Methods, systems, storage media, and devices for enabling automatic measurement of a long jump

Abstract

The embodiment of the application discloses a method for realizing automatic measurement of long jump, which comprises the following steps: first, using a video acquisition device to acquire a long jump video; determining a long jump area in the long jump video, wherein the long jump area contains a plurality of calibrated measuring points; determining a to-be-detected person in the long jump video; determining a current state of the to-be-detected person according to a current area and a motion posture of the to-be-detected person; when the to-be-detected person is in a measuring state, acquiring a performance measuring point of the to-be-detected person; and finally, determining a long jump performance of the to-be-detected person according to the performance measuring point and the plurality of calibrated measuring points. The method provides a light, convenient, more accurate and stable long jump measurement technology, which does not need to be measured by manual operation, directly acquires a video to realize self-recognition and determination, thereby obtaining an accurate long jump performance, and can also be played back and viewed for multiple times, thereby effectively ensuring fairness and accuracy of the long jump performance.

Description

Technical Field

[0001] This invention relates to the field of sports training equipment technology, and in particular to a method, system, storage medium and device for automatically measuring long jump. Background Technology

[0002] The long jump, also known as the running long jump, is a jumping event in track and field. It consists of a combination of actions such as approach run, take-off, flight and landing. The athlete runs in a straight line, takes off with one foot behind the front line of the take-off board, goes through the flight phase, and then lands with both feet in the sandpit.

[0003] The standing long jump is an important event in school physical education tests. It tests the explosive power of the legs. The basic concept of the standing long jump is: a long jump that starts from a standing position without a running start. The position of the athlete's feet is not limited during the competition. Only one jump is allowed. If the athlete does not jump after both feet leave the ground and then jumps again after landing, it is considered as two consecutive jumps and is considered a failed attempt. It is frequently used in track and field training.

[0004] The traditional method for testing long jump involves drawing lines on the ground, including the take-off line and a ruler, and then visually estimating the jump result based on the test subject's landing point. This method, because it relies on human intervention to measure the long jumper's performance, is not very accurate, is prone to human error, and is also very labor-intensive.

[0005] In addition to the aforementioned methods of manual judgment for measuring the long jump or standing long jump, measurements can also be taken using infrared sensors. However, the results obtained by these two methods cannot be reviewed, and the accuracy of manual measurement is low, while infrared equipment is bulky and difficult to set up. Summary of the Invention

[0006] Therefore, it is necessary to propose a method for automatically measuring the long jump in response to the above problems.

[0007] A method for automatically measuring long jump, the method comprising the following steps:

[0008] Use video capture equipment to obtain long jump recordings;

[0009] The long jump area is determined in the long jump video, and the long jump area contains several calibrated measurement points;

[0010] The person to be tested is identified from the long jump video recording;

[0011] Determine the current state of the person being tested based on their current location and movement posture;

[0012] When the person to be tested is in a measurement state, acquire the person's performance measurement points;

[0013] The long jump score of the person being tested is determined based on the aforementioned performance measurement points and several calibrated measurement points.

[0014] In the above scheme, determining the current state of the person to be detected based on their current location and movement posture specifically includes:

[0015] When the person to be tested is first detected in the long jump video, their current status is determined to be that of a tourist and it is determined whether the person to be tested has entered the long jump preparation area.

[0016] If so, change the status of the person to be tested to the ready status;

[0017] When the posture of the person to be tested matches the take-off posture, their current state is changed to the flight state;

[0018] Start the timing function until the person to be tested lands, then change their current status to measurement status.

[0019] In the above scheme, the start-up timing function, which continues until the person to be tested lands, also includes:

[0020] Obtain the flight time of the person to be tested;

[0021] Determine whether the flight time of the person to be tested meets the threshold range;

[0022] If the conditions are met, the current state of the person to be tested is determined to be the measurement state.

[0023] In the above scheme, when the current state of the person to be tested is the measurement state, it is determined whether the person to be tested meets the measurement requirements;

[0024] If the conditions are met, the performance measurement points of the person to be detected are determined according to the global image segmentation algorithm;

[0025] The long jump score of the person being tested is determined based on the matching relationship between the measurement points and the numerical values ​​of the measurement points.

[0026] The measurement requirements in the above scheme specifically include:

[0027] Obtain the horizontal and vertical velocities of the feet of the person to be detected in the current frame;

[0028] The speed score of the key point is determined based on the horizontal and vertical speeds of the feet of the person being tested.

