Auxiliary judgment method and system for badminton sports

Abstract

The invention discloses an auxiliary judgment method and system for badminton sports, and belongs to the technical field of judgment assistance, and the method comprises the steps: building an actual site physical coordinate system according to a badminton site; building a virtual net belt module to obtain net touching information of the badminton and inputting the net touching information to a control center; a camera assembly is installed above the badminton court, and a plurality of two-dimensional code marks are installed at the bottom corners of the boundary of the badminton court; the camera assembly identifies a plurality of two-dimensional code marks to obtain site physical boundary information and inputs the site physical boundary information to the control center; the camera assembly acquires badminton image track information, performs optimization prediction on the badminton image track information in combination with Kalman filtering, acquires actual badminton track information by adopting perspective transformation, and inputs the actual badminton track information to the control center; and the control center judges the score of the badminton by combining the touch net information, the site physical boundary information and the actual track information of the badminton. The method can achieve the efficient, precise and automatic auxiliary judgment, and improves the fairness of the competition and the watching experience of the audience.

Description

Technical Field

[0001] The present invention relates to the technical field of referee assistance, and particularly relates to a badminton referee assistance device and a usage method thereof. Background Art

[0002] Traditional badminton game referees usually rely on manual judgment, which has a certain risk of misjudgment. Especially when the ball is moving at high speed, it may be difficult for the referee to accurately determine whether the badminton is out of bounds. In the prior art, there has been a target detection system based on video analysis to judge the out-of-bounds situation of the ball. However, this method has a certain lag and is mostly used to further judge whether the ball touches the line to score or is out of bounds when the ball landing point is near the boundary and the situation of whether the ball is out of bounds is ambiguous. And the judgment process usually needs to interrupt the game and disrupt the rhythm of the game, affecting the viewing experience of the audience.

[0003] Therefore, developing an efficient, accurate, and automated referee assistance device has become an urgent need to improve the fairness of the game and the game experience. Summary of the Invention

[0004] Aiming at the problems existing in the prior art, the present invention provides an auxiliary referee method and system for badminton sports, which can achieve efficient, accurate, and automated auxiliary refereeing during the badminton sports confrontation, and improve the fairness of the game and the viewing experience of the audience.

[0005] The technical solution of the present invention is as follows:

[0006] In the first aspect of the present invention, an auxiliary referee method for badminton sports is provided, including:

[0007] Establish an actual field physical coordinate system according to the badminton court;

[0008] Build a virtual net belt module to obtain the net-touching information of the badminton and input it into the control center;

[0009] Install a camera component above the field and install several two-dimensional code markers at the boundary bottom corners of the badminton court;

[0010] The camera component identifies the image position information of several two-dimensional code markers based on the target detection algorithm, and uses perspective transformation for the image position information to obtain the field physical boundary information and input it into the control center;

[0011] The camera component obtains the badminton image trajectory information based on the target detection algorithm, and combines Kalman filtering to optimize and predict the badminton image trajectory information, and uses perspective transformation for the optimized badminton image trajectory information to obtain the actual badminton trajectory information and input it into the control center;

[0012] The control center combines catenary information, the physical boundary information of the venue, and the actual trajectory information of the badminton to determine the badminton score.

[0013] In some embodiments of the present invention, the virtual net belt module includes a signal transmitting unit for emitting optoelectronic signals and a signal receiving unit for receiving optoelectronic signals. The signal transmitting unit and the signal receiving unit are symmetrically arranged at a certain distance apart.

[0014] In some embodiments of the present invention, the signal transmitting unit includes a signal transmitting unit bracket, on which a signal transmitting unit alignment component and a plurality of signal transmitting unit optoelectronic components are provided. The signal transmitting unit optoelectronic components can emit optoelectronic signals;

[0015] The signal receiving unit includes a signal receiving unit bracket, on which a signal receiving unit alignment component and a signal receiving unit optoelectronic component are provided. The signal receiving unit optoelectronic component can receive optoelectronic signals;

[0016] The signal transmitting unit alignment component and the signal receiving unit alignment component are arranged to be aligned with each other.

[0017] In some embodiments of the present invention, the virtual net belt module is built at the center position of the venue to obtain the catenary information of the badminton and input it into the control center. Specifically:

[0018] After the signal transmitting unit bracket and the signal receiving unit bracket are symmetrically arranged at a certain distance apart, the signal transmitting unit alignment component and the signal receiving unit alignment component are arranged to be aligned with each other. Then, a plurality of signal transmitting unit optoelectronic components are started to emit optoelectronic signals, and a plurality of signal receiving unit optoelectronic components are started to receive optoelectronic signals to form an optoelectronic signal net belt. When the badminton touches the optoelectronic signal net belt, catenary information is generated and input into the control center.

