Dynamic posture analysis method and system based on sports biomechanics

Abstract

The invention discloses a dynamic posture analysis method and system based on sports biomechanics, and relates to the technical field of biological action recognition, and the method comprises the following steps: obtaining image data in a scanning range, average blood oxygen content data in the scanning range and average temperature data in the scanning range; image data in a scanning range, average blood oxygen content data in the scanning range and average temperature data in the scanning range are preprocessed, contour scanning is conducted on the preprocessed image data, contour scanning evaluation is conducted on a contour scanning result, and human body contour judgment is conducted according to a contour scanning evaluation result. Accurate dynamic posture analysis can be provided by integrating image data, blood oxygen content and temperature data, human body actions can be captured and evaluated more accurately through the multi-dimensional data acquisition and analysis method, and the reliability and accuracy of analysis are improved especially in a complex environment.

Description

Technical Field

[0001] The present invention relates to the technical field of biological motion recognition, and in particular to a dynamic posture analysis method and system based on sports biomechanics. Background Art

[0002] Dynamic posture analysis achieves accurate measurement and analysis of human movements by capturing data such as the three-dimensional spatial position, movement, speed, joint angle and torque of organisms in motion. These systems can monitor and evaluate human movements in real time, provide real-time risk assessment and early warning of movements and postures, and support multiple data collection and encoding methods, including real-time data transmission and later data import and analysis.

[0003] Especially in the assessment of postural disorders caused by mismatch between functional postural behavior and sensation, in summary, the dynamic posture analysis system provides effective tools for multiple fields to improve health, safety and efficiency through its high-precision data acquisition, real-time monitoring, flexibility and technological innovation.

[0004] For example, the invention patent with publication number: CN111951241B discloses a method for measuring and displaying muscle deformation during the movement of aquatic animals. The method obtains a video of aquatic animal behavior, decomposes the captured animal behavior video into photos for use in arranging animal postures during CT tomography, accurately extracts characteristic points of a muscle through the anatomy and CT perspective of real animals, establishes a preliminary muscle characteristic point equation group, and adjusts the equation parameters according to the posture of aquatic animals in field videos to obtain information such as the three-dimensional structural changes of the muscle during the movement of aquatic animals and the continuous deformation of muscle morphology; this information has great reference value for biological and medical research. In addition, a three-dimensional model of a muscle can be calculated based on photos or videos, and then these three-dimensional muscle models can be superimposed on photos or real videos to obtain animal movement videos with muscle activity perspective effects.

[0005] For example, the invention patent with publication number: CN109741371A discloses a method for describing human motion based on inertial sensors, the steps of which are: collecting motion data through inertial sensors, defining a bone hierarchy, and saving it in the motion capture data BVH format; performing coordinate conversion on the BVH Euler angle data, characterizing each frame of motion of each limb in the form of relative world coordinates, and forming a data frame sequence to facilitate subsequent analysis of the spatial orientation of the motion; after obtaining the human posture frame sequence, the key frame is extracted using the key frame extraction technology; finally, the spatial orientation of the human motion is analyzed, and other motion feature parameters of the action key frame are calculated, a mapping between the limb coordinates and the spatial orientation of the human body is established, and a description language is determined to describe the subject content of the human motion; the three subject contents of time, space and force efficiency of the human motion process are described in detail through the action description language, which is conducive to the application of non-computer related personnel in the field of motion analysis and recognition, and provides assistance for human motion analysis and human motion biomechanics.

[0006] When performing dynamic posture detection and analysis, if there are other obstacles around the tester, the presence of the obstacles will be mistakenly identified as part of the tester himself, causing a deviation between the tester's dynamic posture data and the standard motion data, thereby misleading the judgment of professionals. Summary of the invention

[0007] Technical issues solved In view of the shortcomings of the prior art, the present invention provides a dynamic posture analysis method and system based on sports biomechanics, which solves the problem that obstacles around the subject are mistakenly identified as part of the subject himself, thereby causing deviations between the subject's dynamic posture data and standard motion data.

[0008] Technical Solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic posture analysis method and system based on sports biomechanics, comprising the following specific steps: step one: acquiring image data within a scanning range, average blood oxygen content data within a scanning range, and average temperature data within a scanning range, preprocessing the image data within the scanning range, average blood oxygen content data within the scanning range, and average temperature data within the scanning range, performing contour scanning on the preprocessed image data, and performing contour scanning evaluation on the contour scanning results; step two: judging the human body contour according to the contour scanning evaluation results, if it is judged that there is no obstacle contour within the scanning range, locating the human body contour joint points, and acquiring human body dynamic posture data, preprocessing the human body dynamic posture data, and performing contour scanning evaluation on the human body contour joint points; step two: judging the human body contour according to the contour scanning evaluation results, if it is judged that there is no obstacle contour within the scanning range, locating the human body contour joint points, and obtaining human body dynamic posture data, preprocessing the human body dynamic posture data, and performing contour scanning evaluation on the human body contour joint points; step three: judging the human body contour according to the contour scanning evaluation results, and locating the human body contour joint points, and locating the human body dynamic posture data, preprocessing the human body dynamic posture data, and performing contour scanning evaluation on the human body contour joint points; step three: judging the human body contour according to the contour scanning evaluation results, and locating the human body contour joint points, The preprocessed human body dynamic posture data is evaluated; Step three: If it is determined that there is an obstacle outline within the scanning range, the obstacle outline within the scanning range is shielded, the human body outline joints are located, and the human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; Step four: The human body motion trajectory is judged according to the evaluation results of the human body dynamic posture data. If it is determined that the actual motion trajectory of the human body conforms to the preset motion trajectory, the human body dynamic posture data is analyzed and fed back; Step five: If it is determined that the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is issued, and the human body dynamic posture data continues to be obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated.

