Human body motion posture correction method and system based on computer vision
By establishing a multi-view reference coordinate system and action coordinate data set, comparing the coordinate differences between user actions and standard actions, and calculating deviation values, the judgment inaccuracy problem caused by multi-point coincidence in the existing technology is solved, and the correction of human movement postures is improved accuracy.
Patent Information
- Application Number
- CN202411201443.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-12-06
AI Technical Summary
The existing computer vision-based human motion posture correction method cannot effectively determine whether the user's movement is standard when multiple points overlap, reducing the accuracy of the judgment.
By establishing a reference coordinate system for the front, left and right side view, the coordinates of multiple key parts in the standard movement of human body are collected, the corresponding action coordinate data set is formed, and the deviation value is calculated. Determine whether the user's actions are standard and prompt to correct the non-standard actions.
It improves the accuracy of judging user action standards, avoids judgment errors when multiple points overlap, and enhances the accuracy of the system.
Smart Images

Figure CN119092049A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of human body motion posture correction, and in particular to a human body motion posture correction method and system based on computer vision. Background Art
[0002] The human motion posture correction method based on computer vision mainly uses a camera to capture the dynamic posture of the human body, and then uses deep learning and image processing technology to analyze these posture data and identify incorrect motion patterns or postures. The system will provide real-time feedback or suggestions to help users adjust and improve their motion postures to achieve better results and reduce the risk of injury.
[0003] To this end, Chinese patent CN 104200491 A discloses a human motion posture correction system, which constructs human skeleton data and human joint point Euclidean distance data to obtain human motion posture information: the image standard library module stores human standard motion posture data, which includes human skeleton data and human joint point data; the image comparison module compares the human skeleton and human joint point Euclidean distance data constructed by the kinect camera with the standard motion posture data in the standard library to obtain the result of whether the posture is correct, and obtains the error between the relatively correct postures of the human body posture; the image correction output module outputs the result of comparing the collected human body posture with the standard motion posture in the image standard library and the error information;
[0004] However, when multiple points overlap in this way, it is impossible to effectively determine whether the user's action is standard, thereby reducing the accuracy of the judgment. Summary of the invention
[0005] 1. Technical issues to be resolved
[0006] In view of the deficiencies in the prior art, the present invention provides a method and system for correcting human motion posture based on computer vision, which has the advantages of avoiding the inability to effectively judge whether the user's movements are standard when multiple points overlap, improving the accuracy of judgment, and solving the above problems.
[0007] (II) Technical solution
[0008] To achieve the above object, the present invention provides the following technical solution: a method for correcting human body motion posture based on computer vision, comprising the following steps:
[0009] S1. Input a standard human motion and establish a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, wherein the reference coordinate system includes a front reference coordinate system, a left view reference coordinate system and a right view reference coordinate system;
[0010] S2, obtaining multiple front coordinates of the human head, multiple front coordinates of the left hand, multiple front coordinates of the right hand, multiple front coordinates of the left foot, and multiple front coordinates of the right foot in the standard human motion according to the established front reference coordinate system, obtaining multiple left view coordinates of the human head, multiple left view coordinates of the left hand, and multiple left view coordinates of the left foot in the standard human motion according to the established left view reference coordinate system, and obtaining multiple right view coordinates of the human head, multiple right view coordinates of the right hand, and multiple right view coordinates of the right foot according to the established right view reference coordinate system;
[0011] S3, establishing a corresponding front action coordinate data set according to a plurality of human head front coordinates, a plurality of left hand front coordinates, a plurality of right hand front coordinates, a plurality of left foot front coordinates and a plurality of right foot front coordinates in a sequence of steps to complete the corresponding action; establishing a left view action coordinate data set according to a plurality of human head left view coordinates, a plurality of left hand left view coordinates and a plurality of left foot left view coordinates in a sequence of steps to complete the corresponding action; establishing a right view action coordinate data set according to a plurality of human head right view coordinates, a plurality of right hand right view coordinates and a plurality of right foot right view coordinates in a sequence of steps to complete the corresponding action;
[0012] S4, obtaining the frontal control coordinates of the human head, the left hand, the right hand, the left foot and the right foot after each movement completed by the user in accordance with the standard human movement movements, obtaining the left-view control coordinates of the human head, the left-view control coordinates of the left hand and the left-view control coordinates of the left foot after each movement completed by the user in accordance with the standard human movement movements, and obtaining the right-view control coordinates of the human head, the right-view control coordinates of the right hand and the right-view control coordinates of the right foot after each movement completed by the user in accordance with the standard human movement movements;
[0013] S5, comparing and calculating the obtained frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the right view control coordinates of the human head, the right hand, and the right foot with the corresponding frontal action coordinate data set, the left view action coordinate data set, and the right view action coordinate data set one by one;
[0014] S6. Compare the analysis result with a preset threshold value to determine whether the action performed by the user is standard. If a non-standard action occurs, prompt the user to correct the non-standard action.
