A motion trajectory capture and analysis system

Abstract

The present invention belongs to the field of trajectory analysis, and discloses a motion trajectory capture and analysis system, including a motion image acquisition device, a motion image noise reduction device and a motion image analysis device; the motion image acquisition device is used to shoot the sports personnel in the process of the sports personnel performing training movements to obtain multiple motion images; the motion image noise reduction device includes a grayscale module and a noise reduction module; the grayscale module is used to grayscale the motion image to obtain a grayscale image; the noise reduction module is used to perform noise reduction based on the similarity of the RGB color space to obtain a noise-reduced motion image; the motion image analysis device includes a recognition module and an analysis module; the recognition module is used to respectively obtain the joint points in each noise-reduced motion image; the analysis module is used to respectively obtain the motion trajectory of each joint point. The present invention can improve the accuracy of the position of the joint point finally detected.

Description

Technical Field

[0001] The present invention relates to the field of trajectory analysis, and in particular to a motion trajectory capture and analysis system. Background Art

[0002] Motion trajectory capture and analysis refers to the process of using a camera to capture multiple images of the entire action process at high speed while the athlete is performing training movements. The positions of the human body's joints (such as the elbows) are then obtained from each image. Finally, the coordinates of all the joints are connected to obtain the motion trajectory of the joints. By analyzing the motion trajectory, it is possible to determine whether the movements are standard, thereby improving the quality of the movements performed by the athlete.

[0003] In the process of identifying the obtained image to obtain the position of the joint point, it is usually necessary to first perform denoising on the image, and then input the image obtained after denoising to identify the joint point. In the process of denoising in the prior art, denoising is usually performed only based on grayscale values, without considering the original color information. In this way, when two pixels have similar grayscale values ​​in the grayscale image but differ greatly in color in the color image, they will be identified as similar pixels in the denoising process. Therefore, when denoising is performed based on the sliding window, the pixels in the sliding window are given too high a denoising weight, affecting the accuracy of the final denoising result. This will obviously result in the accuracy of the position of the joint point finally detected being not high enough. Summary of the invention

[0004] The purpose of the present invention is to disclose a motion trajectory capture and analysis system to solve the technical problems raised in the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] The present invention provides a motion trajectory capture and analysis system, comprising a motion image acquisition device, a motion image noise reduction device and a motion image analysis device;

[0007] The motion image acquisition device is used to photograph the athlete while the athlete is performing training movements to obtain a plurality of motion images;

[0008] The motion image noise reduction device includes a grayscale module and a noise reduction module;

[0009] The grayscale module is used to grayscale the motion image to obtain a grayscale image;

[0010] The noise reduction module is used to perform noise reduction processing on each pixel in the grayscale image using the following formula to obtain a noise-reduced motion image:

[0011]

[0012] represents the gray value of pixel b after noise reduction, and bu represents the radius centered at pixel b. The set of pixels within the range of represents the adaptive denoising radius of b, and Represent the grayscale values ​​of pixel b and i respectively, represents the first noise reduction parameter, represents the second noise reduction parameter, is the noise reduction weight, Indicates the similarity between pixel i and b in RGB color space; Represents the distance between pixel points i and b;

[0013] The motion image analysis device includes a recognition module and an analysis module;

[0014] The recognition module is used to obtain the joint points in each denoised motion image respectively;

[0015] The analysis module is used to obtain the motion trajectory of each joint point respectively.

[0016] Preferably, the motion image acquisition device includes an acceleration acquisition module, a calculation module and a shooting module;

[0017] The acceleration acquisition module is used to obtain the acceleration of the moving person;

[0018] The calculation module is used to periodically obtain the number of shots per second based on acceleration;

[0019] The shooting module is used to shoot the moving person based on the number of shots per second to obtain moving images.

[0020] Preferably, the acceleration acquisition module includes a wearable acceleration sensor;

[0021] The wearable acceleration sensor is worn by the athlete; after obtaining the acceleration, the wearable acceleration sensor transmits it to the computing module.

[0022] Preferably, the acceleration of the moving person includes any one of X-axis acceleration, Y-axis acceleration and Z-axis acceleration.