[0010] Furthermore, in step one, by setting a camera, image data within the scanning range of the camera is obtained, the image data is denoised and sharpened, an image edge detection algorithm is used to perform edge detection on each frame of the processed image to obtain the object contour in a single frame image, an object contour overlap detection algorithm is used to detect the object contour overlap of the image data, a time series analysis is used to obtain the position value of the object contour in a single frame image, and a pixel tracking algorithm is used to track the position value of the object contour in a single frame image to obtain the object contour movement trajectory; the image data includes the number of image frames, the pixel coordinates of each frame image, the number of object contours in each frame image, the area of ​​each object contour in each frame image, and the area of ​​each object contour in each frame image. The distance from the center point of the body contour to the top endpoint and the distance from the center point of each object contour in each frame image to the bottom endpoint; use the object contour overlap detection algorithm to detect the object contour overlap of the image data; use the object contour segmentation algorithm to segment the overlapping object contours; set the sample area of ​​the human body contour; obtain the ratio of the upper half and the lower half of each object contour in each frame image by the ratio of the distance from the center point of each object contour in each frame image to the top endpoint and the distance from the center point of each object contour in each frame image to the bottom endpoint; obtain the tolerance area value of the human body contour of each object contour in each frame image by the absolute value of the difference between the area of ​​each object contour in each frame image and the sample area of ​​the human body contour ; The contour evaluation value of each object contour in each frame image is obtained according to the ratio value of the upper half and the lower half of the contour of each object in each frame image and the fault tolerance area value of the human body contour of each object contour in each frame image; by setting a blood oxygen sensor, the average blood oxygen content data within the scanning range is obtained, and the average blood oxygen content data within the scanning range is subjected to data fusion and data synchronization processing. The average blood oxygen content data within the scanning range includes the blood oxygen saturation in each frame image and the oxygen partial pressure in each frame image; the blood oxygen saturation of the human body sample is set, and the fault tolerance value of the human body blood oxygen saturation of each object contour in each frame image is obtained according to the absolute value of the difference between the blood oxygen saturation of each object contour in each frame image and the blood oxygen saturation of the human body sample; the human body sample is set Oxygen partial pressure, obtain the tolerance value of human body oxygen partial pressure of each object contour in each frame of image according to the absolute value of the difference between the oxygen partial pressure of each object contour in each frame of image and the oxygen partial pressure of the human sample; obtain the temperature in each frame of image within the scanning range by setting a temperature sensor, and visualize the temperature data; set the body temperature of the human sample, obtain the tolerance value of human body temperature of each object contour in each frame of image according to the absolute value of the difference between the temperature of each object contour in each frame of image and the body temperature of the human sample; obtain the human body evaluation value of each object contour in each frame of image according to the tolerance value of human blood oxygen saturation of each object contour in each frame of image, the tolerance value of human body oxygen partial pressure of each object contour in each frame of image, and the tolerance value of human body temperature of each object contour in each frame of image;The method for obtaining the human body contour evaluation value according to the contour evaluation value of each object contour in each frame image, the human body evaluation value of each object contour in each frame image, the number of object contours in each frame image and the number of image frames is as follows:; ;in represents the human body contour evaluation value, n represents the number of image frames, Represents the contour evaluation value in each frame image, Represents the human evaluation value in each frame image, Indicates the number of object contours in each frame.

[0011] Furthermore, the specific steps of using the image edge detection algorithm to perform edge detection on each frame of the processed image to obtain the object contour in the single frame image are: S1: using the edge detection algorithm in each frame of the image to convolve the image in the horizontal and vertical directions respectively to obtain the horizontal gradient value and the vertical gradient value; S2: using the square root of the sum of squares and the inverse tangent value to obtain the gradient size and direction of each pixel point; S3: in the gradient direction, only the local maximum value is retained for each pixel point; S4: setting a strong edge line threshold and a weak edge line threshold, when the horizontal gradient value and the vertical gradient value are both greater than the strong edge line threshold, the pixel point is retained, when the horizontal gradient value and the vertical gradient value are both greater than the weak edge line value and less than the strong edge line value, and the pixel point of the weak edge line value is adjacent to the pixel point of the strong edge line value, it is retained; S5: connecting the strong edge line pixel point and the weak edge pixel point to form the object contour.

[0012] Furthermore, the specific steps of detecting the object contour overlap of image data using the object contour overlap detection algorithm are as follows: S1: arranging the horizontal gradient value and the vertical gradient value of each frame of the image obtained by the image edge detection algorithm in order from small to large; S2: obtaining the local minimum gradient value, that is, the pixel point close to the edge pixel point, the pixel point has the smallest distance from the edge pixel point, and there are multiple local minimum gradient values ​​in each frame of the image; S3: traversing from the local minimum gradient value to the adjacent pixel point until encountering an adjacent pixel point whose gradient value is less than or equal to the local minimum gradient value, then determining that a segmentation point has been found; S4: if the local minimum gradient value of each pixel point determines that a segmentation point has been found, then connecting multiple segmentation points into a line to obtain a segmentation line of the overlapping contour; S5: continuing to execute steps S1 to S4 until the gradient values ​​in each frame of the image are traversed once.

