Novel curling intelligent monitoring and auxiliary training system

Abstract

The invention relates to the field of curling auxiliary training, and discloses a novel curling intelligent monitoring and auxiliary training system, which comprises a front monitoring module, a side monitoring module and a site monitoring module, and is characterized in that the side monitoring module comprises a first speed calculation unit, a posture detection unit, a motion target detection unit and a motion moment division unit; the site monitoring module is composed of a second speed calculation unit and a site mapping unit; the front monitoring module can be divided into a motion observation unit and a curling stone detection unit; according to the invention, deep learning and binocular vision technologies are utilized to provide more intuitive and professional exercise data and a playback and replaying mode for athletes and coaches, the training of the athletes is assisted, and good interactivity and real-time performance are achieved; all results are summarized to a central server, and information such as running tracks and speeds of curling stones in a single track of a training field, dynamic position distribution conditions and athlete technical and tactical states is presented for coaches through visual data, visual speed curves and joint angle charts, so that an auxiliary tool is provided for guiding training of athletes.

Description

Technical Field

[0001] The invention relates to the field of curling auxiliary training, and in particular to a novel curling intelligent monitoring and auxiliary training system. Background Art

[0002] Curling is an ice sport that requires teamwork and strategy. Players need to slide special curling stones (usually stones made of granite) on the ice and try to accurately send the curling stones to the center of the target area (called the "fortress"). The essence of curling is to control the speed and direction of the curling stones and adjust the sliding trajectory of the curling stones so that the team can occupy the advantageous area in the best way. Curling games are usually played on a piece of ice about 45 meters long and 5 meters wide. Each team has four players who take turns throwing the curling stones to push them to the target area (fortress). The game is played for 10 rounds, and the goal of each round is to keep your own curling stones in the fortress or as close to the center point of the fortress as possible (called the "center of the circle" or "button"). The score depends on which side's curling stone is closest to the center of the circle after each round.

[0003] According to the curling venue conditions and competition rules, the distance between the kicker and the base camp is large, and the closer the curling is to the center of the base camp, the higher the score. This requires that the curling needs to meet certain requirements for the line speed and angular velocity when throwing the curling. Therefore, curling athletes need to go through a lot of training and use reasonable tactics based on stable and accurate throwing techniques to win in the game. Since the coach judges the quality of curling throwing only by subjective consciousness, the implementation of the athlete's throwing action is mainly through proprioception and neuromuscular control, and there is a lack of objective data to quantify the throwing process; resulting in the coach's inability to accurately and comprehensively grasp the overall information of the venue. If an intelligent curling monitoring and auxiliary training system is developed, the coach can observe the athlete's throwing situation and venue information from different perspectives, and provide some objective data to quantify the action quality for adjustment, provide intuitive chart information and playback and review video, and play the role of auxiliary training, improve training efficiency, and be conducive to the promotion of curling and the improvement of scientific training. Summary of the invention

[0004] The purpose of the present invention is to provide a novel curling intelligent monitoring and auxiliary training system to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a new type of curling intelligent monitoring and auxiliary training system, comprising:

[0006] The front monitoring module includes a front visual acquisition control unit, a motion observation unit and a motion detection unit. It is responsible for observing the front situation of the curling arena, recording the process of the athlete from the starting step to the front throwing line from the front, and analyzing the athlete's posture information and the rotation information of the curling stone when it is thrown;

[0007] The side monitoring module includes a side vision acquisition control unit, a first speed calculation unit, a posture detection unit, a moving target detection unit and a motion moment division unit. The moving target detection unit uses a deep learning method to detect moving targets in the curling scene, with the main targets being athletes and curling. The posture detection unit uses a posture detection algorithm to estimate the human body posture and key point coordinates for subsequent calculation of key motion information. The first speed calculation unit calculates the parallax based on the result obtained by the binocular camera using the target detection unit, and after obtaining the three-dimensional coordinates of the target, calculates the speed through the position offset and time separated by several frames. The motion moment division unit divides the athlete's motion process from the starting pedal to the front throwing line into several key moments. At the key moments, the posture detection unit calculates the key human motion information, and presents the results to the coach in the form of data charts through the software interface.