[0019] In some embodiments of the present invention, several two-dimensional code marks are installed at the boundary bottom corners of the badminton court. Specifically:

[0020] When the singles mode is adopted, several two-dimensional codes are pasted along the boundary of the singles court. The positions of several two-dimensional code marks are identified by the camera component, and the physical boundary information of the singles court is determined by combining the perspective transformation technology and input into the control center;

[0021] When the doubles mode is adopted, several two-dimensional codes are pasted along the boundary of the doubles court. The positions of several two-dimensional code marks are identified by the camera component, and the physical boundary information of the doubles court is determined by combining the perspective transformation technology and input into the control center.

[0022] In some embodiments of the present invention, the actual physical coordinates of a number of two-dimensional codes on the site are obtained. After the camera assembly identifies a number of two-dimensional code markers based on the target detection algorithm, the position information of a number of two-dimensional code avatars is obtained, and the image position information is mapped into the actual physical coordinate system by using perspective transformation to determine the boundary information of the site and input it into the control center.

[0023] In some embodiments of the present invention, the camera assembly obtains the badminton image trajectory information based on the target detection algorithm, and optimizes and predicts the badminton image trajectory information by combining Kalman filtering. Specifically:

[0024] The camera assembly obtains the image coordinates in the K-th frame image during the movement of the badminton based on the target detection algorithm, defines and predicts the update of the Kalman filter state of the image coordinates and speed of the K-th frame of the badminton to obtain the image coordinates of the badminton in the (K + 1)-th frame, and obtains the optimized badminton image trajectory information.

[0025] In some embodiments of the present invention, the control center determines the badminton score by combining the net-touching information, the physical boundary information of the site and the actual trajectory information of the badminton. Specifically:

[0026] When the control center receives the net-touching information of the badminton and determines that the badminton touches the net, and the control center determines that the badminton goes out of bounds according to the physical boundary information of the site and the actual trajectory information of the badminton, then no score is obtained;

[0027] When the control center receives the net-touching information of the badminton and determines that the badminton touches the net, and the control center determines that the badminton does not go out of bounds according to the physical boundary information of the site and the actual trajectory information of the badminton, then no score is obtained;

[0028] When the control center does not receive the net-touching information of the badminton and determines that the badminton does not touch the net, and the control center determines that the badminton goes out of bounds according to the physical boundary information of the site and the actual trajectory information of the badminton, then no score is obtained;

[0029] When the control center does not receive the net-touching information of the badminton and determines that the badminton does not touch the net, and the control center determines that the badminton does not go out of bounds according to the physical boundary information of the site and the actual trajectory information of the badminton, then a score is obtained.

[0030] In some embodiments of the present invention, the control center determines that the badminton goes out of bounds according to the physical boundary information of the site and the actual trajectory information of the badminton. Specifically: The control center calculates the intersection coordinates of the actual trajectory information of the badminton and the physical boundary information of the site based on the ray method, and if the intersection coordinates are outside the actual site physical coordinate system, it is determined that the badminton goes out of bounds;

[0031] The control center determines that the badminton has not gone out of bounds based on the physical boundary information of the venue and the actual trajectory information of the badminton. Specifically, the control center calculates the intersection coordinates of the actual trajectory information of the badminton and the physical boundary information of the venue based on the ray method. The intersection coordinates are within the actual physical coordinate system of the venue, and it is determined that the badminton has not gone out of bounds.

[0032] In the second aspect of the present invention, an auxiliary referee system for badminton is provided. The system is used to execute an auxiliary referee method for badminton as described in any one of the above.

[0033] One or more technical solutions of the present invention have the following beneficial effects:

[0034] The present invention can improve the accuracy of badminton score determination: By combining photoelectric switch sensors and QR code markers, it can accurately detect whether the badminton touches the net or goes out of bounds, reducing the error of human judgment. Especially in complex game scenarios, the application of QR code markers and perspective transformation technology greatly improves the recognition accuracy of the venue boundary and the badminton movement trajectory.

[0035] While ensuring the accuracy of score determination, it can realize the intelligence and automation of score determination. By using a camera component and the target detection algorithm YOLO to automatically track the position and trajectory of the badminton, compared with traditional human referees, it can make judgments faster and more accurately, avoiding disputes caused by human errors.

[0036] During the score determination process, it can provide real-time feedback and efficient alarm. By connecting various feedback means such as a sound alarm unit and a display screen in the control center, it can quickly inform the referee and the audience whether the badminton touches the net, goes out of bounds or scores, greatly improving the transparency and viewing pleasure of the game and reducing the pressure on the referee.

[0037] By setting up a virtual net module to replace the traditional physical net, the module uses a photoelectric switch sensor to replace the traditional physical net, avoiding problems such as wear and breakage that may occur in the traditional net, making the game environment more stable and reliable. At the same time, the bottom ends of the signal transmitting unit bracket and the signal receiving unit bracket are both provided with stable foldable support structures, and the upper part of the foldable support structure is connected with a telescopic rod, which is convenient for storage and transportation and can be used conveniently in different venues.

