An image processing-based motion posture evaluation system and method

Abstract

The application discloses a kind of motion posture evaluation system and method based on image processing, it is related to motion posture evaluation technical field, including the motion video of student, the motion approximation between evaluation motion video and teaching video is obtained reference teaching video;Reference teaching video is obtained to the motion video, the reference image set of reference teaching video is generated, the posture abnormality degree of posture image frame in posture image set is evaluated, and abnormal posture image frame is obtained;Abnormal posture image frame in posture image set is obtained, the position data of motion key point in abnormal posture image frame is obtained, and posture abnormality of motion key point in abnormal posture image frame is evaluated in combination with reference image set, and motion posture abnormality data is obtained;Motion posture abnormality data of motion video is obtained, and according to motion posture abnormality data, the motion posture adjustment data of student is generated, and the motion posture of student is intelligently adjusted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of motion posture evaluation, and particularly relates to a motion posture evaluation system and method based on image processing. BACKGROUND

[0002] At present, posture evaluation of students in the process of movement mainly includes visual observation method and video analysis method. The visual observation method is to directly observe the action and posture of the student through the eyes of the evaluator. However, this method is highly subjective and highly dependent on the experience and skills of the evaluator. The video analysis method uses a shooting device to shoot the action of the student, and then the evaluator plays back and analyzes the shot video. Although this method can be observed in detail for multiple times and overcomes the limitations of direct visual observation, it still requires an evaluator with professional knowledge to analyze the video, and is also subjective, which may cause analysis errors. Therefore, the current methods for posture evaluation of students in the process of movement basically all require an evaluator to directly or indirectly evaluate the students. These evaluation methods not only require a large number of talents with professional knowledge and skills, but also consume a large amount of manpower and financial resources. Moreover, the evaluation results of the students may be greatly different from the actual situation due to subjectivity. SUMMARY

[0003] The present application aims to provide a motion posture evaluation system and method based on image processing to solve the problems in the prior art.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical solution: a motion posture evaluation method based on image processing, the method comprising:

[0005] Step S100: constructing a posture evaluation cloud platform, obtaining a motion video of a student, obtaining a teaching video in the cloud platform, evaluating the motion similarity between the motion video and the teaching video, and obtaining a reference teaching video;

[0006] Step S200: obtaining a posture image set of the student, obtaining the reference teaching video of the motion video, generating a reference image set of the reference teaching video, evaluating the posture abnormality degree of the posture image frames in the posture image set, and obtaining abnormal posture image frames;

[0007] Step S300: obtaining the abnormal posture image frames in the posture image set, obtaining the position data of the motion key points in the abnormal posture image frames, combining the reference image set, evaluating the posture abnormality of the motion key points in the abnormal posture image frames, and obtaining motion posture abnormality data;

[0008] Step S400: Obtain the motion posture abnormal data of the motion video, and generate the motion posture adjustment data of the trainee according to the motion posture abnormal data, and intelligently adjust the motion posture of the trainee.

[0009] Further, step S100 includes:

[0010] Step S101: When the trainee uploads the photographed motion video to the cloud platform, the motion video is obtained, and according to the preset feature time length, the key frames in the motion video are extracted every feature time length, the posture image frames are obtained, the posture image frames are preprocessed, and the posture image frames of the motion video are obtained and collected to obtain the posture image set of the motion video;

[0011] Step S102: Randomly select several posture image frames from the posture image set, and mark them as marked posture image frames, collect the several marked posture image frames in the posture image set, and obtain the motion recognition set of the motion video;

[0012] Step S103: Obtain the preset motion key points, establish a two-dimensional coordinate system for the marked posture image frames, obtain the coordinates of the motion key points in the marked posture image frames, and obtain the feature vector F of the marked posture image frames:

[0013]

[0014] Wherein, x1, y1 represent the horizontal and vertical coordinates of the first motion key point in the marked posture image frame; x2, y2 represent the horizontal and vertical coordinates of the second motion key point in the marked posture image frame; x m , y m respectively represent the horizontal and vertical coordinates of the mth motion key point in the marked posture image frame; m represents the total number of motion key points; W represents the image width of the marked posture image frame; H represents the image width of the marked posture image frame;

[0015] Step S104: Calculate the feature aggregation vector F avg of the motion recognition set:

[0016]

[0017] Wherein, ε represents the total number of marked posture image frames in the motion recognition set; F i represents the feature vector of the ith marked posture image frame in the motion recognition set;