[0008] The venue detection module includes a full-field visual acquisition control unit, a second speed calculation unit and a venue mapping unit. The venue mapping unit is calibrated in advance by the coach based on a standard curling venue, and the entire venue is mapped onto a two-dimensional plane map. The moving target detection unit detects key targets, and the second speed calculation unit calculates the target speed based on the known venue size.

[0009] Preferably: the front vision acquisition control unit adopts a combination of software and hardware to control the acquisition of images. The hardware includes 1 high-speed industrial camera, 1 telephoto lens, 1 embedded computer as an image acquisition platform, and a high-precision optical platform. The camera can be freely raised and lowered through the optical platform. The camera is an industrial camera with a resolution greater than 1080p and a frame rate greater than 60fps. The data interface is USB3.0. The camera supports programming control. The lens part adopts a telephoto lens with a focal length of 25mm, an object distance of 30m-40m, an imaging range of 5m in width and 2.5m in height. The embedded acquisition computer connects the camera and the main server, controls the camera to acquire images, and sends the image data to the main server for processing through network programming.

[0010] Preferably, the motion observation unit receives images acquired by each host, compresses the image sequence into a video, and simultaneously displays the video and subsequent processing results through a program software interface.

[0011] Preferably: the motion detection unit uses a target detection deep learning algorithm to detect the motion curling in the acquired image video, determine the rotation direction of the curling when the curling is released from the front throwing line, and provide data reference for the coach.

[0012] Preferably, the speed calculation unit uses a self-built binocular camera to obtain depth information in the scene, and completes binocular ranging by performing partial matching every several frames.

[0013] Preferably, the posture detection unit uses an algorithm to detect the athlete's posture information, calculates key motion information, and presents it to the coach and the athlete to assist them in quantitatively judging the motion process.

[0014] Preferably, the movement moment division unit divides the movement process of the athlete from the starting step to the front throwing line into several key moments, and the posture detection module calculates the key movement information of the human body at the key moments, and presents the results to the coach in the form of data charts through the software interface.

[0015] Preferably: the camera used by the side vision acquisition control unit is consistent with the front vision acquisition control unit, and the lens used by the camera used by the side vision acquisition control unit is a 3.5mm short-focus lens, so that preferably: the field mapping unit is calibrated by the coach in the image in advance, completes the mapping of the real field to the two-dimensional plane, and calculates the speed of the moving target in combination with the moving target detection module and the consistent standard field information; the moving target is dynamically displayed on the small plane map, allowing the coach to observe and make judgments.

[0016] Preferably: the camera used by the full-field vision acquisition control unit is consistent with the front vision acquisition control unit, and the camera used by the full-field vision acquisition control unit uses a lens with a focal length of 12mm. A dual camera is used to form a binocular vision system to complete the ranging function.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention uses deep learning and binocular vision technology to provide athletes and coaches with more intuitive and professional sports data and playback and review methods, assisting athletes in training, and has good interactivity and real-time performance. It can simultaneously observe the site conditions and perform real-time calculation and detection, and can also save the recorded playback for more specific analysis. It adopts a distributed layout method to observe and calculate respectively at three site locations, providing different functions, and finally summarizing all the results to a central server. In the form of intuitive data, visualized speed curves and action charts, feature display programs, and small map interfaces, the presenter can present the training site conditions and athlete status to the coach, making it convenient for the coach to intuitively understand the site distribution and athlete status. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is the overall structure block diagram of the system in the present invention;

[0020] Figure 2 is a system flow chart of the present invention;

[0021] Figure 3 A schematic diagram of the camera position and shooting range of the system of the present invention;

[0022] Figure 4It is a hardware connection framework diagram of the present invention;

[0023] Figure 5 A flow chart of positive monitoring according to an embodiment of the present invention;

