction time = athlete's action time - flying disc release time Shooting time = athlete’s firing time - athlete’s action time; The specific steps of step 4 are: Step 4.1: Detect and model the UFO trajectory based on the pre-processed video data. The specific steps are as follows: Step 4.11, take the time point when the UFO is launched as the starting point of modeling the UFO trajectory; Step 4.12: Automatically model the UFO trajectory based on the starting point and other key time points, and extract parameters to calculate the UFO flight time, horizontal displacement and vertical displacement. The extracted parameters include flight path, initial velocity, gravity acceleration and impact position. That is, different colors are marked for different stages of the UFO operation based on the starting point and other key time points, and the UFO flight time, horizontal displacement and vertical displacement are calculated based on the markings. The calculation formula is: UFO flight time t = UFO hit time - UFO launch time V 0x =V0*cos(θ) V 0y =V0*sin(θ) x=V 0x *t y=V 0y *t-1 / 2gt 2 Among them, V0 represents the initial velocity, V 0x and V 0y represents the horizontal and vertical positions of the hit position, x and y represent the horizontal displacement and vertical displacement respectively, θ represents the angle of 45°, and g represents the acceleration due to gravity; Step 4.2: Stage division and labeling; Step 4.21, the time from the release of the flying disc to the time of the athlete's action is regarded as the reaction stage, and marked; Step 4.22, the time from the athlete's action point to the athlete's firing point is regarded as the gun movement stage, and it is marked; Step 4.23: The time from the athlete's action point to the flying disc hitting point and thereafter is the firing and subsequent stages, and shall be marked.

10. A method for collecting and analyzing double-sided flying saucer data according to claim 6, characterized in that: The master high-speed camera and the slave high-speed camera in step S2 are mounted on a two-meter-high platform ten meters behind the right side of position seven of the skeet shooting range through three-section professional tripods.

Images