[0018] Step S105: Obtain a plurality of standard posture images of each teaching video from the cloud platform, obtain a feature vector of the standard posture image, obtain a motion recognition set, and obtain a motion approximation value of each teaching video, wherein the specific obtaining process of the motion recognition set and the motion approximation value of the a-th teaching video in the cloud platform is as follows:

[0019] Calculate the feature motion approximation value of the motion recognition set and the plurality of standard posture images of the a-th teaching video, wherein the feature motion approximation value R of the motion recognition set and the b-th standard posture image of the a-th teaching video is calculated as follows: a,b :

[0020]

[0021] Wherein, D a,b represents the feature vector of the b-th standard posture image;

[0022] Step S106: Calculate the motion approximation value C a :

[0023]

[0024] Wherein, j represents the total number of standard posture images of the a-th teaching video; R a,z represents the feature motion approximation value of the motion recognition set and the z-th standard posture image of the a-th teaching video;

[0025] Step S107: Evaluate the motion approximation of the motion video and each teaching video in the cloud platform, the specific process is as follows: obtain the maximum value of the motion approximation value of the motion recognition set and each teaching video, obtain the teaching video corresponding to the maximum value, and determine the motion approximation of the motion video and the teaching video in the cloud platform is the largest, and the teaching video is recorded as the reference teaching video of the motion video.

[0026] Further, step S200 includes:

[0027] Step S201: Obtain the reference teaching video of the motion video, extract the key frame every feature time length of the motion video to obtain the reference posture image frame, and collect a plurality of reference posture image frames of the reference teaching video to obtain the reference image set of the reference teaching video;

[0028] Step S202: Obtain the feature vector of the plurality of reference posture image frames, calculate the posture approximation value between each posture image frame in the posture image set of the motion video and the plurality of reference posture image frames in the reference image set, wherein the posture approximation value H α,β between the a-th posture image frame in the posture image set and the b-th reference posture image frame in the reference image set is calculated as follows:

[0029]

[0030] wherein, G α represents a feature vector of the αth posture image frame; Q β represents a feature vector of the βth reference posture image frame;

[0031] Step S203: obtaining the maximum value of the posture approximation between the αth posture image frame in the posture image set and the plurality of reference posture image frames, and taking the reference posture image frame corresponding to the maximum value of the posture approximation as the target reference posture image frame of the αth posture image frame;

[0032] Step S204: evaluating the posture abnormality degree of the αth posture image frame in the posture image set, and the specific evaluation process is that when the posture approximation between the αth posture image frame and the target reference posture image frame of the αth posture image frame is greater than or equal to a preset posture approximation threshold, it is determined that the motion posture of the student in the αth posture image frame is not abnormal.

[0033] Step S205: when the posture approximation between the αth posture image frame and the target reference posture image frame of the αth posture image frame is less than the preset posture approximation threshold, it is determined that the motion posture of the student in the αth posture image frame is abnormal, and the αth posture image frame is recorded as an abnormal posture image frame. Obtain a plurality of abnormal posture image frames in the posture image set.

[0034] Further, step S300 comprises:

[0035] Step S301: obtaining a plurality of abnormal posture image frames in the posture image set, obtaining the target reference posture image frame corresponding to the abnormal posture image frame, respectively establishing a two-dimensional coordinate system for the target reference posture image frame of the abnormal posture image frame, and obtaining the position data of the motion key points of the student from the abnormal posture image frame, wherein the position data comprises the coordinate vector of each motion key point in the abnormal posture image frame.

[0036] Step S302: obtaining the motion key points of the joint nodes of a plurality of preset feature joints in the reference teaching video to which the target reference posture image frame belongs, and recording them as the feature motion key points of the feature joints, and obtaining the adjacent motion key points of the feature motion key points, and recording them as the first motion key point and the second motion key point of the feature joints, respectively.

[0037] Step S303: evaluating the posture abnormality of the motion key points in the abnormal posture image frame, and the specific evaluation process is that the feature joint angle θ δ:

[0038]

[0039] wherein S δ denotes the coordinate vector of the first motion key point of the δth feature joint; S δ 1 denotes the coordinate vector of the first motion key point of the δth feature joint; S δ 2 denotes the coordinate vector of the second motion key point of the δth feature joint;

[0040] Step S304: calculating the feature joint angle deviation value △θ δ of the δth feature joint in the abnormal posture image frame from the feature joint angle of the target reference posture image frame. δ When the feature joint angle deviation value △θ δ is outside the preset angle deviation range, it is determined that the posture of the δth feature joint in the abnormal posture image frame is abnormal, and the δth feature joint is taken as the feature abnormal joint in the abnormal posture image frame. δ

[0041] Step S305: obtaining a plurality of feature abnormal joints of the abnormal posture image frame and collecting them to obtain the motion posture abnormal data of the student in the abnormal posture image frame.