[0024] Figure 6 This is a flow chart of side monitoring according to an embodiment of the present invention;

[0025] Figure 7 This is a flow chart of site monitoring according to an embodiment of the present invention;

[0026] Figure 8 This is a diagram of the moving target detection and posture detection results of the present invention;

[0027] Fig. 9 This is a graph showing the calculation result of the moving target speed of the present invention;

[0028] Fig.10 This is a diagram showing the calculation results of the body posture angle of the present invention;

[0029] Fig.11 This is a diagram of the sports venue mapping and curling trajectory calculation results of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Example

[0032] See also Figure 1 The figure shows a novel curling intelligent monitoring and auxiliary training system, which includes three main modules: a front monitoring module, a side monitoring module and a field monitoring module, which respectively represent the shooting and detection of the curling stadium from three different positions. In order to observe and calculate different types of motion data from different perspectives, the present invention adopts a distributed architecture. The front monitoring module focuses on observing whether the athlete deviates from the center line during the throwing stage. It does not require too much calculation data, and only adds auxiliary lines on the screen for judgment. The side monitoring module focuses on measuring the posture information and movement speed of the athlete when throwing the stone, so a binocular vision system is used to obtain depth information, and target detection and posture detection algorithms are used to monitor moving targets. The field monitoring module focuses on the movement trajectory of curling and the distribution of curling stones in the stadium. The three-dimensional stadium is first mapped to a two-dimensional plane, and then the target detection algorithm is used to collect the known curling venue dimensions to calculate the curling trajectory information and distribution. The system flow chart of the present invention is as follows Figure 2 shown.

[0033] Figure 3 The figure is a schematic diagram of the placement of the cameras of the three modules on the ice rink and their respective observation ranges. The camera of the front monitoring module is located behind the base camp, and the camera optical axis is aligned with the midline to shoot the front posture of the athlete. The observation range is from the starting step to the front throwing line, covering the athlete's back lead to the curling stone leaving the hand. The height of the athlete in the picture is controlled at about 1 / 3. This module uses an industrial camera with a telephoto lens to complete the coverage of the area from the front starting step to the front throwing line. The field detection module receives the image data transmitted by the visual acquisition control system, and uses the program to draw a cross line at the center point of the image. The coach can judge whether each throw is on the midline and whether the curling stone trajectory is offset. The motion detection unit is connected to the field detection module to display the detection results on the image. The specific workflow of the front monitoring module is that the acquisition host controls the camera to collect monocular images and transmits the images to the main server through the network port. The main server converts the image sequence into a video stream and displays it to the user through the field observation module. At the same time, the curling stone detection module deployed on the main server determines the position and direction of the curling stone in the image, and displays it to the user through the software interface. Figure 5 Flowchart of the front monitoring module.

[0034] The camera of the side monitoring module is located on the side of the ice rink, between the starting block and the front throwing line, observing the process of athletes throwing curling stones from the starting block to the front throwing line. Two industrial cameras are installed on the free lifting platform. The baseline can be adjusted according to the actual usage to adapt to different observation ranges. The camera has an imaging width of about 8m at a distance of 5m, which meets the shooting requirements of covering the side from the starting block to the front throwing line. The host controls the collection of images from the two cameras and transmits them to the main server for unified processing. The left and right camera images are first corrected and the detection results are output by the moving target detection unit. The first speed calculation unit matches the same object according to the detection results in the left and right cameras, calculates the parallax d, and calculates the three-dimensional coordinates according to the formula The position of the moving target in the left camera coordinate system is calculated by combining the internal and external parameters of the camera, and the speed is estimated according to the displacement and time of the camera at intervals of several frames. The posture detection unit uses the posture detection algorithm to detect the two-dimensional image position of each key point of the athlete, and then converts it to the three-dimensional coordinate system to calculate the key posture angles of the body, such as the three-dimensional angles formed by key points such as the hip joint, knee joint, shoulder joint, and elbow joint. The motion moment division unit divides the throwing process from the starting pedal to the front throwing line into 7 key moments: the athlete's back lead, the moment when the two feet are parallel, the moment of stepping forward, the curling stone reaches the white edge outside the base camp, the curling stone reaches the white edge inside the base camp, the curling stone reaches the red edge of the base camp, and the curling stone reaches the blue edge of the base camp. After the division is completed, the posture detection unit is connected to calculate the key posture angles of the body at each moment, and all the results are summarized and displayed on the system software interface. Figure 6 This is the flow chart of the side monitoring module. Figure 8 This is the result diagram of moving target detection and posture detection. Fig. 9 This is the result diagram of the moving target speed calculation. Fig.10 This is the result diagram of body posture angle calculation.