[0038] The auxiliary referee method and system for badminton provided by the present invention have low cost and are easy to install. Compared with other high-precision intelligent referee systems, the photoelectric switch sensors and QR code markers used in the present invention are relatively simple and low-cost, which can save costs during large-scale promotion and are easy to install and deploy in different venues. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Flow chart of an auxiliary referee method for badminton provided in Embodiment 1 of the present invention;

[0040] Figure 2 Overall structure diagram of the signal transmitting unit bracket and the signal receiving unit bracket provided in Embodiment 1 of the present invention;

[0041] Figure 3 Structural schematic diagram of the signal transmitting unit bracket provided in Embodiment 1 of the present invention;

[0042] Figure 4 Internal sectional view structural schematic diagram of the signal transmitting unit bracket provided in Embodiment 1 of the present invention;

[0043] Figure 5 Score determination flow chart of an auxiliary referee method for badminton provided in Embodiment 1 of the present invention.

[0044] In the figure: 1. Signal transmitting unit bracket; 11. Signal transmitting unit optoelectronic component; 12. Signal transmitting unit alignment component; 13. Signal transmitting unit bracket base; 2. Signal receiving unit bracket; 21. Signal receiving unit optoelectronic component 21; 22. Signal receiving unit alignment component; 23. Signal receiving unit bracket base. Detailed implementation manners

[0045] The present invention will be further described below in conjunction with the drawings and embodiments.

[0046] Embodiment 1

[0047] In a typical implementation manner of the present invention, an auxiliary referee method for badminton is proposed, including:

[0048] Establish an actual field physical coordinate system according to the badminton court;

[0049] Build a virtual net belt module to obtain the net contact information of the badminton and input it into the control center;

[0050] Install a camera component above the court and install several two-dimensional code markers at the bottom corners of the boundaries of the badminton court;

[0051] The camera component identifies the image position information of several two-dimensional code markers based on the target detection algorithm, and uses perspective transformation for the image position information to obtain the field physical boundary information and input it into the control center;

[0052] The camera component obtains the badminton image trajectory information based on the target detection algorithm, and optimizes and predicts the badminton image trajectory information in combination with Kalman filtering, and uses perspective transformation for the optimized badminton image trajectory information to obtain the actual badminton trajectory information and input it into the control center;

[0053] The control center combines catenary information, physical boundary information of the venue, and actual trajectory information of the badminton to determine the badminton score.

[0054] In this embodiment, first, a virtual net belt module is set up: in the part that replaces the traditional net belt in the center of the badminton court, multiple optoelectronic switch sensors are installed to form a virtual net belt area. The optoelectronic switch sensors can detect in real time whether the badminton touches the virtual net belt and give a prompt through a sound alarm;

[0055] Then, four-corner QR code marking and venue boundary recognition are carried out: corresponding QR code markings are pasted at the four corners of the venue according to the competition format (singles or doubles), and the positions of the QR codes are recognized in real time through a camera. Combining perspective transformation technology, the physical boundaries of the venue are calculated. The QR code markings provide an accurate boundary reference for the venue. When the singles mode is adopted, the QR codes are pasted according to the boundaries of the singles court to realize the calculation of the singles boundaries; when the doubles mode is adopted, the QR codes are pasted according to the boundaries of the doubles court to realize the calculation of the doubles boundaries.

[0056] Next, badminton position and trajectory recognition are carried out: cameras are installed above the venue, and the target detection algorithm YOLO is used to track the position of the badminton in real time.

[0057] YOLO is used for badminton target detection, and Kalman filtering is combined for trajectory prediction and update to make the detection results more stable and accurate, and finally realize the efficient determination of the badminton movement trajectory. Whether the badminton exceeds the venue boundary is judged according to its movement trajectory.

[0058] The YOLO algorithm has significant advantages over traditional methods in badminton trajectory determination. Its single-stage detection method has high real-time performance, can effectively track high-speed moving badminton, and at the same time optimizes small target detection through deep learning to improve the recognition accuracy. In addition, this algorithm has strong adaptability to complex lighting and background interference and can run on embedded devices to meet the efficient detection requirements of sports events and training.

[0059] Finally, out-of-bounds and score determination are carried out: combining the data of the optoelectronic switch sensors and the camera, it is automatically determined whether the badminton crosses the virtual net belt or goes out of bounds. The positions of the badminton and the QR code markings are captured through the camera, and the ray method or the angle method is used to judge whether the badminton exceeds the venue range.

[0060] In addition, the control center is also connected to a real-time feedback and alarm system: when the badminton touches the net, goes out of bounds or scores, the system will give a prompt through sound alarms, display screens and other feedback means so that the referee and the audience can understand the game situation in real time.