[0042] The above steps first obtain the motion key points of the joint nodes of a plurality of preset feature joints, and then calculate the feature joint angle deviation value of the feature joint in the abnormal posture image frame from the feature joint angle of the target reference posture image frame, so as to know the region of the feature joint and whether the student's action is standard, which helps to quickly find the region where the student's motion posture is wrong, thereby better evaluating the student's motion posture.

[0043] Further, the step S400 comprises:

[0044] Step S401: obtaining a plurality of abnormal posture image frames in the motion video, obtaining motion posture abnormal data of the plurality of abnormal posture image frames, and extracting the feature joint angle deviation value of the feature abnormal joint of the abnormal posture image frame from the motion posture abnormal data, respectively.

[0045] Step S402: labeling the plurality of feature abnormal joints in the plurality of abnormal posture image frames and the corresponding target reference posture image frames, respectively, and generating motion posture adjustment data of the student, and displaying the motion posture adjustment data to the student in the program.

[0046] ​In order to better realize the above method, a motion posture evaluation system based on image processing is also proposed, which comprises a motion posture approximate evaluation module, an image posture evaluation module, a motion posture abnormal data module and a posture adjustment module.

[0047] The motion posture approximate evaluation module is used for evaluating the motion approximation between the motion video and the teaching video to obtain the reference teaching video.

[0048] The image posture evaluation module is used for evaluating the posture abnormality degree of the posture image frame in the posture image set to obtain the abnormal posture image frame.

[0049] The motion posture abnormal data module is used for acquiring the position data of the motion key point in the abnormal posture image frame, and combining the reference image set to evaluate the posture abnormality of the motion key point in the abnormal posture image frame to obtain the motion posture abnormal data.

[0050] The posture adjustment module is used for acquiring the motion posture abnormal data of the motion video, and generating the motion posture adjustment data of the student according to the motion posture abnormal data to intelligently adjust the motion posture of the student.

[0051] Further, the motion posture approximate evaluation module comprises a motion approximation value unit and a motion posture approximate evaluation unit.

[0052] The motion approximation value unit is used for calculating the motion approximation value between the motion recognition set and each teaching video in the cloud platform.

[0053] The motion posture approximate evaluation unit is used for evaluating the motion approximation between the motion video and each teaching video in the cloud platform to obtain the reference teaching video in the motion video.

[0054] Further, the image posture evaluation module comprises a posture approximation value unit and an image posture evaluation unit.

[0055] The posture approximation value unit is used for calculating the posture approximation value between each posture image frame in the posture image set of the motion video and the plurality of reference posture image frames of the reference image set.

[0056] The image posture evaluation unit is used for evaluating the motion posture difference degree between the αth posture image frame in the posture image set and the reference image set to obtain the abnormal posture image frame.

[0057] Further, the motion posture abnormal data module comprises a feature joint angle unit and a motion posture abnormal data unit.

[0058] The feature joint angle unit is used for calculating the feature joint angle of the plurality of feature joints in the abnormal posture image frame.

[0059] The motion posture abnormality data unit is used to acquire and aggregate several feature abnormal joints of abnormal posture image frames to obtain motion posture abnormality data.

[0060] Furthermore, the attitude adjustment module includes an attitude adjustment unit;

[0061] The posture adjustment unit is used to acquire several abnormal posture image frames in the motion video, acquire abnormal motion posture data of several abnormal posture image frames, annotate several abnormal joint features in several abnormal posture image frames and corresponding target reference posture image frames, and generate the trainee's motion posture adjustment data, which is then displayed to the trainee in the program.