[0013] Furthermore, in step 2, a human body contour threshold is set and compared with the human body contour evaluation value; when the human body contour evaluation value is less than or equal to the human body contour threshold, it is determined that there is no obstacle contour within the scanning range, the human body contour joint points are located, and human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; the step of evaluating the preprocessed human body dynamic posture data is: using a joint positioning algorithm to locate the human body contour joint points, and detecting the human body key points in the image in real time; using an image edge detection algorithm and a pixel tracking algorithm to identify and track human body movement postures in real time, obtain human body dynamic posture data, and perform feature extraction on the human body dynamic posture data, the human body dynamic posture data includes the pixel coordinates in each frame of the image and the total number of pixels in each frame of the image; the pixel coordinates in each frame of the image are obtained by an optical flow algorithm to obtain the moving direction of the pixel in each frame of the image, and the moving direction of the pixel in each frame of the image is calculated to obtain the angle of the pixel in each frame of the image; the dynamic posture value is obtained according to the angle of the pixel in each frame of the image.

[0014] Furthermore, the specific steps of obtaining the dynamic posture value according to the angle of the pixel point in each frame image are as follows: S1: divide 360 ​​degrees into ten angle intervals, each interval is 36 degrees; S2: the angle interval corresponding to the angle of each pixel point in each frame image is expressed by the formula ;in Indicates the angle interval corresponding to the angle of each pixel in each frame of the image. Represents the angle of each pixel in each frame of the image. It means that the angle of each pixel in each frame of the image is rounded down to the corresponding complete angle interval. Since the numbering starts from 1, 1 should be added; S3: Count the number of pixels in each angle interval in each frame of the image, and then obtain the angle ratio of the pixels in each angle interval in each frame of the image, which is expressed by the formula: ;in Indicates the angle ratio of pixels in each angle interval in each frame image. Indicates the number of pixels in each angle interval in each frame of the image. Indicates the total number of pixels in each frame; S4: Arrange the angle proportions of the pixels in each angle interval of each of the ten angle intervals in each frame in order from large to small; S5: Take the angle interval with the largest proportion in each frame, express the angle of each pixel in these angle intervals as the optimized angle and sum them up to obtain the dynamic posture value; The dynamic posture value acquisition method is: set the dynamic posture sample value, and obtain the dynamic posture evaluation value according to the dynamic posture value and the dynamic posture sample value as follows: ;in represents the dynamic posture evaluation value, represents the dynamic posture value, Represents the dynamic pose sample value.

[0015] Furthermore, in step three, when the human body contour evaluation value is greater than the human body contour threshold, redundant object contours are masked.

[0016] Furthermore, in step four, a dynamic posture threshold is set and compared with the dynamic posture evaluation value; when the dynamic posture evaluation value is less than the dynamic posture threshold, it is determined that the actual motion trajectory of the human body meets the preset motion trajectory, and dynamic posture detection continues.

[0017] Furthermore, in step five, when the dynamic posture evaluation value is greater than or equal to the dynamic posture threshold, the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is given, the human body dynamic posture data continues to be acquired, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated.

[0018] Furthermore, there are a data acquisition module, a data analysis module and a data execution module; the data acquisition module is used to acquire image data, average blood oxygen content data within a scanning range and average temperature data within a scanning range and perform preprocessing, and send the data to the data execution module for contour scanning. When the data execution module locates the joints of the human body contour, it acquires dynamic posture data and performs preprocessing, and sends the dynamic posture data to the data analysis module; the data analysis module is used to analyze the results of the contour scanning and judge the contour, and continue to send the judgment result to the data execution module for result execution, until the data execution module locates the joints of the human body contour, and when the data analysis module acquires the dynamic posture data, it performs action consistency judgment and sends the judgment result to the data execution module; the data execution module executes human body contour judgment and action consistency judgment according to the judgment result of the data analysis module.

[0019] Beneficial Effects

[0020] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: 1. By integrating image data, blood oxygen content and temperature data, it can provide accurate dynamic posture analysis. This multi-dimensional data acquisition and analysis method can more accurately capture and evaluate human movements, especially in complex environments, improving the reliability and accuracy of the analysis.

[0021] 2. Through advanced image processing technology, it can effectively distinguish the outline of the human body and the outline of obstacles, solving the problem of obstacles around the detector being misidentified.

[0022] 3. Able to monitor and evaluate the consistency of human movements in real time, provide instant feedback, and ensure the correctness and effectiveness of the movements.

[0023] 4. When it is detected that the actual movement trajectory of the human body does not match the preset trajectory, the system will issue an early warning prompt in time. This is particularly important in the fields of sports training, safety monitoring and health monitoring, and helps prevent sports injuries and improve training effects.

[0024] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This invention is a flow chart of a dynamic posture analysis method based on sports biomechanics.

[0026] Figure 2 This is the present invention: a structural diagram of a dynamic posture analysis system based on sports biomechanics. DETAILED DESCRIPTION

[0027] 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.