[0061] Further, the virtual net belt module includes a signal transmitting unit for transmitting optoelectronic signals and a signal receiving unit for receiving optoelectronic signals. The signal transmitting unit and the signal receiving unit are symmetrically arranged at a certain distance apart.

[0062] Further, the signal transmitting unit includes a signal transmitting unit bracket 1, on which there are a signal transmitting unit alignment component 12 and a plurality of signal transmitting unit optoelectronic components 11. The signal transmitting unit optoelectronic components 11 can transmit optoelectronic signals;

[0063] The signal receiving unit includes a signal receiving unit bracket 2, on which there are a signal receiving unit alignment component 22 and signal receiving unit optoelectronic components 21. The signal receiving unit optoelectronic components 21 can receive optoelectronic signals;

[0064] The signal transmitting unit alignment component 12 and the signal receiving unit alignment component 22 are arranged to be aligned with each other.

[0065] In this embodiment, at the bottom of the signal transmitting unit bracket 1, there is a triangular foldable signal transmitting unit bracket base 13. In the middle of the signal transmitting unit bracket 1, there is a signal transmitting unit alignment component 12. At the upper part of the bracket, a plurality of signal transmitting unit optoelectronic components 11 are arranged at intervals; at the bottom of the signal receiving unit bracket, there is a triangular foldable signal receiving unit bracket base 23. In the middle of the signal receiving unit bracket 2, there is a signal receiving unit alignment component 22. At the upper part of the bracket, a plurality of signal receiving unit optoelectronic components 21 are arranged at intervals.

[0066] With such an arrangement, it is convenient to place the signal transmitting unit bracket 1 and the signal receiving unit bracket 2 at the corresponding site for use. In addition, both the signal transmitting unit optoelectronic components 11 and the signal receiving unit optoelectronic components 21 use optoelectronic switch sensors.

[0067] By setting up the virtual net belt module to replace the traditional physical net belt, and this module uses optoelectronic switch sensors to replace the traditional physical net belt, it avoids problems such as wear and fracture that may occur in the traditional net belt, making the competition environment more stable and reliable. At the same time, both the bottom ends of the signal transmitting unit bracket 1 and the signal receiving unit bracket are provided with stable foldable support structures, and the upper parts of the foldable support structures are connected with telescopic rods, which is convenient for storage and transportation and can be used conveniently in different sites.

[0068] Further, building a virtual net belt module at the center position of the site to obtain the net contact information of the badminton and input it into the control center, specifically:

[0069] After symmetrically arranging the signal transmitting unit bracket 1 and the signal receiving unit bracket 2 at a certain distance, align the signal transmitting unit alignment component 12 and the signal receiving unit alignment component 22 with each other. Start multiple signal transmitting unit optoelectronic components 11 to emit optoelectronic signals, and start multiple signal receiving unit optoelectronic components 21 to receive optoelectronic signals to form an optoelectronic signal network belt. When the badminton touches the optoelectronic signal network belt, touch network information is generated and input into the control center.

[0070] Further, install several two-dimensional code markers at the boundary bottom corners of the badminton court. Specifically:

[0071] When using the singles mode, several two-dimensional codes are pasted along the boundary of the singles court. The positions of several two-dimensional code markers are identified by the camera component, and the physical boundary information of the singles court is determined by combining the perspective transformation technology and input into the control center.

[0072] When using the doubles mode, several two-dimensional codes are pasted along the boundary of the doubles court. The positions of several two-dimensional code markers are identified by the camera component, and the physical boundary information of the doubles court is determined by combining the perspective transformation technology and input into the control center.

[0073] Further, obtain the actual physical coordinates of several two-dimensional codes on the court. After the camera component identifies several two-dimensional code markers based on the object detection algorithm, obtain the position information of several two-dimensional code avatars, and use perspective transformation to map the image position information into the actual physical coordinate system to determine the boundary information of the court and input it into the control center.

[0074] Install two-dimensional code markers on the badminton court and identify the positions of the two-dimensional codes in real time through the camera, and combine the perspective transformation technology to determine the physical boundary of the court. To achieve this process, first, the position of the two-dimensional code needs to be mapped into the actual physical coordinate system of the court, and this process can be completed through perspective transformation. Specifically:

[0075] First, perform the positioning of the four-corner two-dimensional code markers;

[0076] Assume that the two-dimensional code markers at the four corners of the court are Q1, Q2, Q3, and Q4 respectively, and the actual physical coordinates of these two-dimensional codes are known on the court, which are:

[0077] (X1, Y1) represents the position of the two-dimensional code marker Q1 on the court.

[0078] (X2, Y2) represents the position of the two-dimensional code marker Q2 on the court.

[0079] (X3, Y3) represents the position of the two-dimensional code marker Q3 on the court.

[0080] (X4, Y4) represents the position of the QR code marker Q4 on the site.

