Method for estimating longitudinal jump height

Abstract

The invention discloses a longitudinal jump height estimation method. The method comprises the following steps: acquiring multiple frames of images in a longitudinal jump process at a fixed point; the ground clearance moment of the moving individual is analyzed from the multiple frames of images, a first image collected at a preset first time point after the ground clearance moment of the moving individual is obtained, a second image collected at a preset second time point after the ground clearance moment of the moving individual is obtained, and the first image and the second image are each provided with a first key point of the moving individual; obtaining a first image height coordinate of the first key point on the first image and a second image height coordinate of the first key point on the second image, and calculating a first uplink distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate, and associating the vertical jump height of the moving individual according to the first uplink distance value. According to the method, the longitudinal jump height can be pre-estimated before the moving individual reaches the highest point, so that the aim of outputting the longitudinal jump result in real time or even in advance is fulfilled.

Description

Technical Field

[0001] The present invention relates to the technical field of distance measurement based on machine vision, and in particular to a method for estimating vertical jump height. Background Art

[0002] Vertical jump is one of the basic sports movements. Methods for testing vertical jump height include manually observing the vertical jump height, obtaining the vertical jump height through sensors, or analyzing the vertical jump height based on machine vision. The method of analyzing the vertical jump height based on machine vision has obvious advantages and saves manpower, that is, no manual observation is required, and the sports individual is unaware, so that the sports individual user experience is good. The existing method of obtaining the vertical jump height based on machine vision is based on the principle of obtaining multiple frames of images or video streams during the vertical jump process at a fixed point, analyzing the key frames when the sports individual is at the highest point in the multiple frames of images or video streams, and then calculating the vertical jump height based on the key frames. However, this method will not be able to meet some special scenarios. For example, when the sports individual performs continuous vertical jumps, if it is necessary to simulate the vertical jump process of the sports individual in real time, and continuously display the vertical jump process to the sports individual in a manner associated with each vertical jump height through the output terminal to increase interest. However, the above method of obtaining the vertical jump height has the following problems: since the moving individual is at the highest point during the key frame, the key frame can only be collected when the moving individual reaches the highest point, and it takes time to analyze the key frame, calculate the vertical jump height, there is a delay in signal transmission, and both simulated interaction and output require time. The combination of multiple factors will cause delays in interaction, affecting the user experience, and needs to be improved urgently. Summary of the invention

[0003] In order to solve the technical problem of insufficient accuracy in the background technology, the present invention proposes a method for estimating vertical jump height, comprising:

[0004] Fixed-point acquisition of multiple frames of images during vertical jump;

[0005] Analyze the time when the moving individual leaves the ground from the multiple frames of images, obtain a first image captured at a preset first time point after the moving individual leaves the ground, and obtain a second image captured at a preset second time point after the moving individual leaves the ground, wherein both the first image and the second image have a first key point of the moving individual;

[0006] Obtain the first image height coordinate of the first key point on the first image and the second image height coordinate of the first key point on the second image, calculate the first upward distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate, and associate the vertical jump height of the moving individual according to the first upward distance value.

[0007] The first key point is the center point of the two shoulders.

[0008] The vertical jump height of the individual according to the first upward distance value is specifically associated with:

[0009] A plurality of ascending distance value ranges are pre-divided, and a plurality of vertical jump height values ​​are corresponding to the plurality of ascending distance value ranges one by one;

[0010] The upward distance value range where the first upward distance value is located is analyzed and the vertical jump height value corresponding thereto is associated, and the vertical jump height value is the vertical jump height of the sports individual.

[0011] There is a second ascending distance value within any ascending distance value range, and the vertical jump height calculated by the second ascending distance value is the vertical jump height value corresponding to the ascending distance value range.

[0012] After associating the vertical jump height of the sports individual according to the upward distance value, the method further includes: performing real-time simulation output according to the vertical jump height of the sports individual.

[0013] Calculating the first upward distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate specifically includes:

[0014] Acquire a second proportional relationship between an image height of the reference target in the first image / the second image and a preset height of the reference target;

[0015] A first upward distance value is acquired according to a height difference between the first image height coordinate and the second image height coordinate and a second proportional relationship.

[0016] The reference target includes some individual motion features, and a preset height of the reference target is associated with a height of the reference target in a specified area in the image.

[0017] Analyze the moment when the moving individual leaves the ground from multiple frames of images, including:

[0018] By image analysis, the upward distance of the second key point of the moving individual between consecutive frame images is obtained in sequence;

[0019] When the uplink distance exceeds a preset value, the previous frame in the continuous frames is defined as a ground-leaving frame, and the acquisition time corresponding to the ground-leaving frame is the moment when the moving individual leaves the ground.

[0020] The second key point is the bone key point at the ankle.