[0028] It should be noted that, in this article, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0029] like Figure 1 As shown, the embodiment of the present invention provides a dynamic posture analysis method based on sports biomechanics, which includes the following specific steps: Step 1: acquiring image data within a scanning range, average blood oxygen content data within a scanning range, and average temperature data within a scanning range, preprocessing the image data within the scanning range, the average blood oxygen content data within the scanning range, and the average temperature data within the scanning range, performing contour scanning on the preprocessed image data, and performing contour scanning evaluation on the contour scanning results; By setting a camera, image data within the camera scanning range is obtained, the image data is denoised and sharpened, and the edge detection algorithm is used to detect the edge of each frame of the processed image to obtain the object contour in the single frame image, and the object contour overlap detection algorithm is used to detect the object contour overlap of the image data, and the position value of the object contour in the single frame image is analyzed by time series, and then the pixel tracking algorithm is used to track the position value of the object contour in the single frame image to obtain the object contour movement trajectory; The image data includes the number of image frames, the pixel coordinates of each frame, the number of object contours in each frame, the area of ​​each object contour in each frame, the distance from the center point to the top endpoint of each object contour in each frame, and the distance from the center point to the bottom endpoint of each object contour in each frame; Perform denoising and sharpening on the image data. Denoising can reduce random noise in the image, making the image smooth and clear, while sharpening can enhance the edges and details in the image, making the image outline clear. Image edge detection algorithms include the Sobel edge detection algorithm, which obtains a 3x3 matrix by performing convolution in the horizontal and vertical directions of the image, calculates the gradient size and direction of each pixel in the image, and then determines the edge position based on the size and direction of the gradient value; The specific steps of using the Sobel edge detection algorithm to perform edge detection on each frame of the processed image and obtain the object contour in a single frame image are as follows: S1: In each frame of the image, the edge detection algorithm is used to convolve the image in the horizontal and vertical directions to obtain the horizontal gradient value and the vertical gradient value; S2: The horizontal gradient value and the vertical gradient value are respectively obtained by the square root of the sum of squares and the inverse tangent value to obtain the gradient magnitude and direction of each pixel; S3: In the gradient direction, only the local maximum value is retained for each pixel, non-edge responses are suppressed, and non-edge detection is reduced; S4: Set the strong edge threshold and the weak edge threshold. When the horizontal gradient value and the vertical gradient value are both greater than the strong edge threshold, the pixel point is retained. When the horizontal gradient value and the vertical gradient value are both greater than the weak edge value and less than the strong edge value, and the pixel point with the weak edge value is adjacent to the pixel point with the strong edge value, the pixel point is retained. S5: Connect the strong edge pixels and the weak edge pixels to form the object contour; Detecting the overlapping of object contours of image data using an object contour overlapping detection algorithm including a minimum circumscribed circle contour overlapping detection algorithm; The specific steps of using the object contour overlap detection algorithm to detect the object contour overlap of image data are as follows: S1: Calculate the minimum circumscribed circle of each object contour, and construct a triangle through three edge pixels on the object contour that are not on the same straight line. The pixels that make up the object contour are the edge pixels. S2: Calculate the intersection of the perpendicular bisectors of the three sides of the triangle to get the center of the circle, that is, the circle is tangent to the three sides of the triangle, and the intersection of the perpendicular bisectors of the three sides of the triangle is the center of the circle; S3: The distances from the center of the circle to the three sides of the triangle are equal. This circle is the circumscribed circle. Continue to perform steps S1 to S2 until the minimum distance from the center of the circle to the three sides of the triangle is obtained, which is the minimum circumscribed circle. S4: Compare the distances between the centers of the minimum circumscribed circles of the contours of each object, and use the Euclidean distance formula to obtain the distances between the centers of the minimum circumscribed circles of the contours of each object; S5: When the distance between the centers of the minimum circumscribed circles of the two object contours is less than or equal to the sum of the radii of the two object contours, it is determined that the two object contours overlap, and steps S4 to S5 are continued until the contours of each object in each frame image are compared; S6: Continue to execute steps S1 to S5 until the contours of each object in each frame image are compared; S7: using an object contour segmentation algorithm to segment the overlapping object contours in each frame of the image; The specific steps of using the object contour segmentation algorithm including the watershed algorithm to segment the overlapping object contours are as follows: S1: The horizontal gradient value and the vertical gradient value of each frame of the image obtained by the above-mentioned image edge detection algorithm including the Sobel edge detection algorithm are arranged in order from small to large; S2: Get the local minimum gradient value, that is, the pixel point close to the edge pixel point, which has the shortest distance from the edge pixel point. There are multiple local minimum gradient values ​​in each frame image. S3: Traverse from the local minimum gradient value to the adjacent pixel points. The gradient values ​​of the adjacent pixel points are all greater than the local minimum gradient value. Until the adjacent pixel point gradient value is less than or equal to the local minimum gradient value, the segmentation point is found. S4: Multiple local minimum gradient values ​​are found until multiple segmentation points are found. Connecting multiple segmentation points into a line is the segmentation line of the overlapping contours. S5: Continue to execute steps S1 to S4 until the overlapping object contours in each frame image are segmented; In time series analysis, a single frame image is a single time point in the time series, and the data value at the time point is the position value of the pixel point of each object contour in each frame image; The pixel tracking algorithm includes an optical flow algorithm, which tracks the motion trajectory of each object's outline by calculating the position value changes of the pixel points of each object's outline in each frame of the image; Set the sample area of ​​the human body contour, and obtain the sample area of ​​the human body contour by calling the body shape data of each human body from the human body shape database and calculating the average value; The center point of each object contour in each frame image represents the coordinates of all pixels in the contour, which are accumulated on the x-axis and y-axis, and then divided by the number of pixels in the contour to obtain the center point of the contour; The top endpoint and the bottom endpoint of the object contour are the pixel point with the largest y-axis coordinate and the pixel point with the smallest y-axis coordinate in the contour respectively; The ratio of the distance from the center point of each object contour in each frame image to the top endpoint to the distance from the center point of each object contour in each frame image to the bottom endpoint represents the ratio of the upper half to the lower half of each object contour in each frame image. The method for obtaining the ratio is as follows: ; in Indicates the ratio of the upper and lower parts of the contour of each object in each frame of the image. Indicates the distance from the center point to the top endpoint of the outline of each object in each frame of the image. Indicates the distance from the center point to the bottom endpoint of the outline of each object in each frame of the image; The absolute value of the difference between the area of ​​each object contour in each frame image and the sample area of ​​the human body contour represents the error tolerance area value of the human body contour of each object contour in each frame image. The method for obtaining the error tolerance area value of the human body contour is as follows: ; in Represents the error tolerance area value of the human body contour of each object contour in each frame image, Represents the area of ​​each object contour in each frame of the image, The sample area representing the human body outline; According to the ratio of the upper half to the lower half of the contour of each object in each frame image and the tolerance area value of the human body contour of each object contour in each frame image, the contour evaluation value of each object contour in each frame image is obtained as follows: ; in Represents the contour evaluation value of each object contour in each frame image, Indicates the ratio of the upper and lower parts of the contour of each object in each frame of the image. Represents the error tolerance area value of the human body contour of each object contour in each frame of the image; By setting a blood oxygen sensor, average blood oxygen content data within a scanning range is obtained, and data fusion and data synchronization processing are performed on the average blood oxygen content data within the scanning range, where the average blood oxygen content data within the scanning range includes blood oxygen saturation in each frame image and oxygen partial pressure in each frame image; Data fusion means integrating data from different sources, formats and structures into one data set, which helps reduce errors and uncertainties in the data and improve the reliability of data analysis; Data synchronization means updating and matching data in two or more different locations or systems in real time or regularly to maintain data consistency, which helps reduce the cost of data management and maintenance by reducing duplicate data entry and errors. The maintenance of blood oxygen saturation and oxygen partial