[0081] After the camera identifies the QR code markers through the object detection algorithm YOLO, it obtains their pixel coordinates in the image.

[0082] Suppose these coordinates are respectively:

[0083] (x1, y1) represents the position of the QR code marker Q1 in the image.

[0084] (x2, y2) represents the position of the QR code marker Q2 in the image.

[0085] (x3, y3) represents the position of the QR code marker Q3 in the image.

[0086] (x4, y4) represents the position of the QR code marker Q4 in the image.

[0087] Second, perform perspective transformation. Through perspective transformation, the pixel coordinates of the QR code markers in the image can be mapped to the actual physical coordinate system, thereby determining the boundaries of the site. Perspective transformation is a two-dimensional affine transformation, which can usually be represented by a 3x3 matrix H. The matrix H maps the point (x, y) in the image coordinate system to the point (X, Y) in the physical coordinate system.

[0088] Among them, the perspective transformation formula is as follows:

[0089]

[0090] Among them, H is the perspective transformation matrix. This matrix can be calculated through at least four pairs of known corresponding points (image coordinates and physical coordinates). The specific calculation process usually uses the least squares method to solve the system of linear equations.

[0091] Third, calculate the perspective transformation matrix for the QR code markers (x1, y1), (x2, y2), (x3, y3), (x4, y4) at the four corners of the site and their corresponding actual physical coordinates (X1, Y1), (X2, Y2), (X3, Y3), (X4, Y4).

[0092] The perspective transformation matrix H can be calculated through the following steps:

[0093] Construct the system of equations: Use each pair of known QR code marker positions to obtain two equations (one corresponding to the x coordinate and the other corresponding to the y coordinate). For each pair of points (x i , y i ) and (X i , Y i ), the system of equations is as follows:

[0094]

[0095] Solve for matrix H: Solve for H using the least squares method or other methods through at least four pairs of known points.

[0096] Fourth, calculate the field boundary. After completing the perspective transformation, obtain the mapping relationship from the image coordinate system to the physical coordinate system. Next, the position of the badminton can be recognized by the camera, and whether the badminton is out of the field can be judged by combining the boundary coordinates obtained after the perspective transformation.

[0097] Detect the position of the badminton (x ball , y ball ) in real time through the camera, and convert it into the actual field coordinates (X ball , Y ball ) through the perspective transformation formula.

[0098] Compare the physical coordinates (X ball , Y ball ) of the badminton with the physical boundary of the field to judge whether the badminton has crossed the boundary of the field.

[0099] If (X ball , Y ball ) falls within the valid area of the field, the badminton is not out of bounds; otherwise, it is judged to be out of bounds.

[0100] Furthermore, the camera component obtains the badminton image trajectory information based on the object detection algorithm, and optimizes and predicts the badminton image trajectory information in combination with the Kalman filter. Specifically:

[0101] The camera component obtains the image coordinates in the Kth frame of the badminton movement process based on the object detection algorithm, defines and predicts and updates the Kalman filter state of the image coordinates and speed of the Kth frame of the badminton, obtains the image coordinates of the (K + 1)th frame of the badminton, and obtains the optimized badminton image trajectory information.

[0102] In this embodiment, the method for identifying the position and trajectory of the badminton is specifically:

[0103] (1) Input the result of object detection (YOLO). Assume that the position of the badminton is detected by YOLO for each frame of the image, and the output is a bounding box containing the position of the badminton (x k , y k ), that is, the pixel coordinates in the kth frame. In order to further track and optimize these positions, the Kalman filter needs to combine these detection results with the estimated values of the previous frame to deduce a more accurate badminton trajectory.

[0104] (2) State definition of the Kalman filter. In order to optimize the position using the Kalman filter, we need to take the position and speed of the badminton as the state of the system.

[0105] Assume we are concerned with two-dimensional motion. The state vector can be defined as:

[0106]

[0107] where x k , y k is the position of the badminton at the k-th frame, and is the velocity of the badminton (the velocity components in the x and y directions).

[0108] (3) State transition equation. The core of the Kalman filter is the prediction and update process. First, we need to define the state transition model, that is, how to calculate the state of the current frame from the state of the previous frame.

[0109] Assume the time interval between each frame is Δt. The motion of the badminton can be approximated as uniform linear motion or motion with a constant acceleration in a short period. The state transition matrix F is defined as:

[0110]

[0111] This matrix describes the state update relationship from the (K - 1)-th frame to the K-th frame, where Δt is the time interval between two adjacent frames.

[0112] (4) Prediction step. In the prediction step of the Kalman filter, we predict the state of the current frame based on the state estimate of the previous frame:

[0113]

[0114] P k = F · P k-1 · F T + Q

[0115] is the predicted state of the current frame. Among them, P k is the predicted covariance matrix of the current state, representing the error covariance of the prediction. Q is the process noise matrix, representing the uncertainty of the system.