[0062] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention assesses abnormal postures in student movement videos by first evaluating the similarity between the movement video and the instructional video, automatically finding the corresponding reference instructional video, and then evaluating the degree of posture abnormality in the posture image frames within the posture image set. This identifies abnormal posture image frames and analyzes the angles of the key points of motion where the characteristic joints are located in these abnormal posture image frames. This determines which characteristic joint of the student has a posture problem in the abnormal posture image frame and generates visualized adjustment data for intelligent adjustment of the student's movement posture. This not only avoids errors caused by the subjectivity of human analysis but also greatly saves manpower and resources, significantly improving analysis speed and accuracy. Attached Figure Description

[0063] Fig. 1 This is a flowchart of a motion posture evaluation system and method based on image processing according to the present invention;

[0064] Fig. 2 This is a schematic diagram of a motion posture evaluation system and method based on image processing according to the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Example: Figs. 1-2 As shown, the present invention provides a technical solution, a motion pose evaluation method based on image processing, the method comprising:

[0067] Step S100: constructing a posture evaluation cloud platform, acquiring a motion video of a student, acquiring a teaching video in the cloud platform, evaluating motion approximation between the motion video and the teaching video, and obtaining a reference teaching video;

[0068] Step S100 includes:

[0069] Step S101: when the student uploads the photographed motion video to the cloud platform in the program, the motion video is acquired, the key frames in the motion video are extracted every feature duration according to the preset feature duration, the posture image frames are obtained, the posture image frames are preprocessed, each posture image frame of the motion video is acquired and collected, and a posture image set of the motion video is obtained;

[0070] For example, the preprocessing generally includes size adjustment, normalization and other processing steps;

[0071] Step S102: a plurality of posture image frames are randomly selected from the posture image set, and are recorded as labeled posture image frames, the plurality of labeled posture image frames in the posture image set are acquired and collected, and a motion recognition set of the motion video is obtained;

[0072] Step S103: a plurality of motion key points are acquired, a two-dimensional coordinate system is established for the labeled posture image frames, the coordinates of the plurality of motion key points in the labeled posture image frames are acquired, and a feature vector F of the labeled posture image frames is acquired:

[0073]

[0074] Wherein, x1 and y1 represent the horizontal coordinate and the vertical coordinate of the first motion key point in the labeled posture image frame; x2 and y2 represent the horizontal coordinate and the vertical coordinate of the second motion key point in the labeled posture image frame; x m m m represents the total number of motion key points; W represents the image width of the labeled posture image frame; H represents the image width of the labeled posture image frame;

[0075] For example, the plurality of motion key points include wrist, left shoulder, etc.

[0076] Step S104: calculating a feature aggregation vector F avg :

[0077]

[0078] Wherein, ε represents the total number of labeled posture image frames in the motion recognition set; F i represents the feature vector of the i-th labeled posture image frame in the motion recognition set;​

[0079] Step S105: Obtain a plurality of standard posture images of each teaching video from the cloud platform, obtain a feature vector of the standard posture image, obtain a motion recognition set, and obtain a motion approximation value of each teaching video, wherein the specific obtaining process of the motion recognition set and the motion approximation value of the a-th teaching video in the cloud platform is as follows:

[0080] Calculate the feature motion approximation value of the motion recognition set and the plurality of standard posture images of the a-th teaching video, wherein the feature motion approximation value R a,b :

[0081]

[0082] Wherein, D a,b represents the feature vector of the b-th standard posture image;

[0083] Step S106: Calculate the motion approximation value C a :

[0084]

[0085] Wherein, j represents the total number of standard posture images of the a-th teaching video; R a,z represents the feature motion approximation value of the motion recognition set and the z-th standard posture image of the a-th teaching video;

[0086] For example, the total number of standard posture images j of the first teaching video is 5; the feature motion approximation value R 1,1 of the motion recognition set and the first standard posture image of the first teaching video is 0.8; the feature motion approximation value R 1,2 of the motion recognition set and the second standard posture image of the first teaching video is 0.7; the feature motion approximation value R 1,3 of the motion recognition set and the third standard posture image of the first teaching video is 0.9; the feature motion approximation value R 1,4 of the motion recognition set and the fourth standard posture image of the first teaching video is 0.5; the feature motion approximation value R 1,5 of the motion recognition set and the fifth standard posture image of the first teaching video is 0.8;

[0087] Calculate the motion approximation value C1 of the motion recognition set and the first teaching video:

[0088]

[0089] Step S107: evaluating the motion video, and the motion approximation with each teaching video in the cloud platform, the specific process is, obtaining the motion recognition set, and the maximum value of the motion approximation value of each teaching video, obtaining the teaching video corresponding to the maximum value, and determining that the motion video has the maximum motion approximation with the teaching video in the cloud platform, and the teaching video is recorded as the reference teaching video of the motion video;

[0090] Step S200: obtaining the posture image set of the student, obtaining the reference teaching video of the motion video, generating the reference image set of the reference teaching video, evaluating the posture abnormality degree of the posture image frame in the posture image set, and obtaining the abnormal posture image frame;

[0091] Among them, step S200 includes:

[0092] Step S201: obtaining the reference teaching video of the motion video, extracting the key frame every feature time length of the motion video, obtaining the reference posture image frame, and collecting a plurality of reference posture image frames of the reference teaching video to obtain the reference image set of the reference teaching video;

[0093] Step S202: obtaining the feature vector of a plurality of reference posture image frames, calculating the posture approximation value between each posture image frame in the posture image set of the motion video and a plurality of reference posture image frames in the reference image set, wherein the posture approximation value H α,β :

[0094]

[0095] Among them, G α represents the feature vector of the αth posture image frame; Q β represents the feature vector of the βth reference posture image frame;

[0096] Step S203: obtaining the maximum value of the posture approximation value between the αth posture image frame in the posture image set and a plurality of reference posture image frames, and taking the reference posture image frame corresponding to the maximum value of the posture approximation value as the target reference posture image frame of the αth posture image frame;

[0097] Step S204: evaluating the posture abnormality degree of the αth posture image frame in the posture image set, and the specific evaluation process is that when the posture approximation value between the αth posture image frame and the target reference posture image frame of the αth posture image frame is greater than or equal to the preset posture approximation threshold, it is determined that the motion posture of the student in the αth posture image frame is not abnormal;

[0098] Step S205: When the pose approximation value of the αth posture image frame and the target reference posture image frame of the αth posture image frame is less than the preset pose approximation threshold, it is determined that the motion posture of the student in the αth posture image frame is abnormal, and the αth posture image frame is recorded as an abnormal posture image frame. A plurality of abnormal posture image frames in the posture image set are obtained;

[0099] Step S300: Obtain the abnormal posture image frame in the posture image set, obtain the position data of the motion key point in the abnormal posture image frame, and combine the reference image set to evaluate the posture abnormality of the motion key point in the abnormal posture image frame, to obtain the motion posture abnormality data;

[0100] Wherein, step S300 comprises:

[0101] Step S301: Obtain a plurality of abnormal posture image frames in the posture image set, obtain the target reference posture image frame corresponding to the abnormal posture image frame, respectively establish a two-dimensional coordinate system for the target reference posture image frame of the abnormal posture image frame, obtain the position data of the motion key point of the student from the abnormal posture image frame, and the position data includes the coordinate vector of each motion key point in the abnormal posture image frame;

[0102] Step S302: Obtain the motion key point of the joint of the preset plurality of feature joints in the reference teaching video to which the target reference posture image frame belongs, and record it as the feature motion key point of the feature joint, obtain the adjacent motion key points of the feature motion key point, and record them as the first motion key point and the second motion key point of the feature joint, respectively.

[0103] For example, the plurality of feature joints include shoulder joint, elbow joint, knee joint, etc.

[0104] Step S303: Evaluate the posture abnormality of the motion key point in the abnormal posture image frame. The specific evaluation process is: calculate the feature joint angle of the plurality of feature joints in the abnormal posture image frame, wherein the feature joint angle θ δ :

[0105]

[0106] Wherein, S δ is the coordinate vector of the feature motion key point of the δth feature joint; S δ 1 is the coordinate vector of the first motion key point of the δth feature joint; S δ 2 is the coordinate vector of the second motion key point of the δth feature joint.

[0107] Step S304: calculating the feature joint angle deviation value △θ of the δth feature joint in the abnormal posture image frame and the feature joint angle of the target reference posture image frame δ = θ δ ′-θ δ When the feature joint angle deviation value △θ δ When the feature joint angle deviation value △θ is outside the preset angle deviation range, it is determined that the posture of the δth feature joint in the abnormal posture image frame is abnormal, and the δth feature joint is taken as the feature abnormal joint in the abnormal posture image frame;

[0108] Step S305: obtaining a plurality of feature abnormal joints of the abnormal posture image frame and collecting them to obtain the motion posture abnormal data of the student in the abnormal posture image frame;

[0109] Step S400: obtaining the motion posture abnormal data of the motion video, and generating the motion posture adjustment data of the student according to the motion posture abnormal data, and intelligently adjusting the motion posture of the student;

[0110] The step S400 includes:

[0111] Step S401: obtaining a plurality of abnormal posture image frames in the motion video, obtaining the motion posture abnormal data of the plurality of abnormal posture image frames, and respectively extracting the feature joint angle deviation value of the feature abnormal joint of the abnormal posture image frame from the motion posture abnormal data;

[0112] Step S402: respectively labeling the plurality of feature abnormal joints in the plurality of abnormal posture image frames and the corresponding target reference posture image frames, and generating the motion posture adjustment data of the student, and displaying the student in the program;

[0113] In order to better realize the above method, a motion posture evaluation system based on image processing is also proposed, which includes a motion posture approximate evaluation module, an image posture evaluation module, a motion posture abnormal data module, and a posture adjustment module.