[0023] Preferably, periodically acquiring the number of shots per second based on the acceleration includes:

[0024] The number of shots per second is calculated every preset time period T. The calculation formula for the number of shots per second is as follows:

[0025]

[0026] Indicates the number of shots per second. Indicates the time interval The collection of accelerations obtained within, Indicates the time when the number of shots per second starts to be calculated. express The number of accelerations in express The acceleration in express The average value of acceleration in, spd represents the preset number.

[0027] Preferably, photographing the moving person based on the number of photographs per second to obtain the moving image includes:

[0028] After calculating the number of shots per second, the calculation module transmits it to the shooting module;

[0029] After receiving the number of shots per second, the shooting module shoots the moving person according to the received number of shots per second to obtain a moving image.

[0030] Preferably, grayscale processing is performed on the motion image to obtain a grayscale image, including:

[0031] For a pixel point d in a moving image, the grayscale value is calculated as follows:

[0032]

[0033] represents the gray value of pixel d, , and Respectively represent the first grayscale weight, the second grayscale weight and the third grayscale weight, , and are the values ​​of the red component, green component, and blue component of pixel d in the RGB color space respectively.

[0034] Preferably, the calculation formula of the first noise reduction parameter is:

[0035]

[0036] and They represent the maximum and minimum grayscale values ​​of the pixels in the grayscale image respectively.

[0037] Preferably, the calculation formula of the second noise reduction parameter is:

[0038]

[0039] len represents the length of the diagonal of the grayscale image; , line represents the number of columns of pixels in the grayscale image, and row represents the number of rows of pixels in the grayscale image.

[0040] Preferably, The calculation formula is:

[0041]

[0042] and They represent the red component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and They represent the green component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and They respectively represent the values ​​of the blue component of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image.

[0043] Beneficial effects:

[0044] After obtaining the motion image, the present invention not only considers the information in the grayscale image, but also introduces the similarity information in the RGB color space in the process of denoising the motion image. Therefore, when the grayscale of the pixel point in the window is similar to the pixel point to be denoised, but the color difference is far in the color image, the denoising weight of the pixel point in the window is reduced, so that the pixel point to be denoised can refer to more similarity information for denoising, and obtain a more accurate denoising result, thereby improving the accuracy of the position of the joint point finally detected. It allows coaches and other personnel to guide the movements of athletes based on more accurate movement trajectories of the joint points. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0046] Figure 1 It is a schematic diagram of a motion trajectory capture and analysis system of the present invention. DETAILED DESCRIPTION

[0047] 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.

[0048] The present invention provides a motion trajectory capture and analysis system, comprising a motion image acquisition device, a motion image noise reduction device and a motion image analysis device;

[0049] The motion image acquisition device is used to photograph the athlete while the athlete is performing training movements to obtain a plurality of motion images;

[0050] The motion image noise reduction device includes a grayscale module and a noise reduction module;

[0051] The grayscale module is used to grayscale the motion image to obtain a grayscale image;

[0052] The noise reduction module is used to perform noise reduction processing on each pixel in the grayscale image using the following formula to obtain a noise-reduced motion image:

[0053]

[0054] represents the gray value of pixel b after noise reduction, and bu represents the radius centered at pixel b. The set of pixels within the range of represents the adaptive denoising radius of b, and Represent the grayscale values ​​of pixel b and i respectively, represents the first noise reduction parameter, represents the second noise reduction parameter, is the noise reduction weight, Indicates the similarity between pixel i and b in RGB color space; Represents the distance between pixel points i and b;

[0055] The motion image analysis device includes a recognition module and an analysis module;

[0056] The recognition module is used to obtain the joint points in each denoised motion image respectively;

[0057] The analysis module is used to obtain the motion trajectory of each joint point respectively.