[0021] The present invention provides a method for estimating vertical jump height, which can estimate the vertical jump height before the sports individual reaches the highest point, thereby achieving the purpose of outputting the vertical jump result immediately or even in advance. DETAILED DESCRIPTION

[0022] In order to solve the problem that the first landing point of the target object in the background technology has a large error, the present invention proposes a method for estimating the vertical jump height, comprising:

[0023] S1. Fixed-point acquisition of multiple frames of images during vertical jump. The multiple frames of images are specifically video streams acquired at fixed points.

[0024] S2. Analyze the moment when the moving individual leaves the ground from multiple frames of images.

[0025] Specifically include:

[0026] S21, sequentially obtain the upward distance of the second key point of the moving individual between consecutive frame images through image analysis. Specifically, the second key point is the bone key point at the ankle.

[0027] S22. When the uplink distance exceeds a preset value, the previous frame in the continuous frames is defined as a ground-leaving frame, and the acquisition time corresponding to the ground-leaving frame is the moment when the moving individual leaves the ground.

[0028] During vertical jumping, the feet of the moving individual will not move before leaving the ground. The second key point is, for example, the bone key point at the ankle. Of course, other points with the same attributes can also be used as the second key point. At this stage, the second key point has no displacement in the vertical direction. When the vertical displacement reaches the preset value for the first time between two consecutive frames, the previous frame can be defined as the moment of leaving the ground.

[0029] S3, obtaining a first image captured at a preset first time point after the moving individual leaves the ground, and obtaining a second image captured at a preset second time point after the moving individual leaves the ground.

[0030] Among them, both the first image and the second image have the first key point of the moving individual. The selection of the first key point depends on the moving individual during the upward movement. The upward movement distance of the first key point can be equivalent to the upward movement distance of the moving individual. For example, the center points of the two shoulders can be selected as the first key point. It should be pointed out that the preset first time point refers to a frame obtained M seconds after leaving the ground, and M is a preset known value. Similarly, the preset second time point refers to a frame obtained N seconds after leaving the ground, and N is a preset known value. M and N are not equal.

[0031] S4. Obtain a first image height coordinate of the first key point on the first image and a second image height coordinate of the first key point on the second image.

[0032] S5. Calculate a first upward distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate.

[0033] Specifically include:

[0034] S51, obtaining a first image height of a reference target in a specified area and associating a relative height of the reference target.

[0035] After the image acquisition terminal is arranged at a fixed point, the designated area will be calibrated within the acquisition area of ​​the image acquisition terminal, and the moving individual will start vertical jumping within the designated area. In this scheme, for objects in the designated area, the first proportional relationship between the image height and the actual height is assumed to be a fixed value. Since the first proportional relationship is related to the distance and angle between the object and the image acquisition terminal, the smaller the designated area, the more accurate the result. When the image acquisition terminal is arranged at a fixed point and the designated area is demarcated, the first proportional relationship can also be determined. Therefore, after obtaining the first image height of the reference target in the designated area and the known or pre-configured first proportional relationship, the actual height of the reference target can be associated. In addition, the first proportional relationship can be modified arbitrarily and the relative height of the reference target can be associated therefrom.

[0036] It should be pointed out that the reference target must have the following characteristics: it always maintains the same position as the moving individual, and in the perspective of the image acquisition terminal, the reference target itself does not deform in the height direction. Based on the above, the reference target at least contains some features of the moving individual, such as the head frame of the moving individual.

[0037] S52: Acquire a second image height of the reference target in the first image or the second image.

[0038] During a vertical jump of an individual, its horizontal displacement can be ignored.

[0039] S53: Calculate a second proportional relationship between the second image height and the reference target relative height.

[0040] S54, calculating the height difference between the first image height coordinate and the second image height coordinate.

[0041] S55, calculating a first ascending distance value of the moving individual between the first time point and the second time point according to the height difference and the second proportional relationship.

[0042] The above height difference is the height difference collected in the image, not the actual height difference.

[0043] The second image height is the image height of the reference target in the first image and the second image. Since the moving individual may jump out of the designated area during multiple vertical jumps, it is obvious that when the moving individual jumps out of the designated area, it will definitely get close to or away from the image acquisition terminal. The default first proportional relationship of the designated area is still used for calculation, which will definitely cause a large error. Moreover, as the moving individual deviates from the designated area, the error will become larger. In this embodiment, the default first proportional relationship is not directly used for calculation, but a reference target with the above-mentioned attributes is selected as a bridge for calculation. When the moving individual jumps out of the designated area and causes the distance between the moving individual and the image acquisition terminal to change, the second image height of the reference target will also increase or decrease, and the height difference between the first image height coordinate and the second image height coordinate will also change in equal proportion. Based on this equal proportion change, the actual or relative first upward distance value of the moving individual between the first time point and the second time point can be calculated by the second proportional relationship after the height difference changes.

[0044] S6. Associating the vertical jump height of the sports individual according to the first ascending distance value.