[0114] The motion posture approximate evaluation module is used to evaluate the motion approximation between the motion video and the teaching video to obtain the reference teaching video.

[0115] The image posture evaluation module is used to evaluate the posture abnormality degree of the posture image frame in the posture image set to obtain the abnormal posture image frame.

[0116] The motion posture abnormal data module is used to obtain the position data of the motion key point in the abnormal posture image frame, and combine the reference image set to evaluate the posture abnormality of the motion key point in the abnormal posture image frame to obtain the motion posture abnormal data.

[0117] The posture adjustment module is configured to acquire motion posture abnormal data of the motion video, and generate motion posture adjustment data of the student according to the motion posture abnormal data, so as to intelligently adjust the motion posture of the student.

[0118] The motion posture approximate evaluation module includes a motion approximate value unit and a motion posture approximate evaluation unit.

[0119] The motion approximate value unit is configured to calculate motion approximate values of the motion recognition set and each teaching video in the cloud platform.

[0120] The motion posture approximate evaluation unit is configured to evaluate motion approximation of the motion video and each teaching video in the cloud platform, and obtain a reference teaching video in the motion video.

[0121] The image posture evaluation module includes a posture approximate value unit and an image posture evaluation unit.

[0122] The posture approximate value unit is configured to calculate posture approximate values between each posture image frame in a posture image set of the motion video and a plurality of reference posture image frames in a reference image set.

[0123] The image posture evaluation unit is configured to evaluate a motion posture difference degree between an αth posture image frame in the posture image set and the reference image set, and obtain an abnormal posture image frame.

[0124] The motion posture abnormal data module includes a feature joint angle unit and a motion posture abnormal data unit.

[0125] The feature joint angle unit is configured to calculate feature joint angles of a plurality of feature joints in the abnormal posture image frame.

[0126] The motion posture abnormal data unit is configured to acquire and collect a plurality of feature abnormal joints of the abnormal posture image frame, and obtain motion posture abnormal data.

[0127] The posture adjustment module includes a posture adjustment unit.

[0128] The posture adjustment unit is configured to acquire a plurality of abnormal posture image frames in the motion video, acquire motion posture abnormal data of the plurality of abnormal posture image frames, label a plurality of feature abnormal joints in the plurality of abnormal posture image frames and corresponding target reference posture image frames respectively, and generate motion posture adjustment data of the student, and display the motion posture adjustment data to the student in a program.

[0129] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. The embodiments should therefore be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims with respect to the figures of the patent document.