[0035] The venue monitoring module is responsible for observing the curling trajectory from the front throw line to the base camp and the distribution of curling stones in the base camp. Two industrial cameras are used to collect images. Since the curling venue is close to 50m, it is difficult for the general viewing angle to cover the entire venue. The module is equipped with a highly scalable optical platform in hardware, which can support a maximum of 8m lifting and lowering to ensure that the viewing angle range meets the requirements to the greatest extent. The module uses a lens with a focal length between 8mm and 15mm to ensure that the moving target can be observed at the farthest and closest distances. Before use, the embedded acquisition platform first takes a picture of the venue and sends it to the main server. The coach completes the venue calibration, and the coach needs to mark it in the picture. 2 front throwing lines and 2 side lines. After completion, the venue mapping unit completes the mapping from the three-dimensional venue space to the two-dimensional plane map, obtains the image from the image acquisition control end, and detects the object through the moving target detection unit. Combined with the length and width information of the known venue, the speed calculation unit measures the movement speed. The first speed calculation unit calculates the parallax by matching the left and right camera targets based on the binocular vision principle to obtain the three-dimensional coordinates. The second speed calculation unit calculates the three-dimensional coordinates and speed through the mapped two-dimensional map combined with the known venue size information, and obtains the curling trajectory and the final base camp curling position map, and displays the curling trajectory and base camp distribution on the software interface. Figure 7 This is a flow chart of site monitoring. Fig.11 This is a map of the sports field and the curling trajectory calculation results.

[0036] Figure 4 This is the hardware connection framework diagram of the present invention. The camera and the embedded image acquisition platform, and the acquisition platform and the general server with a computing / storage server are connected by network cables. The cameras at three positions are installed on a free-lifting tripod. Two industrial cameras are installed in the side monitoring module and the site monitoring module, and one industrial camera is installed in the front monitoring module.

[0037] The characteristic of the present invention is that the new intelligent monitoring and auxiliary training system for curling is a distributed system, and different modules are deployed at different positions of the competition venue to complete the monitoring of the curling venue from different perspectives; it detects and tracks targets based on computer vision, calculates athlete posture information, uses binocular vision to obtain three-dimensional coordinates, further calculates speed information, and uses venue mapping combined with known venue information to estimate the specific position of the target in the competition venue, and presents the post-processing information in combination with the original video to the coach for the coach to review and analyze, and at the same time provides specific data and charts to quantify the quality of the action or throwing process for the coach to refer to and make corrections, which plays a role in auxiliary training, improves training efficiency, and is conducive to the promotion of curling.