[0045] Specifically include:

[0046] S61, pre-dividing a plurality of ascending distance value ranges, and corresponding the plurality of ascending distance value ranges to a plurality of vertical jump height values ​​one by one.

[0047] There is a second ascending distance value within any ascending distance value range, and the vertical jump height calculated by the second ascending distance value is the vertical jump height value corresponding to the ascending distance value range.

[0048] The explanation of calculating the corresponding vertical jump height value through the second upward distance value is as follows:

[0049] After calculating the moment of liftoff, since the first time point and the second time point are known, the time difference between the first time point and the second time point can also be obtained. The second upward distance value is the upward distance value of the moving individual between the first time point and the second time point. After liftoff, the acceleration of the moving individual can be defined as the acceleration due to gravity, which is known. At the highest point, the speed of the moving individual is zero. Based on the above parameters, the vertical jump height value corresponding to the second upward distance value can be calculated.

[0050] S62, analyzing the upward distance value range where the first upward distance value is located and associating the vertical jump height value corresponding thereto, the vertical jump height value being the vertical jump height of the sports individual.

[0051] It is easy to see that the selection of the first image and the second image is related to the selection of the first time point and the second time point. After obtaining the first image and the second image, the vertical jump height of the moving individual can be estimated. The first image and the second image are both two frames after the moving individual leaves the ground, rather than the key frames in the background technology. Obviously, when the first image and the second image are collected, the occurrence time of the key frames in the background technology has not yet arrived, that is, the moving individual has not reached the highest point. This implementation method can already estimate the vertical jump height, which can save time, thereby achieving the purpose of outputting the vertical jump result instantly or even in advance.

[0052] S7. Perform real-time simulation output according to the vertical jump height of the sports individual.

[0053] In order to improve user experience, increase fun, and enhance the sports passion of individual athletes, simulated animation output can be performed to interact with individual athletes. For example, when outputting animation, when the vertical jump height reaches different upward distance value ranges, different rewards are fed back to the individual athletes. The higher the jump, the richer the reward. Since this implementation can realize instant or even advance output of vertical jump results, instant interaction is achieved to improve user experience.

[0054] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and inventive concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for estimating vertical jump height, characterized in that: include: Fixed-point acquisition of multiple frames of images during vertical jump; Analyze the time when the moving individual leaves the ground from the multiple frames of images, obtain a first image captured at a preset first time point after the moving individual leaves the ground, and obtain a second image captured at a preset second time point after the moving individual leaves the ground, wherein both the first image and the second image have a first key point of the moving individual; Obtain the first image height coordinate of the first key point on the first image and the second image height coordinate of the first key point on the second image, calculate the first upward distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate, and associate the vertical jump height of the moving individual according to the first upward distance value.

2. The method for estimating vertical jump height according to claim 1, characterized in that: The first key point is the center point of the two shoulders.

3. The method for estimating vertical jump height according to claim 1, characterized in that: The vertical jump height of the individual according to the first upward distance value is specifically associated with: A plurality of ascending distance value ranges are pre-divided, and a plurality of vertical jump height values ​​are corresponding to the plurality of ascending distance value ranges one by one; The upward distance value range where the first upward distance value is located is analyzed and the vertical jump height value corresponding thereto is associated, and the vertical jump height value is the vertical jump height of the sports individual.

4. The method for estimating vertical jump height according to claim 3, characterized in that: There is a second ascending distance value within any ascending distance value range, and the vertical jump height calculated by the second ascending distance value is the vertical jump height value corresponding to the ascending distance value range.

5. The method for estimating vertical jump height according to claim 1, characterized in that: After associating the vertical jump height of the sports individual according to the upward distance value, the method further includes: performing real-time simulation output according to the vertical jump height of the sports individual.

6. The method for estimating vertical jump height according to claim 1, characterized in that: Calculating the first upward distance value of the moving individual between the first time point and the second time point according to the first image height coordinate and the second image height coordinate specifically includes: Obtaining the first image height of the reference target in the specified area and associating the relative height of the reference target; Acquire a second image height of a reference target in the first image or the second image; Calculating a second proportional relationship between the height of the second image and the relative height of the reference target; Calculate the height difference between the first image height coordinate and the second image height coordinate; A first ascending distance value of the moving individual between the first time point and the second time point is calculated according to the height difference and the second proportional relationship.

7. The method for estimating vertical jump height according to claim 6, characterized in that: The reference target contains some individual characteristics of the movement.

8. The method for estimating vertical jump height according to claim 1, characterized in that: Analyze the moment when the moving individual leaves the ground from multiple frames of images, including: By image analysis, the upward distance of the second key point of the moving individual between consecutive frame images is obtained in sequence; When the uplink distance exceeds a preset value, the previous frame in the continuous frames is defined as a ground-leaving frame, and the acquisition time corresponding to the ground-leaving frame is the moment when the moving individual leaves the ground.

9. The method for estimating vertical jump height according to claim 1, characterized in that: The second key point is the bone key point at the ankle.