Claims

1. A method for motion posture assessment based on image processing, characterized in that, The method comprises: Step S100: constructing a posture evaluation cloud platform, acquiring a motion video of a student, acquiring a teaching video in the cloud platform, evaluating motion approximation between the motion video and the teaching video, and obtaining a reference teaching video; Step S200: acquiring a posture image set of the student, acquiring a reference teaching video of the motion video, generating a reference image set of the reference teaching video, evaluating posture abnormality degree of posture image frames in the posture image set, and obtaining abnormal posture image frames; Step S300: acquiring the abnormal posture image frames in the posture image set, acquiring position data of motion key points in the abnormal posture image frames, and combining the reference image set to evaluate posture abnormality of the motion key points in the abnormal posture image frames, and obtaining motion posture abnormality data; Step S400: acquiring motion posture abnormality data of the motion video, and generating motion posture adjustment data of the student according to the motion posture abnormality data, and intelligently adjusting the motion posture of the student; The step S100 comprises: Step S101: when the student uploads the shot motion video to the cloud platform, the motion video is acquired, the key frames in the motion video are extracted every feature time according to the preset feature time, the posture image frames are obtained, the posture image frames are preprocessed, and the posture image set of the motion video is obtained by collecting the posture image frames of the motion video; Step S102: a plurality of posture image frames are randomly selected from the posture image set and are recorded as marked posture image frames, and the plurality of marked posture image frames in the posture image set are acquired and collected to obtain a motion recognition set of the motion video; Step S103: acquiring preset motion key points, establishing a two-dimensional coordinate system for the marked posture image frames, acquiring coordinates of the motion key points in the marked posture image frames, and acquiring a feature vector F of the marked posture image frames: ， Wherein, x1, y1 respectively represent the horizontal coordinate and the vertical coordinate of the first motion key point in the marker posture image frame; x2, y2 respectively represent the horizontal coordinate and the vertical coordinate of the second motion key point in the marker posture image frame; x m , y m respectively represent the horizontal coordinate and the vertical coordinate of the mth motion key point in the marker posture image frame; m represents the total number of the motion key points; W represents the image width of the marker posture image frame; H represents the image width of the marker posture image frame; Step S104: calculating a feature aggregation vector F of the motion recognition set avg : ， wherein ε represents the total number of labeled gesture image frames in the motion recognition set; F i represents the feature vector of the i-th labeled gesture image frame in the motion recognition set; Step S105: acquiring a plurality of standard posture images of each teaching video from the cloud platform, acquiring a feature vector of the standard posture image, acquiring the motion recognition set, and the motion approximation value of each teaching video, wherein the specific acquisition process of the motion recognition set and the motion approximation value of the a-th teaching video in the cloud platform is as follows: calculating the motion recognition set, and the feature motion approximation value of the plurality of standard posture images of the a-th teaching video, wherein the motion recognition set, and the feature motion approximation value R of the b-th standard posture image of the a-th teaching video a,b : ， wherein D a,b is a feature vector representing the bth standard pose image; Step S106: Calculate the motion recognition set, and the motion approximation value C of the a-th teaching video a : ， wherein j represents the total number of standard posture images of the a-th teaching video; R a,z represents the feature motion approximation value of the z-th standard posture image of the a-th teaching video in the motion recognition set. Step S107: evaluating motion approximation of the motion video and each teaching video in the cloud platform, and the specific process is as follows: acquiring a maximum value of the motion approximation value of the motion recognition set and each teaching video, acquiring a teaching video corresponding to the maximum value, and determining that the motion approximation of the motion video and the teaching video in the cloud platform is the largest, and the teaching video is recorded as the reference teaching video of the motion video; The step S200 comprises: Step S201: Obtain a reference teaching video of the sports video, extract key frames of the sports video every feature duration to obtain reference posture image frames, and collect a plurality of reference posture image frames of the reference teaching video to obtain a reference image set of the reference teaching video; Step S202: obtaining feature vectors of the several reference posture image frames, and calculating posture approximations between each posture image frame in the posture image set of the motion video and the several reference posture image frames in the reference image set, wherein a posture approximation H between an αth posture image frame in the posture image set and a βth reference posture image frame in the reference image set is calculated according to the following formula: α,β : ， wherein G α represents a feature vector of the first a-th pose image frame; Q β represents a feature vector of the first β-th reference pose image frame; Step S203: Obtain the maximum value of the posture approximation value between the αth posture image frame in the posture image set and the plurality of reference posture image frames, and take the reference posture image frame corresponding to the maximum value of the posture approximation value as the target reference posture image frame of the αth posture image frame; Step S204: Evaluate the posture abnormality degree of the αth posture image frame in the posture image set, and the specific evaluation process is that when the posture approximation value between the αth posture image frame and the target reference posture image frame of the αth posture image frame is greater than or equal to a preset posture approximation threshold, it is determined that the movement posture of the student in the αth posture image frame is not abnormal; Step S205: When the posture approximation value between the αth posture image frame and the target reference posture image frame of the αth posture image frame is less than a preset posture approximation threshold, it is determined that the movement posture of the student in the αth posture image frame is abnormal, and the αth posture image frame is recorded as an abnormal posture image frame. Obtain a plurality of abnormal posture image frames in the posture image set; The step S300 comprises: Step S301: Obtain a plurality of abnormal posture image frames in the posture image set, obtain the target reference posture image frame corresponding to the abnormal posture image frame, respectively establish a two-dimensional coordinate system for the target reference posture image frame of the abnormal posture image frame, obtain the position data of the movement key points of the student from the abnormal posture image frame, and the position data comprises the coordinate vector of each movement key point in the abnormal posture image frame; Step S302: Obtain the movement key points of the joint nodes of a plurality of feature joints in the reference teaching video to which the target reference posture image frame belongs, and record them as feature movement key points of the feature joints, and obtain adjacent movement key points of the feature movement key points, and record them as first movement key points and second movement key points of the feature joints respectively; Step S303: evaluating the posture abnormality of the motion key point in the abnormal posture image frame, and the specific evaluation process is: calculating the feature joint angle of several feature joints in the abnormal posture image frame, wherein the feature joint angle θ δ of the δth feature joint in the abnormal posture image frame is calculated according to the following formula: θ δ = arccos (cos (θ δ-1) * cos (θ δ+1) - sin (θ δ-1) * sin (θ δ+1)), wherein θ δ-1 is the feature joint angle of the (δ-1)th feature joint in the abnormal posture image frame, and θ δ+1 is the feature joint angle of the (δ+1)th feature joint in the abnormal posture image frame. δ : ， wherein S δ represents a coordinate vector of the first motion key point of the δth feature joint; S δ 1 represents a coordinate vector of the first motion key point of the δth feature joint; S δ 2 represents a coordinate vector of the second motion key point of the δth feature joint; Step S304: calculating a feature joint angle deviation value Δθ of a δth feature joint in the abnormal posture image frame and a feature joint angle of the target reference posture image frame δ = θ δ ´- θ δ When the feature joint angle deviation value Δθ δ When the feature joint angle deviation value Δθ is outside the preset angle deviation range, it is determined that the posture of the δth feature joint in the abnormal posture image frame is abnormal, and the δth feature joint is taken as a feature abnormal joint in the abnormal posture image frame. Step S305: Obtain a plurality of feature abnormal joints of the abnormal posture image frame and collect them to obtain the movement posture abnormality data of the student in the abnormal posture image frame; The step S400 comprises: Step S401: Obtain a plurality of abnormal posture image frames in the sports video, obtain the movement posture abnormality data of the plurality of abnormal posture image frames, and respectively extract the feature joint angle deviation value of the feature abnormal joint of the abnormal posture image frame from the movement posture abnormality data; Step S402: Label the plurality of feature abnormal joints in the plurality of abnormal posture image frames and the corresponding target reference posture image frames respectively, and generate the movement posture adjustment data of the student, and display it to the student in the program.