[0038] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0039] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A new type of curling intelligent monitoring and auxiliary training system, characterized in that: include: The front monitoring module includes a front visual acquisition control unit, a motion observation unit and a motion detection unit. It is responsible for observing the front situation of the curling arena, recording the process of the athlete from the starting step to the front throwing line from the front, and analyzing the athlete's posture information and the rotation information of the curling stone when it is thrown; A side monitoring module includes a side vision acquisition control unit, a first speed calculation unit, a posture detection unit, a moving target detection unit and a moving moment division unit. The moving target detection unit uses a deep learning method to detect moving targets in a curling scene, with the main targets being athletes and curling; The posture detection unit uses a posture detection algorithm to estimate the human posture and key point coordinates for subsequent calculation of key motion information; the first speed calculation unit calculates the parallax based on the result obtained by the binocular camera using the target detection unit, and calculates the speed through the position offset and time separated by several frames after obtaining the three-dimensional coordinates of the target; the motion moment division unit divides the athlete's motion process from the starting pedal to the front throwing line into several key moments, and the posture detection unit calculates the key motion information of the human body at the key moment, and presents the result to the coach in the form of a data chart through the software interface; The venue detection module includes a full-field visual acquisition control unit, a second speed calculation unit and a venue mapping unit. The venue mapping unit is calibrated in advance by the coach based on a standard curling venue, and the entire venue is mapped onto a two-dimensional plane map. The moving target detection unit detects key targets, and the second speed calculation unit calculates the target speed based on the known venue size.

2. A novel curling intelligent monitoring and auxiliary training system according to claim 1, characterized in that: The front vision acquisition control unit adopts a combination of software and hardware to control the acquisition of images. The hardware includes 1 high-speed industrial camera, 1 telephoto lens, 1 embedded computer as an image acquisition platform, and a high-precision optical platform. The camera can be freely raised and lowered through the optical platform. The camera is an industrial camera with a resolution greater than 1080p and a frame rate greater than 60fps. The data interface is USB3.

0. The camera supports programming control. The lens part adopts a telephoto lens with a focal length of 25mm, an object distance of 30m-40m, an imaging range of 5m wide and 2.5m high. The embedded acquisition computer connects the camera and the main server, controls the camera to acquire images, and sends the image data to the main server for processing through network programming.

3. A novel curling intelligent monitoring and auxiliary training system according to claim 2, characterized in that: The motion observation unit receives images acquired by each host, compresses the image sequence into a video, and simultaneously displays the video and subsequent processing results through a program software interface.

4. A novel curling intelligent monitoring and auxiliary training system according to claim 3, characterized in that: The motion detection unit uses a target detection deep learning algorithm to detect the moving curling stone in the acquired image video, determine the direction of rotation of the curling stone when it is released from the front throw line, and provide data reference for the coach.

5. A novel curling intelligent monitoring and auxiliary training system according to claim 4, characterized in that: The speed calculation unit uses a self-built binocular camera to obtain depth information in the scene, and completes binocular ranging by performing partial matching every several frames.

6. A novel curling intelligent monitoring and auxiliary training system according to claim 5, characterized in that: The posture detection unit uses an algorithm to detect the athlete's posture information, calculates key movement information, and presents it to the coach and the athlete to assist them in quantitatively judging the movement process.

7. A novel curling intelligent monitoring and auxiliary training system according to claim 6, characterized in that: The movement moment division unit divides the athlete's movement process from the starting step to the front throwing line into several key moments. At the key moments, the posture detection module calculates the key movement information of the human body and presents the results to the coach in the form of data charts through the software interface.

8. The novel curling intelligent monitoring and auxiliary training system according to claim 7 is characterized by: The camera used by the side vision acquisition control unit is the same as that of the front vision acquisition control unit, and the lens used by the camera used by the side vision acquisition control unit is a 3.5mm short-focus lens. A dual camera is used to form a binocular vision system to complete the speed measurement function.

9. A novel curling intelligent monitoring and auxiliary training system according to claim 8, characterized in that: The field mapping unit is used by the coach to calibrate the field in the image in advance, complete the mapping of the real field to the two-dimensional plane, and calculate the speed of the moving target in combination with the moving target detection module and the consistent standard field information; the moving target is dynamically displayed on the plane small map, allowing the coach to observe and make judgments.

10. A novel curling intelligent monitoring and auxiliary training system according to claim 9, characterized in that: The camera used by the full-field vision acquisition control unit is consistent with the front vision acquisition control unit, and the camera used by the full-field vision acquisition control unit uses a lens with a focal length of 12mm. A dual camera is used to form a binocular vision system to complete the ranging function.

Images