[0015] Preferably, the frontal action coordinate data set expression is:
[0016]
[0017] in, Indicates all the acquired human head frontal coordinate data, Indicates all the left-hand front coordinates obtained, Indicates all the right-hand front coordinates obtained, Indicates all the obtained left foot front coordinates, Indicates all the acquired right foot front coordinates.
[0018] Preferably, the left-view action coordinate data set expression is as follows:
[0019]
[0020] in, Indicates all the obtained left-view coordinates of the human head. Indicates all the left-hand left-view coordinates obtained, Represents all the obtained left foot left view coordinates.
[0021] Preferably, the right-view action coordinate data set expression is as follows:
[0022]
[0023] in, Indicates all the acquired right view coordinates of the human head. Indicates all the right-handed right-view coordinates obtained. Represents all the right foot right view coordinates obtained.
[0024] Preferably, the frontal coordinate expression of the human head is: (DZTX i ,DZTY i ), the left hand frontal coordinate expression is: (DZZSX i ,DZZSY i ), the right-hand frontal coordinate expression is: (DZYSX i ,DZYSY i ), the left foot front coordinate expression is: (DZZJX i ,DZZJY i ), the coordinate expression of the right foot front view is: (DZZJX i ,DZZJY i ), the left view coordinate expression of the human head is: (ZZTX i ,ZZTY i ), the left-hand left-view coordinate comparison expression is: (ZZZSX i ,ZZZSY i ), the left foot left view coordinate expression is: (ZZZJX i ,ZZZJYi ), the human head right view coordinate expression is: (ZYTX i ,ZYTY i ), the right-hand right-view coordinate expression is: (ZZYSX i ,ZZYSY i ), the right foot right view coordinate expression is: (ZZYJX i ,ZZYJY i ).
[0025] Preferably, when any one or more of the frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the left foot, the right view control coordinates of the human head, the right hand, the right foot, the right foot, the right view control coordinates of the human head, the right hand, and the right foot exceed the coordinates of the corresponding steps in the corresponding frontal action coordinate data set, the left view action coordinate data set, or the right view action coordinate data set, it is determined that the user action deviates from the standard action of the human body movement, which is the analysis result.
[0026] Preferably, when it is determined that the user's action deviates from the standard action of human body movement, the deviation value is calculated, and the calculation expression of the deviation value is as follows:
[0027]
[0028] In the formula, AX, AY represent the frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the left foot, the right view control coordinates of the human head, the right hand, or the right foot, and BX, BY represent the coordinate data in the frontal action coordinate data set, the left view action coordinate data set, or the right view action coordinate data set corresponding to the control coordinates.
[0029] Preferably, the expression for determining whether the action performed by the user is standard is as follows:
[0030] When (1±0.02)CZ>threshold, it is judged that the user's action is not standard.
[0031] A human motion posture correction system based on computer vision, comprising:
[0032] The data collection module is used to collect and input standard human motion movements, and establish a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, and accordingly form a front motion coordinate data set, a left-view motion coordinate data set, and a right-view motion coordinate data set;
[0033] The data acquisition module is used to obtain the frontal control coordinates of the human head, the left hand, the right hand, the left foot and the right foot after the user performs a movement according to the standard human movement movements, obtain the left-view control coordinates of the human head, the left-view control coordinates of the left hand and the left-view control coordinates of the left foot after the user performs a movement according to the standard human movement movements, and obtain the right-view control coordinates of the human head, the right-view control coordinates of the right hand and the right-view control coordinates of the right foot after the user performs a movement according to the standard human movement movements;
[0034] The data analysis module is used to compare and calculate the acquired frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand and the left foot, the right view control coordinates of the human head, the right hand and the right foot with the corresponding frontal action coordinate data set, the left view action coordinate data set and the right view action coordinate data set, and compare the analysis results with the preset threshold value to determine whether the action performed by the user is standard. When non-standard actions occur, the user is prompted to correct the non-standard actions.