pressure ensures that cells obtain oxygen for aerobic metabolism and generates energy, thus providing protection for the normal movement of muscles; Set the blood oxygen saturation of human samples, and obtain the blood oxygen saturation of human samples by calling the blood oxygen saturation data of each human body from the human blood oxygen saturation database and calculating the average value; The absolute value of the difference between the blood oxygen saturation of each object contour in each frame image and the blood oxygen saturation of the human sample represents the error tolerance value of the human blood oxygen saturation of each object contour in each frame image. The method for obtaining the error tolerance value is as follows: ; in Indicates the tolerance value of human blood oxygen saturation of each object contour in each frame of the image, Indicates the blood oxygen saturation of each object contour in each frame of the image, Indicates the blood oxygen saturation of human samples; Set the oxygen partial pressure of the human sample, and obtain the oxygen partial pressure of the human sample by calling the oxygen partial pressure data of each human body from the human oxygen partial pressure database and calculating the average value; The absolute value of the difference between the oxygen partial pressure of each object contour in each frame image and the oxygen partial pressure of the human body sample represents the error tolerance value of the human body oxygen partial pressure of each object contour in each frame image. The method for obtaining the error tolerance value is as follows: ; in Indicates the tolerance value of human oxygen partial pressure of each object contour in each frame of the image, Represents the oxygen partial pressure of each object contour in each frame of the image, Indicates the oxygen partial pressure of human sample; By setting a temperature sensor, the average temperature data within the scanning range is obtained, and the temperature data is visualized. The average temperature data within the scanning range includes the temperature in each frame of the image; Visual processing means converting data into an intuitive and easy-to-understand form through charts, graphs, etc., which improves the efficiency of information transmission; Set the body temperature of human samples, and obtain the body temperature of human samples by calling the data of normal body temperature and exercise body temperature of each human body from the human body temperature database and calculating the average value; The absolute value of the difference between the temperature of each object contour in each frame image and the body temperature of the human sample represents the error tolerance value of the human body temperature of each object contour in each frame image. The method for obtaining the error tolerance value is as follows: ; in Indicates the tolerance value of the human body temperature of each object contour in each frame of the image, Indicates the temperature of each object contour in each frame of the image, Indicates the body temperature of human samples; According to the error tolerance value of human blood oxygen saturation of each object contour in each frame image, the error tolerance value of human oxygen partial pressure of each object contour in each frame image, and the error tolerance value of human body temperature of each object contour in each frame image, the method for obtaining the human body evaluation value of each object contour in each frame image is as follows: ; in Represents the human body evaluation value of each object contour in each frame image, Indicates the tolerance value of human blood oxygen saturation of each object contour in each frame of the image, Indicates the tolerance value of human oxygen partial pressure of each object contour in each frame of the image, The tolerance value of human body temperature representing the contour of each object in each frame of the image; The method for obtaining the human body contour evaluation value according to the contour evaluation value of each object contour in each frame image, the human body evaluation value of each object contour in each frame image, the number of object contours in each frame image and the number of image frames is as follows: ; in represents the human body contour evaluation value, =1,2,3…n, n represents the number of image frames, Represents the contour evaluation value in each frame image, Represents the human evaluation value in each frame image, Indicates the number of object contours in each frame; Step 2: judging the human body contour according to the contour scanning evaluation result. If it is judged that there is no obstacle contour within the scanning range, locating the joint points of the human body contour, obtaining the human body dynamic posture data, preprocessing the human body dynamic posture data, and evaluating the preprocessed human body dynamic posture data; Setting a human body contour threshold and comparing it with the human body contour evaluation value, the human body contour threshold is obtained from the laboratory and stored in the database, and then called from the database; When the human body contour evaluation value is less than or equal to the human body contour threshold, there is no obstacle contour within the scanning range, the human body contour joint points are located, and the human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; Use joint positioning algorithms including the OpenPose algorithm to locate the joint points of the human body contour, and use deep learning-based human posture technology to detect human key points in the image in real time, including posture information of the body, face, and hands; The Sobel edge detection algorithm and the optical flow algorithm are used to identify and track human body movements in real time, obtain dynamic posture data, and extract features from the dynamic posture data. The dynamic posture data includes the coordinates of the pixels in each frame and the total number of pixels in each frame. By extracting features, the dimension of dynamic posture data is reduced, which improves the efficiency of Sobel edge detection algorithm and optical flow algorithm. The coordinates of the pixels in each frame are obtained by the optical flow algorithm to obtain the moving direction of the pixels in each frame, and the moving direction of the pixels in each frame is calculated to obtain the angle of the pixels in each frame. The angle of the pixel in each frame is the angle between the ray from the position of the pixel in the previous frame to the position of the pixel in the next frame and the horizontal axis; Using statistical principles, 360 degrees are evenly divided into ten angle intervals, and the distribution of the angles of the pixels in each frame in the ten angle intervals is counted. The distribution proportions of the ten angle intervals are arranged in descending order, and the angle interval with the largest proportion in each frame image is summed to obtain the dynamic posture value; The specific steps to obtain the dynamic posture value are: S1: Divide 360 ​​degrees into ten angle intervals, each interval is 36 degrees; S2: The angle interval corresponding to the angle of each pixel in each frame image is expressed by the formula ;in Indicates the angle interval corresponding to the angle of each pixel in each frame of the image. Represents the angle of each pixel in each frame of the image. It means that the angle of each pixel in each frame of the image is rounded down to the corresponding full angle range. Since the numbering starts from 1, 1 should be added; S3: Count the number of pixels in each angle interval in each frame image, and then obtain the angle ratio of the pixels in each angle interval in each frame image, which is expressed by the formula: ;in Indicates the angle ratio of pixels in each angle interval in each frame image. Indicates the number of pixels in each angle interval in each frame of the image. Indicates the total number of pixels in each frame of the image; S4: arranging the angle proportions of the pixels in each angle interval of each of the ten angle intervals in each frame of the image in descending order; S5: Take the angle interval with the largest proportion in each frame image, express the angle of each pixel point in each frame image within these angle intervals as the optimized angle, and sum them to obtain the dynamic posture value as follows: ; in represents the dynamic posture value, =1,2,3…n, n represents the number of image frames, represents the optimization angle; Set the dynamic posture sample value, and obtain the dynamic posture sample value by calling the data of each human body dynamic posture from the human body dynamic posture database and calculating the average value; The method of obtaining the dynamic posture evaluation value based on the dynamic posture value and the dynamic posture sample value is as follows: ; in represents the dynamic posture evaluation value, represents the dynamic posture value, Represents the dynamic posture sample value; Step 3: If it is determined that there is an obstacle outline within the scanning range, the obstacle outline within the scanning range is shielded, the joint points of the human body outline are located, and the human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; When the human body contour evaluation value is greater than the human body contour threshold, a mask is created to cover the contour whose human body contour evaluation value is greater than the human body contour threshold. When the value of the pixel point corresponding to the contour of the mask is 0, the contour is shielded; Step 4: judging the human body motion trajectory according to the human body dynamic posture data evaluation results. If it is judged that the actual motion trajectory of the human body meets the preset motion trajectory, the human body dynamic posture data is analyzed and fed back; Set a dynamic posture threshold and compare it with the dynamic posture evaluation value. The dynamic posture threshold is obtained from the experiment and stored in the database, and historical data is called from it; When the dynamic posture evaluation value is less than the dynamic posture threshold, the actual motion trajectory of the human body conforms to the preset motion trajectory, and the dynamic posture of the human body is praised; Step 5: If it is determined that the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is issued, and the human body dynamic posture data continues to be acquired, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; When the dynamic posture evaluation value is greater than or equal to the dynamic posture threshold, the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is given, and the standard dynamic posture is played once. The human body dynamic posture data continues to be obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated.