[0029] Determine whether the vertical velocity of the feet of the person being tested is less than a first fixed value;

[0030] If so, process the foot image of the person to be detected according to the foot segmentation algorithm to obtain the foot segmentation region; obtain the horizontal velocity and vertical velocity of the foot segmentation region;

[0031] The velocity fraction of the segmented area is determined based on the horizontal and vertical velocities of the segmented area.

[0032] The horizontal velocity of the subject's feet and the horizontal velocity of the segmented area of ​​the feet are measured.

[0033] Weighted summation is performed to obtain the horizontal velocity score of the part to be detected.

[0034] Obtain the geometric center point of the segmented region and calculate the velocity score of the geometric center point;

[0035] The landing score is obtained by weighted summing of the velocity scores of the key points, the segmented regions, the horizontal velocity scores of the parts to be detected, and the geometric center point.

[0036] When the landing score meets the scoring conditions, the current frame is determined to be the landing frame.

[0037] In the above scheme, the step of changing the current state to a flight state when the posture of the person to be tested matches the take-off posture specifically includes:

[0038] Obtain the current body tilt angle, forward arm swing amplitude, leg extension score, and foot movement of the person being tested;

[0039] The weighted sum of the body tilt angle, the amplitude of the forward arm swing, the score of the straight leg, and the amount of foot movement is calculated, and it is determined whether the take-off posture threshold is met.

[0040] If the conditions are met, determine whether the person being tested is continuously exerting force based on the key posture points at the current moment.

[0041] If so, determine whether the current posture of the person being tested matches the take-off posture.

[0042] In the above scheme, after determining the long jump area in the long jump video, the method further includes:

[0043] Obtain the long jump area from the long jump video;

[0044] The long jump area is divided to obtain several long jump area modules;

[0045] The exchange matrix is ​​processed on the several long jump region modules to obtain several perspective transformation diagrams;

[0046] The several perspective transformation images are stitched together to obtain the complete measurement area;

[0047] Each pixel in the measurement area corresponds one-to-one with the actual calibrated measurement point;

[0048] The actual long jump distance is determined based on the pixel distance in the measurement area.

[0049] This application also proposes a system for automatically measuring long jump, characterized in that the system includes: a video acquisition unit, a search unit, a status determination unit, and a long jump result acquisition unit;

[0050] The video acquisition unit is used to acquire long jump recordings using video acquisition equipment;

[0051] The search unit is used to determine the long jump area in the long jump video, the long jump area containing several calibrated measurement points, and to determine the person to be tested in the long jump video.

[0052] The state determination unit is used to determine the current state of the person to be tested based on the current area and movement posture of the person to be tested, and to obtain the performance measurement points of the person to be tested when the person to be tested is in the measurement state.

[0053] The long jump performance acquisition unit is used to determine the long jump performance of the person to be tested based on the performance measurement points and several calibrated measurement points.

[0054] This application also proposes a readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps:

[0055] Use video capture equipment to obtain long jump recordings;

[0056] The long jump area is determined in the long jump video, and the long jump area contains several calibrated measurement points;

[0057] The person to be tested is identified from the long jump video recording;

[0058] Determine the current state of the person being tested based on their current location and movement posture;

[0059] When the person to be tested is in a measurement state, acquire the person's performance measurement points;

[0060] The long jump score of the person being tested is determined based on the aforementioned performance measurement points and several calibrated measurement points.

[0061] This application also proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor in the following steps:

[0062] Use video capture equipment to obtain long jump recordings;

[0063] The long jump area is determined in the long jump video, and the long jump area contains several calibrated measurement points;

[0064] The person to be tested is identified from the long jump video recording;

[0065] Determine the current state of the person being tested based on their current location and movement posture;

[0066] When the person to be tested is in a measurement state, acquire the person's performance measurement points;

[0067] The long jump score of the person being tested is determined based on the aforementioned performance measurement points and several calibrated measurement points.