[0035] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements any step of a method for correcting human motion posture based on computer vision when executing the computer program.
[0036] Compared with the prior art, the present invention provides a method and system for correcting human motion posture based on computer vision, which has the following beneficial effects:
[0037] The present invention collects the coordinates of various key parts when the user completes an action from different perspectives, establishes a corresponding surface reference coordinate system according to standard human body motion movements, collects the coordinates of various parts of the human body when completing each action, and establishes a corresponding coordinate set. When the user performs a corresponding action according to the standard human body motion movement, whether the user's action is standard is judged by whether the coordinate data generated by the completed action steps is the same as the coordinate data in the corresponding coordinate set. When non-standard occurs, the deviation value is calculated to judge whether the error of the user's action is within the allowable deviation, and a corresponding warning is issued to prompt the user to compare the coordinate data of the same parts of the human body on multiple planes to judge whether the user's action is standard, thereby avoiding the situation where multiple points overlap and it is impossible to effectively judge whether the user's action is standard, thereby improving the accuracy of the judgment. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is a schematic diagram of the steps of the method of the present invention. DETAILED DESCRIPTION
[0039] 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.
[0040] In order to solve the problem in the prior art that when multiple acquisition points overlap, it is impossible to effectively judge whether the user's action is standard, thereby reducing the accuracy of judgment, a human motion posture correction method based on computer vision is proposed. Figure 1 , the method is as follows:
[0041] S1. Input a standard human motion and establish a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, wherein the reference coordinate system includes a front reference coordinate system, a left view reference coordinate system and a right view reference coordinate system;
[0042] S2. Acquire multiple front coordinates of the human head, multiple front coordinates of the left hand, multiple front coordinates of the right hand, multiple front coordinates of the left foot, and multiple front coordinates of the right foot in the standard human motion action according to the established front reference coordinate system; acquire multiple left-view coordinates of the human head, multiple left-view coordinates of the left hand, and multiple left-view coordinates of the left foot in the standard human motion action according to the established left-view reference coordinate system; acquire multiple right-view coordinates of the human head, multiple right-view coordinates of the right hand, and multiple right-view coordinates of the right foot according to the established right-view reference coordinate system; acquire coordinates corresponding to different step actions according to the established reference coordinate system, thereby obtaining coordinate data corresponding to the standard action, so as to compare with the coordinate data generated for the subsequent action completed by the user, and determine whether the user action is standard;
[0043] S3, according to the multiple front coordinates of the human head, multiple front coordinates of the left hand, multiple front coordinates of the right hand, multiple front coordinates of the left foot and multiple front coordinates of the right foot, a corresponding front action coordinate data set is established in the order of steps to complete the corresponding action; according to the multiple left view coordinates of the human head, multiple left view coordinates of the left hand and multiple left view coordinates of the left foot, a left view action coordinate data set is established in the order of steps to complete the corresponding action; according to the multiple right view coordinates of the human head, multiple right view coordinates of the right hand and multiple right view coordinates of the right foot, a right view action coordinate data set is established in the order of steps to complete the corresponding action, wherein:
[0044] The expression of the front action coordinate data set is:
[0045]
[0046] in, Indicates all the acquired human head frontal coordinate data, Indicates all the left-hand front coordinates obtained, Indicates all the right-hand front coordinates obtained, Indicates all the obtained left foot front coordinates, Indicates all the acquired right foot front coordinates;
[0047] The left-view action coordinate data set expression is as follows:
[0048]
[0049] in, Indicates all the obtained left-view coordinates of the human head. Indicates all the left-hand left-view coordinates obtained, Indicates all the obtained left foot left view coordinates;
[0050] The expression of the right-view action coordinate data set is as follows:
[0051]
[0052] in, Indicates all the acquired right view coordinates of the human head. Indicates all the right-handed right-view coordinates obtained. Represents all the right foot right view coordinates obtained.
[0053] By forming a data set with the coordinate data generated by the standard action of the corresponding surface, each part is formed into a standard action corresponding to the coordinate data of the body part according to the action sequence, so as to facilitate the subsequent comparison of the coordinate values of each part when the user makes an action.