[0116] (5) Update step. In each frame, we obtain the detected position (x k , y k ) of the badminton through object detection (such as YOLO), and use it to update the state estimate of the Kalman filter. Assume the output of YOLO is a two-dimensional observation vector that is, the position of the badminton in the image.

[0117] Kalman gain:

[0118] K k = Pk ·H T ·(H·P k ·H T +R) -1

[0119] Among them, H is the observation matrix, and R is the observation noise covariance matrix, representing the noise of target detection.

[0120] State update:

[0121]

[0122] Among them, is the updated state estimate, representing position and velocity.

[0123] Covariance update:

[0124] P k =(I-K k ·H)·P k

[0125] Among them, I is the identity matrix, representing the updated covariance matrix.

[0126] Through the above prediction and update steps, we can smooth and optimize the position of the badminton. Whenever YOLO detects the current position of the badminton, the Kalman filter will output a more accurate position estimate based on the previous frame's estimate and the current measurement results. In practical applications, the Kalman filter will automatically reduce the impact of detection errors and noise, generating a smooth trajectory. By continuously iterating (for each frame), the trajectory of the badminton will be more accurate.

[0127] Furthermore, the control center combines the net-touching information, the physical boundary information of the court, and the actual trajectory information of the badminton to determine the score of the badminton, specifically as follows:

[0128] When the control center receives the net-touching information of the badminton and determines that the badminton touches the net, and the control center determines that the badminton goes out of bounds according to the physical boundary information of the court and the actual trajectory information of the badminton, then no score is awarded;

[0129] When the control center receives the net-touching information of the badminton and determines that the badminton touches the net, and the control center determines that the badminton does not go out of bounds according to the physical boundary information of the court and the actual trajectory information of the badminton, then no score is awarded;

[0130] When the control center does not receive the net-touching information of the badminton and determines that the badminton does not touch the net, and the control center determines that the badminton goes out of bounds according to the physical boundary information of the court and the actual trajectory information of the badminton, then no score is awarded;

[0131] If the control center does not receive the information that the badminton touches the net, it is determined that the badminton does not touch the net. If the control center determines that the badminton does not go out of bounds based on the physical boundary information of the court and the actual trajectory information of the badminton, then a score is obtained.

[0132] Further, when the control center determines that the badminton goes out of bounds based on the physical boundary information of the court and the actual trajectory information of the badminton, specifically: the control center obtains the intersection coordinates by calculating the actual trajectory information of the badminton and the physical boundary information of the court based on the ray method. If the intersection coordinates are outside the actual physical coordinate system of the court, it is determined that the badminton goes out of bounds;

[0133] When the control center determines that the badminton does not go out of bounds based on the physical boundary information of the court and the actual trajectory information of the badminton, specifically: the control center obtains the intersection coordinates by calculating the actual trajectory information of the badminton and the physical boundary information of the court based on the ray method. If the intersection coordinates are within the actual physical coordinate system of the court, it is determined that the badminton does not go out of bounds.

[0134] In this embodiment, the method for determining whether the badminton goes out of bounds and for scoring is specifically as follows:

[0135] The boundary of the court can be regarded as a polygon, and the position of the badminton is a point. If the coordinates of the badminton exceed the boundary of the polygon (i.e., the court area demarcated by the four QR codes), it is determined to be out of bounds.

[0136] The scoring determination mainly depends on whether the badminton touches the virtual net (detected by the photoelectric switch sensor) and whether it touches the scoring area of the court.

[0137] (1) Determination of touching the virtual net: The virtual net area is composed of multiple photoelectric switch sensors, located in the center of the court, replacing the traditional net. The photoelectric switch sensor can detect in real time whether the badminton touches the virtual net and give an alarm prompt. At the same time, the trajectory is used to determine the last hitting side, and thus the final determination is achieved.

[0138] (2) Comprehensive determination combining the photoelectric switch and the camera: In order to more accurately judge whether a score is obtained, the method of fusing the data of the photoelectric switch and the camera is very effective. The photoelectric switch sensor provides information on touching the net, while the camera judges whether the badminton has crossed the scoring line through the target detection algorithm.

[0139] The specific steps are as follows:

[0140] Determination of touching the net: The photoelectric switch sensor is used to detect in real time whether the badminton touches the virtual net.

[0141] Determination of going out of bounds: The ray method is used in combination with the QR code marking to judge whether the badminton has crossed the boundary of the court.

[0142] The ray method is used to determine whether a badminton goes out of bounds. The core lies in calculating the intersection points of the badminton trajectory and the court boundary, and judging whether the intersection points are outside the court range. The following is the specific mathematical model and equation derivation.