2. An image processing-based motion posture evaluation system for performing the image processing-based motion posture evaluation method of claim 1, characterized by The system comprises a motion posture approximation evaluation module, an image posture evaluation module, a motion posture abnormal data module, and a posture adjustment module. The motion posture approximation evaluation module is configured to evaluate the motion approximation between the motion video and the teaching video to obtain a reference teaching video. The image posture evaluation module is configured to evaluate the posture abnormality degree of the posture image frames in the posture image set to obtain abnormal posture image frames. The motion posture abnormal data module is configured to obtain the position data of the motion key points in the abnormal posture image frames, and evaluate the posture abnormality of the motion key points in the abnormal posture image frames in combination with the reference image set to obtain motion posture abnormal data. The posture adjustment module is configured to obtain the motion posture abnormal data of the motion video, and generate motion posture adjustment data of the student according to the motion posture abnormal data to intelligently adjust the motion posture of the student.

3. The motion gesture evaluation system based on image processing according to claim 2, characterized in that, The motion posture approximation evaluation module comprises a motion approximation value unit and a motion posture approximation evaluation unit. The motion approximation value unit is configured to calculate the motion approximation values between the motion recognition set and each teaching video in the cloud platform. The motion posture approximation evaluation unit is configured to evaluate the motion approximation between the motion video and each teaching video in the cloud platform to obtain a reference teaching video in the motion video.

4. The motion posture evaluation system based on image processing according to claim 2, characterized in that, The image posture evaluation module comprises a posture approximation value unit and an image posture evaluation unit. The posture approximation value unit is configured to calculate the posture approximation values between each posture image frame in the posture image set of the motion video and the reference posture image frames in the reference image set. The image posture evaluation unit is configured to evaluate the motion posture difference degree between the αth posture image frame in the posture image set and the reference image set to obtain abnormal posture image frames.

5. The motion gesture evaluation system based on image processing according to claim 2, wherein, The motion posture abnormal data module comprises a feature joint angle unit and a motion posture abnormal data unit. The feature joint angle unit is configured to calculate the feature joint angles of the feature joints in the abnormal posture image frames. The motion posture abnormal data unit is configured to obtain and collect the feature abnormal joints of the abnormal posture image frames to obtain motion posture abnormal data.

6. The motion gesture evaluation system based on image processing according to claim 2, wherein, The posture adjustment module comprises a posture adjustment unit. The posture adjustment unit is configured to obtain a plurality of abnormal posture image frames in the motion video, obtain the motion posture abnormal data of the plurality of abnormal posture image frames, label a plurality of feature abnormal joints in the plurality of abnormal posture image frames and corresponding target reference posture image frames, and generate motion posture adjustment data of the student for display to the student in the program.

Images