[0068] The embodiments of this invention have the following beneficial effects: First, a long jump video is acquired using a video capture device; a long jump area is determined in the video, which includes several calibrated measurement points; the person to be tested is identified in the video; the current state of the person to be tested is determined based on their current location and movement posture; when the person to be tested is in a measurement state, the performance measurement points of the person to be tested are acquired; finally, the long jump performance of the person to be tested is determined based on the performance measurement points and several calibrated measurement points. This method eliminates the need for manual measurement, directly acquiring video for autonomous recognition and judgment to obtain accurate long jump performance. It combines posture recognition, state judgment, and foot segmentation algorithms, which can consider more dimensions of information and allows for multiple playback reviews, effectively ensuring the fairness and accuracy of the long jump performance. Attached Figure Description

[0069] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] in:

[0071] Figure 1 This is a flowchart illustrating a method for automatically measuring the long jump in one embodiment;

[0072] Figure 2 This is a flowchart illustrating a method for determining the current state of a person being tested based on their current location and posture.

[0073] Figure 3 This is a schematic diagram of the human body's take-off posture in one embodiment;

[0074] Figure 4 This is a partial image of the landing frame of the person to be detected in one embodiment;

[0075] Figure 5 To and Figure 4 The complete image of the landing frame of the person to be detected;

[0076] Figure 6 A flowchart illustrating a method for determining the actual long jump distance in one embodiment;

[0077] Figure 7 This is a schematic diagram of a long jump measurement based on an original captured image in one embodiment;

[0078] Figure 8 This is a schematic diagram of long jump measurement based on the original captured images after separate calculation and processing for multiple regions. Detailed Implementation

[0079] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0080] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention; however, it will be apparent to those skilled in the art that the invention may be practiced without one or more of these details; in other instances, certain technical features well-known in the art have not been described in order to avoid confusion with the invention. It should be understood that the invention can be implemented in different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided to make the disclosure more comprehensive. For ease of understanding, the relevant terminology involved in this application will be introduced below.

[0081] To facilitate understanding, the relevant terms used in this application will be introduced below.

[0082] (1) Global image segmentation algorithm: Image segmentation is the first step in image analysis, the foundation of computer vision, an important part of image understanding, and also one of the most difficult problems in image processing. Image segmentation refers to dividing an image into several non-overlapping regions based on features such as grayscale, color, spatial texture, and geometric shape, so that these features show consistency or similarity in the same region, while showing obvious differences between different regions. Simply put, it is to separate the target from the background in an image.

[0083] (2) Weighted summation: Weighted summation refers to the process of weighting a series of values ​​according to certain weights and then summing them. The weights can be any real numbers, used to reflect the importance of each value in the sum. Generally, the larger the weight, the greater the proportion of the value in the sum.

[0084] (3) Perspective Transformation Diagram: The real world is three-dimensional, while images are two-dimensional (at least for now). If we want to describe the three-dimensional world with a two-dimensional image and make it look realistic enough, then the process of transforming the three-dimensional world into a two-dimensional image needs to satisfy certain geometric projection relationships, namely perspective relationships. A two-dimensional image that corresponds to the three-dimensional world and satisfies these perspective relationships is a perspective transformation diagram.

[0085] To fully understand the present invention, a detailed structure will be presented in the following description in order to illustrate the technical solution proposed by the present invention; optional embodiments of the present invention are described in detail below, however, in addition to these detailed descriptions, the present invention may have other embodiments.

[0086] like Figure 1 As shown, in one embodiment, a method for automatically measuring the long jump is provided. This method includes steps S101 to S106, which are detailed below:

[0087] S101. Obtain long jump video using video capture equipment;

[0088] The aforementioned video acquisition equipment includes electronic devices such as video recorders, cameras, and mobile phones. It converts analog video into digital video and saves it in the format of digital video files, making it convenient for manual playback and viewing later.

[0089] S102. Determine the long jump area in the long jump video. The long jump area contains several calibrated measurement points.

[0090] Among them, several calibrated measurement points can be preset according to different scenarios. Image processing technology is used to find the long jump area in the highest quality (high contrast, high brightness) frame of the long jump video from multiple frames.

[0091] S103. Identify the person to be tested from the long jump video;

[0092] Preferably, the person to be tested is determined by facial recognition algorithm or manually selected by the invigilator. If no selection is made, the person to be tested is automatically determined based on the area where the person is located. That is, when the person is in the preparation area or the long jump area, he / she will be identified as the person to be tested.

[0093] S104. Determine the current status of the person to be tested based on their current location and movement posture;

[0094] Specifically, the status of the person being tested includes tourist status, preparation status, flight status, and measurement status, which can represent the entire situation.