[0054] S4, obtaining the frontal coordinates of the human head, the frontal coordinates of the left hand, the frontal coordinates of the right hand, the frontal coordinates of the left foot, and the frontal coordinates of the right foot after each movement completed by the user in accordance with the standard human movement movements, obtaining the left-view coordinates of the human head, the left-view coordinates of the left hand, and the left-view coordinates of the left foot after each movement completed by the user in accordance with the standard human movement movements, and obtaining the right-view coordinates of the human head, the right-view coordinates of the right hand, and the right-view coordinates of the right foot after each movement completed by the user in accordance with the standard human movement movements, wherein:
[0055] The coordinate expression of the human head frontal reference is: (DZTX i ,DZTY i ), the left hand front coordinate expression is: (DZZSX i ,DZZSYi ), the coordinate expression of the right-hand front comparison is: (DZYSX i ,DZYSY i ), the left foot front coordinate expression is: (DZZJX i ,DZZJY i ), the coordinate expression of the right foot front view is: (DZZJX i ,DZZJY i ), the left view coordinate expression of the human head is: (ZZJX i ,ZYTY i ), the left-hand left-view coordinate comparison expression is: (ZZZSX i ,ZYTY i ), the left foot left view coordinate expression is: (ZZZJX i ,ZZZJY i ), the coordinate expression of the right view of the human head is: (ZYTX i ,ZYTY i ), the right-hand right-view coordinate expression is: (ZZYSX i ,ZZYSY i ), the coordinate expression of the right foot and right view is: (ZZYJX i ,ZZYJY i ), i in each coordinate data represents the specific step number in the standard action corresponding to the action completed by the user, so that the coordinate data generated by the steps completed by the user can be compared with the coordinate data corresponding to the standard action that has been collected and generated to determine whether the action is standard;
[0056] S5, comparing and calculating the obtained human head frontal control coordinates, left hand frontal control coordinates, right hand frontal control coordinates, left foot frontal control coordinates, right foot frontal control coordinates, human head left-view control coordinates, left hand left-view control coordinates and left foot left-view control coordinates, human head right-view control coordinates, right right-view control coordinates and right foot right-view control coordinates with the corresponding frontal action coordinate data set, left-view action coordinate data set and right-view action coordinate data set one by one; when any one or more coordinates of the human head frontal control coordinates, left hand frontal control coordinates, right hand frontal control coordinates, left foot frontal control coordinates, right foot frontal control coordinates, human head left-view control coordinates, left hand left-view control coordinates and left foot left-view control coordinates, human head right-view control coordinates, right right-view control coordinates and right foot right-view control coordinates exceed the coordinates of the corresponding steps in the corresponding frontal action coordinate data set, left-view action coordinate data set or right-view action coordinate data set, it is determined that the user action deviates from the standard action of human motion, which is the analysis result. At this time, it is necessary to calculate the size of the deviation value to determine whether the user's action is within an acceptable action error range;
[0057] When it is determined that the user's action deviates from the standard action of human body movement, the deviation value is calculated. The calculation expression of the deviation value is as follows:
[0058]
[0059] In the formula, AX, AY represent the frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the left foot, the right view control coordinates of the human head, the right hand, or the right foot, and BX, BY represent the coordinate data in the frontal action coordinate data set, the left view action coordinate data set, or the right view action coordinate data set corresponding to the control coordinates.
[0060] S6. Compare the analysis result with the preset threshold value to determine whether the action performed by the user is standard. If a non-standard action occurs, prompt the user to correct the non-standard action. The expression for determining whether the action performed by the user is standard is as follows:
[0061] When (1±0.02)CZ>threshold, the user's action is judged to be non-standard, and (1±0.02) represents the allowable action error size.
[0062] By collecting the coordinates of each key part of the user when completing the action from different perspectives, and establishing the corresponding surface reference coordinate system according to the standard human body movement action, and collecting the coordinates of each part of the human body when completing each action, and establishing the corresponding coordinate set, when the user makes the corresponding action according to the standard human body movement action, the coordinate data generated by the completed action steps is compared with the coordinate data in the corresponding coordinate set to judge whether the user's action is standard. When it is not standard, the deviation value is calculated to judge whether the error of the user's action is within the allowable deviation, and the corresponding warning is issued to prompt the user, and the coordinate data of the same part of the human body on multiple surfaces are compared to judge whether the user's action is standard, so as to avoid the situation that when multiple points overlap, it is impossible to effectively judge whether the user's action is standard, thereby improving the accuracy of the judgment.