[0030] like Figure 2 As shown: A dynamic posture analysis system based on sports biomechanics, including: a data acquisition module, a data analysis module and a data execution module; The data acquisition module is used to acquire image data, average blood oxygen content data within the scanning range, and average temperature data within the scanning range and perform preprocessing, and send the data to the data execution module for contour scanning. When the data execution module locates the joint points of the human body contour, it acquires dynamic posture data and preprocesses it, and sends the dynamic posture data to the data analysis module; The data analysis module is used to analyze the results of the contour scanning and judge the contour, and send the judgment results to the data execution module for result execution until the data execution module locates the joint points of the human body contour. When the data analysis module obtains the dynamic posture data, it judges the consistency of the action and sends the judgment results to the data execution module; The data execution module performs human body contour judgment and action consistency judgment according to the judgment results of the data analysis module.

[0031] The preferred embodiments of the present invention disclosed above are only used to help explain 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 dynamic posture analysis method based on sports biomechanics, characterized by: The specific steps include: Step 1: acquiring image data within a scanning range, average blood oxygen content data within a scanning range, and average temperature data within a scanning range, preprocessing the image data within the scanning range, the average blood oxygen content data within the scanning range, and the average temperature data within the scanning range, performing contour scanning on the preprocessed image data, and performing contour scanning evaluation on the contour scanning results; Step 2: judging the human body contour according to the contour scanning evaluation result. If it is judged that there is no obstacle contour within the scanning range, locating the joint points of the human body contour, obtaining the human body dynamic posture data, preprocessing the human body dynamic posture data, and evaluating the preprocessed human body dynamic posture data; Step 3: If it is determined that there is an obstacle outline within the scanning range, the obstacle outline within the scanning range is shielded, the joint points of the human body outline are located, and the human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; Step 4: judging the human body motion trajectory according to the human body dynamic posture data evaluation results. If it is judged that the actual motion trajectory of the human body meets the preset motion trajectory, the human body dynamic posture data is analyzed and fed back; Step 5: If it is determined that the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is issued, the human body dynamic posture data continues to be acquired, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated.