[0095] like Figure 2 As shown, in some embodiments, the current state of the person being detected is determined based on their current location and movement posture, specifically including:

[0096] S401. When the person to be tested is first detected in the long jump video, determine that their current status is tourist status and determine whether the person to be tested has entered the long jump preparation area.

[0097] S402. If so, change the status of the person to be tested to the ready status.

[0098] S403. When the posture of the person to be tested matches the take-off posture, change their current state to the flight state.

[0099] S404. Start the timing function until the person to be tested lands, and change their current state to the measurement state.

[0100] Preferably, after the person to be tested enters the preparation area relatively stably and changes the person's status to the preparation state, it is still necessary to determine whether the person has undergone a status transition and whether the person has crossed the line.

[0101] If the person being tested is detected to have left the preparation area and the direction of departure is not the direction of the long jump, and the departure from the preparation area is detected stably, then the person's status will be changed to tourist status; if the direction of departure is the direction of the long jump, then their status will be immediately changed to flight status.

[0102] The determination of whether someone has stepped on the line is based on the relationship between the foot section diagram and the position of the take-off line. If someone steps on the line, the score of that person will be invalidated.

[0103] In some embodiments, activating the timing function until the person being tested lands also includes:

[0104] Obtain the flight time of the person to be tested;

[0105] Determine whether the flight time of the person to be tested meets the threshold range;

[0106] If the conditions are met, the current state of the person to be tested is determined to be the measurement state.

[0107] Furthermore, if the flight time of the person to be tested does not meet the threshold range, their status will be changed to tourist status. If it is found that the person to be tested is unable to leave the measurement area within the specified time, it will be determined that the previous status transition (from preparation status to flight status) was incorrect, and the status of the person to be tested will be changed back to preparation status.

[0108] Furthermore, after determining that the current state of the person to be tested is the measurement state, their long jump performance is measured. When the measurement is completed or the person to be tested leaves the measurement area, their state is changed to tourist state.

[0109] In some embodiments, when the posture of the person being tested matches the take-off posture, their current state is changed to a flight state, specifically including:

[0110] (1) Obtain the body tilt angle, forward arm swing amplitude, straight leg score, and foot movement of the person being tested at the current moment;

[0111] (2) The weighted sum of the body tilt angle, the amplitude of the forward arm swing, the score of the straight leg and the amount of foot movement is calculated, and it is determined whether the take-off posture threshold is met.

[0112] (3) If satisfied, determine whether the person being tested is continuously exerting force at the current moment based on the key points of the posture;

[0113] Judging the sustained force can prevent feints. Without this judgment, when the test subject makes a fake jump, it may be mistaken for a flight state.

[0114] (4) If so, determine whether the posture of the person to be tested at the current moment matches the jumping posture.

[0115] like Figure 3 The diagram shown is a schematic of the human body's take-off posture. This diagram is used to illustrate (1) to (4) in this embodiment.

[0116] The diagram includes points identified based on the human skeletal key point model, as well as some necessary reference auxiliary lines.

[0117] Using vector methods, the law of cosines, and the law of sines, we obtain a directional angle ((-180~+180] or (-π~+π]) (clockwise is positive and counterclockwise is negative).

[0118] Wherein, vector: ,

[0119] Law of Cosines and Law of Sines: ,

[0120] Finally, the angle is calculated using inverse trigonometric functions: and

[0121] 1. Calculate the body tilt angle: Determine BE The angle between the vertical reference line and the vertical reference line, as shown in the diagram. ∠ BEF Only consider the size, set an upper and lower limit such as [10°, 40°], less than 10 is 0 points, greater than 40 is full points, and values ​​in the middle are scored proportionally;

[0122] 2. Calculate the amplitude of the forward swing arm: Calculate ∠ CBA ,from CB Rotate to AB Taking a jump to the right as an example: set an upper and lower limit, such as [50°, 135°]. A score less than 10 is 0 points, and a score greater than 40 is full marks. Values ​​in between are scored proportionally. If jumping to the left, the corresponding upper and lower limits become [-50°, -135°].

[0123] 3. Calculate the score for the straightened leg: Calculate ∠ CDE Size, from CD Rotate to ED For the angle, also set upper and lower limits [135°, 180°] to calculate the score, and calculate ∠. DCB Size, from DC Rotate to BC Similarly, calculate the score for ∠CDE and ∠DCB, and then sum the scores for ∠CDE and ∠DCB by weight to obtain the final score.