[0063] The above method steps also require the use of a data collection module for collecting input of standard human motion movements, and establishing a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, and correspondingly forming a frontal motion coordinate data set, a left-view motion coordinate data set, and a right-view motion coordinate data set;
[0064] The data acquisition module is used to obtain the frontal control coordinates of the human head, the left hand, the right hand, the left foot and the right foot after the user performs a movement according to the standard human movement movements, obtain the left-view control coordinates of the human head, the left-view control coordinates of the left hand and the left-view control coordinates of the left foot after the user performs a movement according to the standard human movement movements, and obtain the right-view control coordinates of the human head, the right-view control coordinates of the right hand and the right-view control coordinates of the right foot after the user performs a movement according to the standard human movement movements;
[0065] The data analysis module is used to compare and calculate the acquired frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand and the left foot, the right view control coordinates of the human head, the right hand and the right foot with the corresponding frontal action coordinate data set, the left view action coordinate data set and the right view action coordinate data set, and compare the analysis results with the preset threshold value to determine whether the action performed by the user is standard. When non-standard actions occur, the user is prompted to correct the non-standard actions.
[0066] A computer device comprises a memory and a processor. The memory stores a computer program. When the processor executes the computer program, any step of a method for correcting human body motion posture based on computer vision is implemented.
[0067] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for correcting human body motion posture based on computer vision, characterized in that: The following steps are involved: S1. Input a standard human motion and establish a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, wherein the reference coordinate system includes a front reference coordinate system, a left view reference coordinate system and a right view reference coordinate system; S2, obtaining multiple front coordinates of the human head, multiple front coordinates of the left hand, multiple front coordinates of the right hand, multiple front coordinates of the left foot, and multiple front coordinates of the right foot in the standard human motion according to the established front reference coordinate system, obtaining multiple left view coordinates of the human head, multiple left view coordinates of the left hand, and multiple left view coordinates of the left foot in the standard human motion according to the established left view reference coordinate system, and obtaining multiple right view coordinates of the human head, multiple right view coordinates of the right hand, and multiple right view coordinates of the right foot according to the established right view reference coordinate system; S3, establishing a corresponding front action coordinate data set according to a plurality of human head front coordinates, a plurality of left hand front coordinates, a plurality of right hand front coordinates, a plurality of left foot front coordinates and a plurality of right foot front coordinates in a sequence of steps to complete the corresponding action; establishing a left view action coordinate data set according to a plurality of human head left view coordinates, a plurality of left hand left view coordinates and a plurality of left foot left view coordinates in a sequence of steps to complete the corresponding action; establishing a right view action coordinate data set according to a plurality of human head right view coordinates, a plurality of right hand right view coordinates and a plurality of right foot right view coordinates in a sequence of steps to complete the corresponding action; S4, obtaining the frontal control coordinates of the human head, the left hand, the right hand, the left foot and the right foot after each movement completed by the user in accordance with the standard human movement movements, obtaining the left-view control coordinates of the human head, the left-view control coordinates of the left hand and the left-view control coordinates of the left foot after each movement completed by the user in accordance with the standard human movement movements, and obtaining the right-view control coordinates of the human head, the right-view control coordinates of the right hand and the right-view control coordinates of the right foot after each movement completed by the user in accordance with the standard human movement movements; S5, comparing and calculating the obtained frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the right view control coordinates of the human head, the right hand, and the right foot with the corresponding frontal action coordinate data set, the left view action coordinate data set, and the right view action coordinate data set one by one; S6. Compare the analysis result with a preset threshold value to determine whether the action performed by the user is standard. If a non-standard action occurs, prompt the user to correct the non-standard action.
2. The method for correcting human body motion posture based on computer vision according to claim 1, characterized in that: The frontal action coordinate data set expression is: in, Indicates all the acquired human head frontal coordinate data, Indicates all the left-hand front coordinates obtained, Indicates all the right-hand front coordinates obtained, Indicates all the obtained left foot front coordinates, Indicates all the acquired right foot front coordinates.
3. The method for correcting human body motion posture based on computer vision according to claim 2, characterized in that: The left-view action coordinate data set expression is as follows: in, Indicates all the obtained left-view coordinates of the human head. Indicates all the left-hand left-view coordinates obtained. Represents all the obtained left foot left view coordinates.