2. The method for dynamic posture analysis based on sports biomechanics according to claim 1, characterized in that: In step one, by setting a camera, image data within the camera scanning range is obtained, the image data is denoised and sharpened, an image edge detection algorithm is used to detect the edge of each frame of the processed image to obtain the object contour in the single frame image, an object contour overlap detection algorithm is used to detect the object contour overlap of the image data, a time series analysis is used to obtain the position value of the object contour in the single frame image, and a pixel tracking algorithm is used to track the position value of the object contour in the single frame image to obtain the object contour movement trajectory; The image data includes the number of image frames, the pixel coordinates of each frame, the number of object contours in each frame, the area of ​​each object contour in each frame, the distance from the center point to the top endpoint of each object contour in each frame, and the distance from the center point to the bottom endpoint of each object contour in each frame; Using an object outline overlap detection algorithm to detect the object outline overlap of image data; Use the object contour segmentation algorithm to segment the overlapping object contours; Set the sample area of ​​the human body contour; The ratio of the upper half to the lower half of the contour of each object in each frame image is obtained by the ratio of the distance from the center point of the contour of each object in each frame image to the top end point to the distance from the center point of the contour of each object in each frame image to the bottom end point; The error tolerance area value of the human body contour of each object contour in each frame image is obtained by the absolute value of the difference between the area of ​​each object contour in each frame image and the sample area of ​​the human body contour; Obtaining a contour evaluation value of each object contour in each frame of image according to a ratio value of the upper half to the lower half of each object contour in each frame of image and a tolerance area value of the human body contour of each object contour in each frame of image; By setting a blood oxygen sensor, average blood oxygen content data within a scanning range is obtained, and data fusion and data synchronization processing are performed on the average blood oxygen content data within the scanning range, where the average blood oxygen content data within the scanning range includes blood oxygen saturation in each frame image and oxygen partial pressure in each frame image; The blood oxygen saturation of the human sample is set, and the tolerance value of the human blood oxygen saturation of each object contour in each frame image is obtained according to the absolute value of the difference between the blood oxygen saturation of each object contour in each frame image and the blood oxygen saturation of the human sample; The oxygen partial pressure of the human body sample is set, and the tolerance value of the human body oxygen partial pressure of each object contour in each frame image is obtained according to the absolute value of the difference between the oxygen partial pressure of each object contour in each frame image and the oxygen partial pressure of the human body sample; By setting a temperature sensor, the temperature in each frame of the image within the scanning range is obtained, and the temperature data is visualized; Set the body temperature of the human sample, and obtain the tolerance value of the human body temperature of each object contour in each frame of the image according to the absolute value of the difference between the temperature of each object contour in each frame of the image and the body temperature of the human sample; Obtaining a human evaluation value for each object contour in each frame of image according to a human blood oxygen saturation tolerance value for each object contour in each frame of image, a human oxygen partial pressure tolerance value for each object contour in each frame of image, and a human body temperature tolerance value for each object contour in each frame of image; The method for obtaining the human body contour evaluation value according to the contour evaluation value of each object contour in each frame image, the human body evaluation value of each object contour in each frame image, the number of object contours in each frame image and the number of image frames is as follows: ； in represents the human body contour evaluation value, n represents the number of image frames, Represents the contour evaluation value in each frame image, Represents the human evaluation value in each frame image, Indicates the number of object contours in each frame.

3. A dynamic posture analysis method based on sports biomechanics according to claim 2, characterized in that: The specific steps of performing edge detection on each processed image frame using an image edge detection algorithm to obtain the contour of the object in a single frame image are as follows: S1: In each frame of the image, the edge detection algorithm is used to convolve the image in the horizontal and vertical directions to obtain the horizontal gradient value and the vertical gradient value; S2: The horizontal gradient value and the vertical gradient value are respectively obtained by the square root of the sum of squares and the inverse tangent value to obtain the gradient magnitude and direction of each pixel; S3: In the gradient direction, only the local maximum value is retained for each pixel; S4: Set a strong edge threshold and a weak edge threshold. When both the horizontal gradient value and the vertical gradient value are greater than the strong edge threshold, the pixel point is retained. When both the horizontal gradient value and the vertical gradient value are greater than the weak edge value and less than the strong edge value, and the pixel point with the weak edge value is adjacent to the pixel point with the strong edge value, the pixel point is retained. S5: Connect the strong edge pixels and the weak edge pixels to form the object contour.