[0124] 4. Calculate key points of the posture to determine if continuous force is being applied: continuously calculate the angle ∠ CDE and ∠ DCB The difference between two angles gives the angular velocity, and the difference between two angular velocities gives the angular acceleration. Angular acceleration equals torque divided by moment of inertia. Since the change in moment of inertia during the two angle calculations is very small, ideally considered constant, the existence of angular acceleration means the existence of torque, which means that force is being applied. The remaining task is to determine whether the force is being applied continuously.

[0125] 5. Calculate foot movement: Record continuously. E The coordinates of the point are used to calculate the distance moved in each jump. A score of 0 is given for jumps in the opposite direction of the jump; otherwise, the distance traveled is multiplied by 2. sigmoid ( d / k The score is calculated using the function )-1, where d It refers to the distance of the movement in pixels. k It is a coefficient set by humans.

[0126] Preferably, the tracking algorithm can also be used to track the person to be detected. Traditional tracking algorithms only consider the detection box. The tracking algorithm of this method combines skeletal key points and integrates the scene posture determination of long jump. First, the human detection box is calculated with the traditional tracking algorithm to calculate the matching score. Then, the skeletal prediction point is predicted based on the changes of historical skeletal key points. During tracking, the posture distance score between the detected skeletal point and the prediction point is calculated. Finally, the scene conformity score is calculated according to the long jump scenario (for example, in the long jump scenario, when the current person to be detected is in an upright state, if the person's posture changes significantly next time, then the next posture is more likely to be crouching, because this is the take-off action, and the direction of movement of the person to be detected is more likely to be moving towards the long jump direction). Based on the above three scores, the highest score is taken as the tracking result.

[0127] In essence, tracking the person to be detected according to the tracking algorithm is to calculate the detection result of the previous image, and then iteratively judge whether all the detection results of the next image meet certain conditions. If multiple images meet the conditions, the similarity is calculated, and the one with the highest similarity is taken as the final tracking result.

[0128] S105. When the person to be tested is in the measurement state, obtain the performance measurement points of the person to be tested;

[0129] Preferably, when the person to be tested is in a measurement state, it is determined whether the person to be tested meets the measurement requirements;

[0130] If the conditions are met, the performance measurement points of the person to be detected are determined according to the global image segmentation algorithm;

[0131] The long jump performance of the person being tested is determined by matching the measurement points with the numerical values ​​of those points.

[0132] In some embodiments, the measurement requirements specifically include:

[0133] Obtain the horizontal and vertical velocities of the feet of the person to be detected in the current frame;

[0134] The speed score of the key point is determined based on the horizontal and vertical speeds of the feet of the person being tested.

[0135] Determine whether the vertical velocity of the person's feet is less than a first fixed value;

[0136] If so, process the foot image of the person to be detected according to the foot segmentation algorithm to obtain the foot segmentation region; obtain the horizontal velocity and vertical velocity of the foot segmentation region;

[0137] The velocity fraction of the segmented area is determined based on the horizontal and vertical velocities of the segmented area.

[0138] The horizontal velocity of the person's foot and the horizontal velocity of the segmented area of ​​the foot are weighted and summed to obtain the horizontal velocity score of the part to be tested.

[0139] Obtain the geometric center point of the segmented region and calculate the velocity score of the geometric center point;

[0140] The landing score is obtained by weighted summing of the velocity scores of key points, segmented regions, horizontal parts to be detected, and geometric center points.

[0141] When the landing score meets the scoring conditions, the current frame is determined to be the landing frame.

[0142] Preferably, the speed score at the geometric center point among the above four scores can be replaced by the speed score at the toes of the person being tested.

[0143] In some embodiments, the selection of the landing frame can be predicted using the following methods:

[0144] (1) Calculate the score of the human body's backward tilt angle.

[0145] (2) Calculate the score for leg bending.

[0146] (3) Calculate the ankle horizontal movement speed score

[0147] The three points above are weighted and summed. When the final score is greater than the landing threshold, it means that the person to be detected is about to land. That is, the landing frame will appear shortly afterward, and we can start to calculate more accurately whether the person to be detected has landed.

[0148] In fact, foot segmentation algorithms have a global perspective and are more accurate in segmentation. Typical vision-based long jump algorithms obtain ankle coordinates from skeletal keypoints, then assume the correct foot is located within that area, and crop that local region for segmentation. This approach is fast, covers a large target area, and is easy to implement. However, this method lacks a global perspective and is prone to errors in judgment when occlusion is present; even humans cannot accurately determine the correct foot position.