4. The method for correcting human body motion posture based on computer vision according to claim 3, characterized in that: The right-view action coordinate data set expression is as follows: in, Indicates all the acquired right view coordinates of the human head. Indicates all the right-handed right-view coordinates obtained. Represents all the right foot right view coordinates obtained.
5. The method and system for correcting human body motion posture based on computer vision according to claim 4, characterized in that: The frontal coordinate expression of the human head is: (DZTX i ,DZTY i ), the left hand frontal coordinate expression is: (DZZSX i ,DZZSY i ), the right-hand frontal coordinate expression is: (DZYSX i ,DZYSY i ), the left foot front coordinate expression is: (DZZJX i ,DZZJY i ), the coordinate expression of the right foot front view is: (DZZJX i ,DZZJY i ), the left view coordinate expression of the human head is: (ZZTX i ,ZZTY i ), the left-hand left-view coordinate comparison expression is: (ZZZSX i ,ZZZSY i ), the left foot left view coordinate expression is: (ZZZJX i ,ZZZJY i ), the human head right view coordinate expression is: (ZYTX i , ZYTY i ), the right-hand right-view coordinate expression is: (ZZYSX i ,ZZYSY i ), the right foot right view coordinate expression is: (ZZYJX i ,ZZYJY i ).
6. The method and system for correcting human body motion posture based on computer vision according to claim 5, characterized in that: When any one or more of the frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the left foot, the right view control coordinates of the human head, the right hand, the right foot, the right foot, the right view control coordinates of the human head, the right hand, and the right foot exceed the coordinates of the corresponding steps in the corresponding frontal action coordinate data set, the left view action coordinate data set, or the right view action coordinate data set, it is determined that the user action deviates from the standard human movement action, which is the analysis result.
7. The method and system for correcting human body motion posture based on computer vision according to claim 6, characterized in that: When it is determined that the user's action deviates from the standard action of human body movement, the deviation value is calculated. The calculation expression of the deviation value is as follows: In the formula, AX, AY represent the frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand, the left foot, the left foot, the right view control coordinates of the human head, the right hand, or the right foot, and BX, BY represent the coordinate data in the frontal action coordinate data set, the left view action coordinate data set, or the right view action coordinate data set corresponding to the control coordinates.
8. The method and system for correcting human body motion posture based on computer vision according to claim 7, characterized in that: The expression for determining whether the action performed by the user is standard is as follows: When (1±0.02)CZ>threshold, it is judged that the user's action is not standard.
9. A human body motion posture correction system based on computer vision, based on a human body motion posture correction method based on computer vision according to claims 1 to 8, characterized in that: include: The data collection module is used to collect and input standard human motion movements, and establish a reference coordinate system based on the waist center position of the human body in the standard human motion as the origin, and accordingly form a front motion coordinate data set, a left-view motion coordinate data set, and a right-view motion coordinate data set; Data acquisition module; Used to obtain the frontal control coordinates of the human head, the left hand, the right hand, the left foot and the right foot after the user completes each movement by performing a movement according to the standard human movement movements, obtain the left-view control coordinates of the human head, the left-view control coordinates of the left hand and the left-view control coordinates of the left foot after the user completes each movement by performing a movement according to the standard human movement movements, and obtain the right-view control coordinates of the human head, the right-view control coordinates of the right hand and the right-view control coordinates of the right foot after the user completes each movement by performing a movement according to the standard human movement movements; The data analysis module is used to compare and calculate the acquired frontal control coordinates of the human head, the left hand, the right hand, the left foot, the right foot, the left view control coordinates of the human head, the left hand and the left foot, the right view control coordinates of the human head, the right hand and the right foot with the corresponding frontal action coordinate data set, the left view action coordinate data set and the right view action coordinate data set, and compare the analysis results with the preset threshold value to determine whether the action performed by the user is standard. When non-standard actions occur, the user is prompted to correct the non-standard actions.
10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a method for correcting human motion posture based on computer vision according to any one of claims 1 to 8 are implemented.
Citation Information
Patent Citations
Motion posture correcting system for human body
CN104200491A
Human motion process posture correction method and system based on Internet of Things
CN115497596A
Human body behavior analysis method and system based on joint points
CN117351570A