[0143] Assume that the badminton court is a rectangle, composed of four boundary lines:

[0144] Left boundary: x = x min

[0145] Right boundary: x = x max

[0146] Front court line: y = y max

[0147] Back court line: y = y min

[0148] Assume that the movement trajectory of the badminton is defined by the position points (X1, Y1) and (X2, Y2) of two consecutive frames. We can represent its trajectory with the straight-line equation:

[0149] y = mx + b

[0150] Among them, the slope m is calculated as:

[0151]

[0152] The intercept b is calculated as:

[0153] b = Y1 - mX1

[0154] To judge whether the badminton goes out of bounds, we need to calculate the intersection points of the trajectory and the court boundary.

[0155] For the left and right boundaries, two intersection points P L (x min , mx min + b) and P R (x max , mx max + b) can be obtained through equation calculation. If the y coordinate of the intersection point is within the court range (y min < y < y max ), then the badminton does not go out of bounds, otherwise it is judged to go out of bounds.

[0156] For the upper and lower boundaries, two intersection points and can be obtained through equation calculation. If the x coordinate of the intersection point is within the court range (x min < x < x max ), then the badminton does not go out of bounds, otherwise it is judged to go out of bounds.

[0157] As an important auxiliary judgment tool, in real-time matches and high-precision detection systems, combined with the landing point detection, the ray method can not only provide an out-of-bounds warning when YOLO drops frames, improve the credibility of the ruling, but also assist in tactical analysis, study the badminton flight trajectory, and implement a more robust and intelligent out-of-bounds judgment scheme.

[0158] To ensure the accuracy of the judgment, we not only calculate the intersection points of the trajectory and the boundary, but also need to make a judgment in combination with the actual position of the landing point.

[0159] Score judgment method combining the badminton trajectory system of the virtual net:

[0160] In this embodiment, the photoelectric switch virtual net is combined with the badminton trajectory system analyzed by the camera to achieve accurate badminton net crossing and score judgment. Combining the information of both can ensure the accuracy of the judgment and improve the intelligence level of the system.

[0161] Among them, the function of the photoelectric switch virtual net is: The photoelectric switch is installed in the plane of the net. Through the photoelectric sensor, a virtual net is formed to monitor in real time whether the badminton touches the net. If the photoelectric signal is not triggered, the badminton normally crosses the net without touching the net; if the photoelectric signal is triggered, the badminton hits the net and may fall back or roll over the net into the opponent's court.

[0162] Among them, the camera component and the YOLO trajectory analysis method are:

[0163] The camera component detects the badminton trajectory based on the YOLO algorithm and optimizes the trajectory prediction in combination with the Kalman filter, which is used to determine whether the badminton enters the opponent's court; calculates the landing point coordinates to judge whether it is out of bounds; combines the photoelectric switch signal to confirm whether it touches the net.

[0164] Fusing the photoelectric switch and the trajectory analysis method, when the badminton crosses the net, enters the opponent's court and is not out of bounds, it is recorded as a valid score; Trajectory analysis: The landing point of the badminton is within the court range. Photoelectric switch: Not triggered, indicating that the ball does not touch the net.

[0165] When the badminton touches the net but still enters the valid area, it is recorded as an invalid score. Trajectory analysis: The landing point of the badminton is within the court range. Photoelectric switch: Triggered, indicating that the ball has touched the net.

[0166] When the badminton is out of bounds, it is recorded as an invalid score. Trajectory analysis: The landing point of the badminton exceeds the boundary line. Photoelectric switch: Irrelevant, directly judged according to the trajectory.

[0167] In this embodiment, an external alarm system can also be connected to the virtual net for badminton out-of-bounds alarm.

[0168] In a second aspect of the present invention, there is provided an auxiliary umpiring system for badminton, which is used to execute an auxiliary umpiring method for badminton according to any one of the above.

[0169] Although the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, they do not limit the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solution of the present invention, various modifications or deformations that can be made by those skilled in the art without creative labor are still within the protection scope of the present invention.

Claims

1. An auxiliary umpiring method for badminton sports, characterized in that, Including: Establish an actual site physical coordinate system according to the badminton court; Build a virtual net belt module to obtain the net-touching information of the badminton and input it into the control center; Install a camera component above the site and install several two-dimensional code markers at the boundary bottom corners of the badminton court; The camera component identifies the image position information of several two-dimensional code markers based on the target detection algorithm, and uses perspective transformation to obtain the site physical boundary information and input it into the control center; The camera component obtains the badminton image trajectory information based on the target detection algorithm, and optimizes and predicts the badminton image trajectory information in combination with Kalman filtering, and uses perspective transformation to obtain the actual badminton trajectory information and input it into the control center; The control center determines the badminton score in combination with the net-touching information, the site physical boundary information and the actual badminton trajectory information.

2. The auxiliary referee method for badminton sports according to claim 1, wherein, The virtual net belt module includes a signal transmitting unit for transmitting photoelectric signals and a signal receiving unit for receiving photoelectric signals, and the signal transmitting unit and the signal receiving unit are symmetrically arranged at a certain distance interval.