[0149] like Figure 4 As shown, when the image is cropped to a shape like this, it is difficult to determine which area is the foot. Even if one can roughly guess that the white area is the foot, the thought that key points may be misidentified, and that the image may not even have a foot, will lead to doubts about the previous judgment. This is precisely because there is too little information in the image.

[0150] However, if full-image input is used, such as Figure 5As shown, in the same scene, the algorithm can make a more certain judgment on the white area based on the human's posture. The segmentation algorithm can see whether there is occlusion and whether there is a second leg extending downward from the body. This can further confirm the conjecture about the white area. Often, such occluded small areas are the points that need to be measured in the final score or the judgment points for whether the line is stepped on. Using this full image input can greatly improve the foot segmentation effect, thereby greatly improving the robustness of the algorithm.

[0151] Preferably, the full-image input method significantly increases computational load, necessitating a reduction in the algorithm model's specifications. However, this leads to a decrease in accuracy. To achieve the original detection accuracy, some suggestive data can be provided during input. For example, detected key points can be input into the algorithm model along with the image from the previous frame for the current region. If the previous frame contains segmentation results, those results can be included. mask (Values ​​range from 0 to 1) are also entered; otherwise, all values ​​are 0.5. mask In this process, key points can be converted into an image (of course, they can be converted into images, but converting them into images will give the model more spatial information). For example, the skeletal key points can be drawn onto a black image. The mask itself is also an image. This stacking of data on channels will not affect the final speed, but it can improve accuracy.

[0152] S106. Determine the long jump score of the person to be tested based on the performance measurement points and several calibrated measurement points.

[0153] like Figure 6 As shown, in some embodiments, after determining the long jump area in the long jump video, the method further includes:

[0154] S601. Obtain the long jump area from the long jump video recording;

[0155] S602. Divide the long jump area and obtain several long jump area modules;

[0156] S603. Perform matrix exchange on several long jump area modules to obtain several perspective transformation diagrams;

[0157] S604. Combine several perspective transformation diagrams to obtain the complete measurement area;

[0158] S605. Each pixel in the measurement area corresponds one-to-one with the actual calibrated measurement point.

[0159] S606. Determine the actual long jump distance based on the pixel distance in the measurement area.

[0160] like Figure 7As shown, common measurement methods are based on the original captured image. These methods often involve dividing the region into equal parts, assuming that identical pixels represent the same distance. However, due to perspective distortion, this measurement accuracy is low and heavily relies on the number of calibration points. Another method involves measuring images obtained after calibration and perspective transformation. However, regardless of whether distortion correction is applied or a distortion-free camera is used, distortion still exists, leading to imperfect transformation matrices generated from calibration points. This can sometimes result in significant errors in certain areas.

[0161] like Figure 8 As shown, this method uses multi-region separate calculation of the transformation matrix. Because it employs cell-level calculation, the error within this range is very small, and this error will not be superimposed after the merging of multiple regions. Although the error still exists, it has been controlled to a very small range. The matrix calculation method can use... OpenCV In findHomography The function proceeds, and this transformation matrix can... Figure 7 become Figure 8 Ideally, Figure 8 The horizontal distance between each yellow dot is equal, making it easy to convert pixel distance to actual distance. However, in practice, uneven spacing among the yellow dots can easily occur. Therefore, [the text abruptly ends here]. Figure 8 The points (1, 2, 6, 7) are treated as a small region. A transformation matrix is ​​calculated, and after perspective transformation, the desired result can be obtained. Figure 3 In the image (1, 2, 6, 7), we calculate each small region separately, obtaining a perspective-transformed small image for each region. When we stitch all the small images together, we get a complete image. Figure 3 Furthermore, due to the segmentation, the error was greatly reduced, resulting in higher accuracy than before.

[0162] In summary, the attitude estimation method in this application can determine whether the device is in flight state in the first instance, avoiding misjudgment. Different measurement parameters are used for different areas, which is more accurate than a single set of global measurement parameters. In the selection of landing frames, attitude estimation is used to determine the process, taking into account more factors. The foot segmentation algorithm has a global perspective and is more accurate than general segmentation algorithms. In addition, to improve speed, suggestive parameters are also provided, allowing the foot segmentation algorithm model to achieve the effect of the original large model with fewer parameters. This provides a more accurate and convenient long jump measurement technology.