[0058] In the above analysis process, after obtaining the motion image, in the process of denoising the motion image, in addition to considering the information in the grayscale image, the similarity information in the RGB color space is additionally introduced. The denoising function of the present invention is composed of two parts. The left part of the plus sign is based on the similarity of the grayscale value and the distance between two pixels in the grayscale image for calculation, while the right part is calculated in the color image. If the difference between the grayscale values ​​of the pixels i and b is smaller, the distance is closer, and the similarity in the RGB color space is greater, the weight of the pixel i in the denoising process is greater; if the pixel i and b are closer, the similarity in the RGB color space is greater, and .... The greater the difference in gray value between b and i, the farther the distance is, the smaller the similarity in the RGB color space is, and the smaller the weight of the pixel i in the denoising process is; therefore, the present invention also introduces the similarity information of the motion image in the RGB color space into the denoising process, so that when the grayscale of the pixel in the window is similar to that of the pixel to be denoised, but in the color image, the color difference is far, the denoising weight of the pixel in the window is reduced, so that the pixel to be denoised can refer to more similarity information for denoising, and obtain a more accurate denoising result, thereby improving the accuracy of the position of the joint point finally detected. It allows coaches and other personnel to guide the movements of athletes based on more accurate motion trajectories of joint points.

[0059] Preferably, The calculation formula is:

[0060]

[0061] ehu represents the set of pixels in the 8-neighborhood of b, and aveg represents the average grayscale value of the pixels in ehu. The preset radius.

[0062] The sliding window of the present invention is different from the existing rectangular window. The window of the present invention is a circular window, so that the distance between the pixels in the window can be avoided to be too far apart, which is conducive to introducing more effective pixels for noise reduction. Because the distance between the pixels near the two end points of the diagonal of the rectangular window and the center of the window will be significantly greater than the remaining pixels, so when reducing noise, the reference is of little significance, but will reduce the speed of noise reduction. Therefore, by introducing a circular window, the present invention can effectively reduce the gap between the distance between the pixel and the center of the window. While controlling the number of pixels, the effect of noise reduction is improved.

[0063] The present invention uses an adaptive noise reduction radius when reducing noise. Therefore, when the grayscale value of the pixel points in the 8-neighborhood of the pixel point b changes more dramatically, the radius of the window can be made smaller, so that more original boundary details can be retained. When the grayscale value changes more slowly, a larger radius is used to improve the noise reduction effect.

[0064] Specifically, The value of is 10.

[0065] Preferably, the noise reduction weight is .

[0066] Preferably, the motion image acquisition device includes an acceleration acquisition module, a calculation module and a shooting module;

[0067] The acceleration acquisition module is used to obtain the acceleration of the moving person;

[0068] The calculation module is used to periodically obtain the number of shots per second based on acceleration;

[0069] The shooting module is used to shoot the moving person based on the number of shots per second to obtain moving images.

[0070] Specifically, existing shooting equipment generally adopts a fixed number of shots per second for shooting. This shooting method will obtain too many motion images when the movement of the athlete is relatively slow, which increases the calculation amount of the motion image noise reduction device and the motion image analysis device, thereby resulting in a slower speed of obtaining the motion trajectory. Therefore, the present invention can effectively reduce the number of motion images finally obtained by obtaining acceleration and calculating the number of shots per second based on acceleration, while avoiding affecting the accuracy of the finally obtained trajectory due to the small size of the obtained image.

[0071] Preferably, the acceleration acquisition module includes a wearable acceleration sensor;

[0072] The wearable acceleration sensor is worn by the athlete; after obtaining the acceleration, the wearable acceleration sensor transmits it to the computing module.

[0073] Specifically, the wearable acceleration sensor can be a sports bracelet, a foot ring, etc.

[0074] Preferably, the acceleration of the moving person includes any one of X-axis acceleration, Y-axis acceleration and Z-axis acceleration.

[0075] Specifically, the acceleration sensor generally obtains acceleration of the X-axis, Y-axis, and Z-axis simultaneously.

[0076] In addition, the acceleration of the moving person may also be the maximum value among the X-axis acceleration, the Y-axis acceleration, and the Z-axis acceleration.

[0077] Preferably, periodically acquiring the number of shots per second based on the acceleration includes:

[0078] The number of shots per second is calculated every preset time period T. The calculation formula for the number of shots per second is as follows:

[0079]

[0080] Indicates the number of shots per second. Indicates the time interval The collection of accelerations obtained within, Indicates the time when the number of shots per second starts to be calculated. express The number of accelerations in express The acceleration in express The average value of acceleration in, spd represents the preset number.