4. The method for dynamic posture analysis based on sports biomechanics according to claim 2, characterized in that: The specific steps of detecting the object outline overlap situation of the image data by using the object outline overlap detection algorithm are as follows: S1: The horizontal gradient value and the vertical gradient value of each frame of the image obtained by the image edge detection algorithm are arranged in order from small to large; S2: Get the local minimum gradient value, that is, the pixel point close to the edge pixel point, which has the shortest distance from the edge pixel point. There are multiple local minimum gradient values ​​in each frame image. S3: Traverse from the local minimum gradient value to the adjacent pixel points until the adjacent pixel point gradient value is less than or equal to the local minimum gradient value, then it is determined that the segmentation point is found; S4: If the local minimum gradient value of each pixel point determines that a segmentation point is found, multiple segmentation points are connected into a line to obtain a segmentation line of the overlapping contour; S5: Continue to execute steps S1 to S4 until the gradient values ​​in each frame image are traversed.

5. The method for dynamic posture analysis based on sports biomechanics according to claim 1, characterized in that: In step 2, a human body contour threshold is set and compared with the human body contour evaluation value; When the human body contour evaluation value is less than or equal to the human body contour threshold, it is determined that there is no obstacle contour within the scanning range, the human body contour joint points are located, and the human body dynamic posture data is obtained, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated; The steps of evaluating the preprocessed human body dynamic posture data are as follows: Use joint positioning algorithm to locate the joint points of human body contour and detect the key points of human body in the image in real time; The image edge detection algorithm and the pixel tracking algorithm are used to identify and track human body movements and postures in real time, and human body dynamic posture data is obtained. The human body dynamic posture data is subjected to feature extraction. The human body dynamic posture data includes the pixel coordinates in each frame image and the total number of pixels in each frame image. The coordinates of the pixels in each frame are obtained by the optical flow algorithm to obtain the moving direction of the pixels in each frame, and the moving direction of the pixels in each frame is calculated to obtain the angle of the pixels in each frame. The dynamic posture value is obtained according to the angle of the pixel points in each frame image.

6. A dynamic posture analysis method based on sports biomechanics according to claim 5, characterized in that: The specific steps of obtaining the dynamic posture value according to the angle of the pixel points in each frame of the image are: S1: Divide 360 ​​degrees into ten angle intervals, each interval is 36 degrees; S2: The angle interval corresponding to the angle of each pixel in each frame image is expressed by the formula ;in Indicates the angle interval corresponding to the angle of each pixel in each frame of the image. Represents the angle of each pixel in each frame of the image. It means that the angle of each pixel in each frame of the image is rounded down to the corresponding full angle range. Since the numbering starts from 1, 1 should be added; S3: Count the number of pixels in each angle interval in each frame image, and then obtain the angle ratio of the pixels in each angle interval in each frame image, which is expressed by the formula: ;in Indicates the angle ratio of the pixels in each angle interval in each frame image. Indicates the number of pixels in each angle interval in each frame of the image. Indicates the total number of pixels in each frame of the image; S4: arranging the angle proportions of the pixels in each angle interval of each of the ten angle intervals in each frame of the image in descending order; S5: Take the angle interval with the largest proportion in each frame image, express the angle of each pixel point in these angle intervals as the optimized angle, and sum them up to obtain the dynamic posture value; The dynamic posture value acquisition method is: Set the dynamic posture sample value. The method for obtaining the dynamic posture evaluation value based on the dynamic posture value and the dynamic posture sample value is as follows: ； in represents the dynamic posture evaluation value, represents the dynamic posture value, Represents the dynamic pose sample value.

7. The method for dynamic posture analysis based on sports biomechanics according to claim 1, characterized in that: In step three, when the human body contour evaluation value is greater than the human body contour threshold, redundant object contours are masked.

8. The method for dynamic posture analysis based on sports biomechanics according to claim 1, characterized in that: In step 4, a dynamic posture threshold is set and compared with the dynamic posture evaluation value; When the dynamic posture evaluation value is less than the dynamic posture threshold, it is determined that the actual motion trajectory of the human body meets the preset motion trajectory, and the dynamic posture detection continues.

9. The method for dynamic posture analysis based on sports biomechanics according to claim 1, characterized in that: In step five, when the dynamic posture evaluation value is greater than or equal to the dynamic posture threshold, the actual motion trajectory of the human body does not conform to the preset motion trajectory, an early warning prompt is given, the human body dynamic posture data continues to be acquired, the human body dynamic posture data is preprocessed, and the preprocessed human body dynamic posture data is evaluated.

10. A dynamic posture analysis system based on sports biomechanics, used in a dynamic posture analysis method based on sports biomechanics as claimed in any one of claims 1 to 9, characterized in that: It includes a data acquisition module, a data analysis module and a data execution module; The data acquisition module is used to acquire image data, average blood oxygen content data within the scanning range, and average temperature data within the scanning range and perform preprocessing, and send the data to the data execution module for contour scanning. When the data execution module locates the joint points of the human body contour, it acquires dynamic posture data and preprocesses it, and sends the dynamic posture data to the data analysis module; The data analysis module is used to analyze the results of the contour scanning and judge the contour, and send the judgment results to the data execution module for result execution until the data execution module locates the joint points of the human body contour. When the data analysis module obtains the dynamic posture data, it judges the consistency of the action and sends the judgment results to the data execution module; The data execution module performs human body contour judgment and action consistency judgment according to the judgment results of the data analysis module.

Images