[0163] This application also proposes a system for automatically measuring long jump, the system including: a video acquisition unit, a search unit, a status determination unit, and a long jump result acquisition unit;

[0164] The video capture unit is used to acquire long jump recordings using video capture equipment.

[0165] The search unit is used to determine the long jump area in the long jump video, which contains several calibrated measurement points, and to identify the person to be tested in the long jump video.

[0166] The status determination unit is used to determine the current status of the person being tested based on the area and movement posture of the person being tested, and to obtain the performance measurement points of the person being tested when the person being tested is in the measurement state.

[0167] The long jump performance acquisition unit is used to determine the long jump performance of the person being tested based on the performance measurement points and several calibrated measurement points.

[0168] This application also proposes a readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps:

[0169] Use video capture equipment to obtain long jump recordings;

[0170] The long jump area is determined from the long jump video, and the long jump area contains several calibrated measurement points;

[0171] Identify the individuals to be tested from the long jump video recording;

[0172] The current status of the person being tested is determined based on their current location and movement posture.

[0173] When the person being tested is in the measurement state, acquire the person's score measurement points;

[0174] The long jump score of the person being tested is determined based on the performance measurement points and several calibrated measurement points.

[0175] This application also proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor in the following steps:

[0176] Use video capture equipment to obtain long jump recordings;

[0177] The long jump area is determined from the long jump video, and the long jump area contains several calibrated measurement points;

[0178] Identify the individuals to be tested from the long jump video recording;

[0179] The current status of the person being tested is determined based on their current location and movement posture.

[0180] When the person being tested is in the measurement state, acquire the person's score measurement points;

[0181] The long jump score of the person being tested is determined based on the performance measurement points and several calibrated measurement points.

[0182] Those skilled in the art will understand that implementing all or part of the processes in the above embodiments can be accomplished by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory. (ROM) Programmable ROM (PROM) Electrically programmable ROM (EPROM) Electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory. (RAM) Or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static... RAM (SRAM) ,dynamic RAM (DRAM) ,synchronous DRAM (SDRAM) Double data rate SDRAM (DDR SDRAM) Enhanced SDRAM (ESDRAM) Synchronization Link Synchlink DRAM (SLDRAM) Memory bus (Rambus) direct RAM (RDRAM) Direct Memory Bus Dynamics RAM (DRDRAM) and memory bus dynamics RAM (RDRAM) wait.

[0183] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0184] The embodiments described above are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application's patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. The embodiments disclosed above are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made according to the claims of this invention are still within the scope of this invention.

Claims

1. A method for automatically measuring the long jump, characterized in that, The method includes the following steps: Use video capture equipment to obtain long jump recordings; The long jump area is determined in the long jump video, and the long jump area contains several calibrated measurement points; The person to be tested is identified from the long jump video recording; The current state of the person being tested is determined based on their current location and posture, specifically including: When the person to be tested is first detected in the long jump video, their current status is determined to be that of a tourist and it is determined whether the person to be tested has entered the long jump preparation area. If so, change the status of the person to be tested to the ready status; When the posture of the person to be tested matches the take-off posture, their current state is changed to the flight state; Start the timing function until the person to be tested lands, then change their current status to the measurement status. The start-up timing function, which continues until the person to be tested lands, also includes: Obtain the flight time of the person to be tested; Determine whether the flight time of the person to be tested meets the threshold range; If the conditions are met, the current state of the person to be tested is determined to be the measurement state; When the person to be tested is in a measurement state, acquire the person's performance measurement points; When the person to be tested is in a measurement state, determine whether the person to be tested meets the measurement requirements; If the conditions are met, the performance measurement points of the person to be detected are determined according to the global image segmentation algorithm; The long jump performance of the person being tested is determined based on the matching relationship between the performance measurement points and the digit values ​​of the measurement points; The measurement requirements specifically include: Obtain the horizontal and vertical velocities of the feet of the person to be detected in the current frame; The speed score of the key point is determined based on the horizontal and vertical speeds of the feet of the person being tested. Determine whether the vertical velocity of the feet of the person being tested is less than a first fixed value; If so, process the foot image of the person to be detected according to the foot segmentation algorithm to obtain the foot segmentation region; obtain the horizontal velocity and vertical velocity of the foot segmentation region; The velocity fraction of the segmented area is determined based on the horizontal and vertical velocities of the segmented area. The horizontal velocity of the subject's feet and the horizontal velocity of the segmented area of ​​the feet are measured. Weighted summation is performed to obtain the horizontal velocity score of the part to be detected. Obtain the geometric center point of the segmented region and calculate the velocity score of the geometric center point; The landing score is obtained by weighted summing of the velocity scores of the key points, the segmented regions, the horizontal velocity scores of the parts to be detected, and the geometric center point. When the landing score meets the scoring condition, the current frame is determined to be the landing frame; The long jump score of the person being tested is determined based on the aforementioned performance measurement points and several calibrated measurement points.