3. The auxiliary referee method for badminton sports according to claim 2, wherein, The signal transmitting unit includes a signal transmitting unit bracket, and a signal transmitting unit alignment component and a plurality of signal transmitting unit photoelectric components are arranged on the signal transmitting unit bracket, and the signal transmitting unit photoelectric components can emit photoelectric signals; The signal receiving unit includes a signal receiving unit bracket, and a signal receiving unit alignment component and a signal receiving unit photoelectric component are arranged on the signal receiving unit bracket, and the signal receiving unit photoelectric component can receive photoelectric signals; The signal transmitting unit alignment component and the signal receiving unit alignment component are arranged to be aligned with each other.

4. The auxiliary umpiring method for badminton sports according to claim 4, wherein, Building a virtual net belt module at the center position of the site to obtain the net-touching information of the badminton and input it into the control center, specifically: After symmetrically arranging the signal transmitting unit bracket and the signal receiving unit bracket at a certain distance interval, align the signal transmitting unit alignment component and the signal receiving unit alignment component with each other, start a plurality of signal transmitting unit photoelectric components to emit photoelectric signals, and start a plurality of signal receiving unit photoelectric components to receive photoelectric signals to form a photoelectric signal net belt. When the badminton touches the photoelectric signal net belt, net-touching information is generated and input into the control center.

5. The auxiliary referee method for badminton sports according to claim 1, characterized in that Installing several two-dimensional code markers at the boundary bottom corners of the badminton court, specifically: When in the singles mode, several two-dimensional codes are pasted according to the boundary of the singles court, the positions of several two-dimensional code markers are identified by the camera component, and the physical boundary information of the singles court is determined in combination with the perspective transformation technology and input into the control center; When in the doubles mode, several two-dimensional codes are pasted according to the boundary of the doubles court, the positions of several two-dimensional code markers are identified by the camera component, and the physical boundary information of the doubles court is determined in combination with the perspective transformation technology and input into the control center.

6. The auxiliary referee method for badminton sports according to claim 1, characterized in that, Obtain the actual physical coordinates of several two-dimensional codes on the site. After the camera component identifies several two-dimensional code markers based on the target detection algorithm, obtain the image position information of several two-dimensional code avatars, and use perspective transformation to map the image position information into the actual physical coordinate system to determine the boundary information of the site and input it into the control center.

7. The auxiliary umpiring method for badminton sports according to claim 6, wherein, The camera component obtains the badminton image trajectory information based on the target detection algorithm, and optimizes and predicts the badminton image trajectory information by combining Kalman filtering. Specifically: The camera component obtains the image coordinates in the K-th frame of the badminton during the movement based on the target detection algorithm, defines the Kalman filtering state and predicts and updates the image coordinates and speed of the K-th frame of the badminton to obtain the image coordinates of the badminton in the (K + 1)-th frame, and obtains the optimized badminton image trajectory information.

8. The auxiliary umpiring method for badminton sports according to claim 1, wherein, The control center determines the badminton score by combining the net-touching information, the site physical boundary information, and the actual badminton trajectory information. Specifically: When the control center receives the badminton net-touching information and determines that the badminton touches the net, and the control center determines that the badminton goes out of bounds according to the site physical boundary information and the actual badminton trajectory information, then no score is awarded. When the control center receives the badminton net-touching information and determines that the badminton touches the net, and the control center determines that the badminton does not go out of bounds according to the site physical boundary information and the actual badminton trajectory information, then no score is awarded. When the control center does not receive the badminton net-touching information and determines that the badminton does not touch the net, and the control center determines that the badminton goes out of bounds according to the site physical boundary information and the actual badminton trajectory information, then no score is awarded. When the control center does not receive the badminton net-touching information and determines that the badminton does not touch the net, and the control center determines that the badminton does not go out of bounds according to the site physical boundary information and the actual badminton trajectory information, then a score is awarded.

9. The auxiliary umpiring method for badminton sports according to claim 8, wherein The control center determines that the badminton goes out of bounds according to the site physical boundary information and the actual badminton trajectory information. Specifically: The control center calculates the intersection coordinates of the actual badminton trajectory information and the site physical boundary information based on the ray method. If the intersection coordinates are outside the actual site physical coordinate system, it is determined that the badminton goes out of bounds. The control center determines that the badminton does not go out of bounds according to the site physical boundary information and the actual badminton trajectory information. Specifically: The control center calculates the intersection coordinates of the actual badminton trajectory information and the site physical boundary information based on the ray method. If the intersection coordinates are within the actual site physical coordinate system, it is determined that the badminton does not go out of bounds.

10. An auxiliary referee system for badminton sports, characterized in that, The system is used to execute an auxiliary referee method for badminton movement according to any one of claims 1-9.

Images