[0081] Specifically, the number of shots of the present invention can be calculated with the severity of the change in acceleration in the time interval. When the change is more severe, the probability of a large change in the movement of the athlete is greater, thereby increasing the number of shots per second to provide more position information for subsequent trajectory generation; when the change is more gradual, the probability of a large change in the movement of the athlete is smaller, thereby reducing the number of shots per second to avoid obtaining too many moving images. Therefore, the number of shots of the present invention can be adaptively changed with the change in acceleration, and the total number of moving images obtained can be reduced while maintaining the accuracy of the final trajectory, which is conducive to improving the efficiency of trajectory analysis of the movement of the athlete.

[0082] Preferably, the value of T is 2 seconds.

[0083] Specifically, the preset number is 100.

[0084] Preferably, photographing the moving person based on the number of photographs per second to obtain the moving image includes:

[0085] After calculating the number of shots per second, the calculation module transmits it to the shooting module;

[0086] After receiving the number of shots per second, the shooting module shoots the moving person according to the received number of shots per second to obtain a moving image.

[0087] The shooting module can acquire motion images based on the latest number of shots per second, thereby effectively controlling the number of acquired motion images.

[0088] Preferably, grayscale processing is performed on the motion image to obtain a grayscale image, including:

[0089] For a pixel point d in a moving image, the grayscale value is calculated as follows:

[0090]

[0091] represents the gray value of pixel d, , and Respectively represent the first grayscale weight, the second grayscale weight and the third grayscale weight, , and are the values ​​of the red component, green component, and blue component of pixel d in the RGB color space respectively.

[0092] Specifically, the first grayscale weight, the second grayscale weight and the third grayscale weight are respectively , and .

[0093] Preferably, the calculation formula of the first noise reduction parameter is:

[0094]

[0095] and They represent the maximum and minimum grayscale values ​​of the pixels in the grayscale image respectively.

[0096] The first noise reduction parameter of the present invention can change with the change of the grayscale value of the grayscale image, so as to further improve the noise reduction effect. Because if a fixed first noise reduction parameter is set, it is very likely that the first noise reduction parameter will be too large or too small. When the first noise reduction parameter is too large, pixels with large grayscale value differences may also be taken into account, resulting in excessive color mixing, and the image may lose color accuracy and richness. When the first noise reduction parameter is too small, it may only reduce the noise of pixels with very small grayscale value differences, and the noise may not be effectively removed, resulting in an unclear filtering effect.

[0097] Preferably, the calculation formula of the second noise reduction parameter is:

[0098]

[0099] len represents the length of the diagonal of the grayscale image; , line represents the number of columns of pixels in the grayscale image, and row represents the number of rows of pixels in the grayscale image.

[0100] The second noise reduction parameter of the present invention can change adaptively with the change of the number of rows and columns of the grayscale image, thereby avoiding the situation where the second noise reduction parameter is too large or too small.

[0101] If the second denoising parameter is too large, the filter will consider neighboring pixels in a larger range, which may blur details, make the image too smooth, and lose edge information. If it is too small, it is easy to only consider very local pixels, and effective denoising may not be possible, and the noise in the image may still be obvious.

[0102] Preferably, The calculation formula is:

[0103]

[0104] and They represent the red component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and They represent the green component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and They respectively represent the values ​​of the blue component of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image.

[0105] The similarity of the present invention is calculated from the differences of the three components respectively. The smaller the difference is, the greater the similarity is.

[0106] Preferably, obtaining the motion trajectory of each joint point separately includes:

[0107] For the Z-type joint points, the coordinates of the Z-type joint points of each motion image are marked in the same image;

[0108] All coordinates are connected in the order of the shooting time of the motion image to obtain the motion trajectory of the Z-type joint point.

[0109] Specifically, in the same image, the coordinates of the Z-type joint points in each motion image are sequentially connected in the order of shooting time of the motion images from early to late, so as to obtain the motion trajectory.