2. The method for automatically measuring the long jump according to claim 1, characterized in that, When the posture of the person to be tested matches the take-off posture, their current state is changed to a flight state, specifically including: Obtain the current body tilt angle, forward arm swing amplitude, leg extension score, and foot movement of the person being tested; The weighted sum of the body tilt angle, the amplitude of the forward arm swing, the score of the straight leg, and the amount of foot movement is calculated, and it is determined whether the take-off posture threshold is met. If the conditions are met, determine whether the person being tested is continuously exerting force based on the key posture points at the current moment. If so, determine whether the current posture of the person being tested matches the take-off posture.

3. The method for automatically measuring the long jump according to any one of claims 1 to 2, characterized in that, After determining the long jump area in the long jump video, the process further includes: Obtain the long jump area from the long jump video; The long jump area is divided to obtain several long jump area modules; The exchange matrix is ​​processed on the several long jump region modules to obtain several perspective transformation diagrams; The several perspective transformation images are stitched together to obtain the complete measurement area; Each pixel in the measurement area corresponds one-to-one with the actual calibrated measurement point; The actual long jump distance is determined based on the pixel distance in the measurement area.

4. A system for automatically measuring long jump, characterized in that, The system includes: a video acquisition unit, a search unit, a status determination unit, and a long jump performance acquisition unit; The video acquisition unit is used to acquire long jump recordings using video acquisition equipment; The search unit is used to determine the long jump area in the long jump video, the long jump area containing several calibrated measurement points, and to determine the person to be tested in the long jump video. The state determination unit is used to determine the current state of the person to be detected based on their current location and movement posture, specifically including: When the person to be tested is first detected in the long jump video, their current status is determined to be that of a tourist and it is determined whether the person to be tested has entered the long jump preparation area. If so, change the status of the person to be tested to the ready status; When the posture of the person to be tested matches the take-off posture, their current state is changed to the flight state; Start the timing function until the person to be tested lands, then change their current status to the measurement status. The start-up timing function, which continues until the person to be tested lands, also includes: Obtain the flight time of the person to be tested; Determine whether the flight time of the person to be tested meets the threshold range; If the conditions are met, the current state of the person to be tested is determined to be the measurement state; When the person to be tested is in a measurement state, acquire the person's performance measurement points; When the person to be tested is in a measurement state, determine whether the person to be tested meets the measurement requirements; If satisfied, the performance measurement points of the person to be tested are determined according to the global image segmentation algorithm; The long jump performance of the person being tested is determined based on the matching relationship between the performance measurement points and the digit values ​​of the measurement points; The measurement requirements specifically include: Obtain the horizontal and vertical velocities of the feet of the person to be detected in the current frame; The speed score of the key point is determined based on the horizontal and vertical speeds of the feet of the person being tested. Determine whether the vertical velocity of the feet of the person being tested is less than a first fixed value; If so, process the foot image of the person to be detected according to the foot segmentation algorithm to obtain the foot segmentation region; obtain the horizontal velocity and vertical velocity of the foot segmentation region; The velocity fraction of the segmented area is determined based on the horizontal and vertical velocities of the segmented area. The horizontal velocity of the feet of the person being tested and the horizontal velocity of the segmented area of ​​the feet are measured. Weighted summation is performed to obtain the horizontal velocity score of the part to be detected. Obtain the geometric center point of the segmented region and calculate the velocity score of the geometric center point; The landing score is obtained by weighted summing of the velocity scores of the key points, the segmented regions, the horizontal velocity scores of the parts to be detected, and the geometric center point. When the landing score meets the scoring condition, the current frame is determined to be the landing frame; The long jump performance acquisition unit is used to determine the long jump performance of the person to be tested based on the performance measurement points and several calibrated measurement points.

5. A readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 3.

6. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 3.

Images