[0110] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A motion trajectory capture and analysis system, characterized in that: It includes a motion image acquisition device, a motion image noise reduction device and a motion image analysis device; The motion image acquisition device is used to photograph the athlete while the athlete is performing training movements to obtain a plurality of motion images; The motion image noise reduction device includes a grayscale module and a noise reduction module; The grayscale module is used to grayscale the motion image to obtain a grayscale image; The noise reduction module is used to perform noise reduction processing on each pixel in the grayscale image using the following formula to obtain a noise-reduced motion image: represents the gray value of pixel b after noise reduction, and bu represents the radius centered at pixel b. The set of pixels within the range of represents the adaptive denoising radius of b, and Represent the grayscale values ​​of pixel b and i respectively, represents the first noise reduction parameter, represents the second noise reduction parameter, is the noise reduction weight, Indicates the similarity between pixel i and b in RGB color space; Represents the distance between pixel points i and b; The motion image analysis device includes a recognition module and an analysis module; The recognition module is used to obtain the joint points in each denoised motion image respectively; The analysis module is used to obtain the motion trajectory of each joint point respectively; The calculation formula is: and They represent the red component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and They represent the green component values ​​of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image, respectively. and Respectively represent the values ​​of the blue component of the pixel with the same coordinates as i and the pixel with the same coordinates as b in the RGB color space in the moving image; The calculation formula is: ehu represents the set of pixels in the 8-neighborhood of b, aveg represents the average grayscale value of the pixels in ehu, The preset radius.

2. A motion trajectory capture and analysis system according to claim 1, characterized in that: The motion image acquisition device includes an acceleration acquisition module, a calculation module and a shooting module; The acceleration acquisition module is used to obtain the acceleration of the moving person; The calculation module is used to periodically obtain the number of shots per second based on acceleration; The shooting module is used to shoot the moving person based on the number of shots per second to obtain moving images.

3. A motion trajectory capture and analysis system according to claim 2, characterized in that: The acceleration acquisition module includes a wearable acceleration sensor; The wearable acceleration sensor is worn by the athlete; after obtaining the acceleration, the wearable acceleration sensor transmits it to the computing module.

4. A motion trajectory capture and analysis system according to claim 2, characterized in that: The acceleration of the moving person includes any one of X-axis acceleration, Y-axis acceleration, and Z-axis acceleration.

5. A motion trajectory capture and analysis system according to claim 4, characterized in that: Periodically obtain the number of shots per second based on acceleration, including: The number of shots per second is calculated every preset time period T. The calculation formula for the number of shots per second is as follows: Indicates the number of shots per second, Indicates the time interval The collection of accelerations obtained within, Indicates the time when the number of shots per second starts to be calculated. express The number of accelerations in express The acceleration in express The average value of acceleration in, spd represents the preset number.

6. A motion trajectory capture and analysis system according to claim 2, characterized in that: Based on the number of shots per second, the moving person is photographed to obtain moving images, including: After calculating the number of shots per second, the calculation module transmits it to the shooting module; After receiving the number of shots per second, the shooting module shoots the moving person according to the received number of shots per second to obtain a moving image.

7. The motion trajectory capture and analysis system according to claim 1, characterized in that: Grayscale processing is performed on the motion image to obtain a grayscale image, including: For a pixel point d in a moving image, the grayscale value is calculated as follows: represents the gray value of pixel d, , and Respectively represent the first grayscale weight, the second grayscale weight and the third grayscale weight, , and are the values ​​of the red component, green component, and blue component of pixel d in the RGB color space respectively.

8. The motion trajectory capture and analysis system according to claim 1, characterized in that: The calculation formula of the first noise reduction parameter is: and They represent the maximum and minimum grayscale values ​​of the pixels in the grayscale image respectively.

9. The motion trajectory capture and analysis system according to claim 1, characterized in that: The calculation formula of the second noise reduction parameter is: len represents the length of the diagonal of the grayscale image; , line represents the number of columns of pixels in the grayscale image, and row represents the number of rows of pixels in the grayscale image.

Images