A method, apparatus, device, and medium for testing squat count

Abstract

The application provides a deep squat test counting method, device, equipment and medium, which can monitor the normativity of the movement process in real time, whether the hands are kept straight, whether the thighs are parallel to the ground during the knee bending process, and whether the thighs are straight during the knee stretching process. Through detection of illegal actions and reminders, the tester can exercise more effectively, and only the deep squats with normative actions are counted during the counting process, so that the counting is more accurate.

Description

Technical Field

[0001] This invention relates to a method, apparatus, device, and medium for testing and counting squats. Background Technology

[0002] With social and economic development, people are increasingly emphasizing physical exercise. Squats, as an effective but challenging exercise, can effectively improve flexibility, strength, and joint stability, promoting coordinated physical development. However, traditional squat counting methods are highly subjective, still counting squats with incorrect form as valid, which is detrimental to squat accuracy. Furthermore, the lack of reminders about incorrect form increases the risk of injury.

[0003] Currently, some systems use cameras combined with artificial intelligence technology for counting, but lack detection of whether the movement process is standardized. Application number CN202110710542.8, entitled "Squat Exercise Counting Method, System, Computer Equipment, and Storage Medium," discloses a counting method that determines the threshold baseline in the movement image through device parameters; when the baseline is crossed twice consecutively, it is counted once. The technical solution in the prior art has the following two problems: 1. When the tester is changed, the height ratio is different, requiring the threshold baseline to be reset, which leads to poor adaptability; 2. The counting method counts once for crossing the baseline twice; if the user does not return to the initial state but repeatedly crosses the baseline, it will continue to count. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a squat test counting method, device, equipment and medium, which enables the test subject to exercise more effectively by detecting and reminding the test subject of improper movements, and makes the counting more accurate by only counting squats with proper movements.

[0005] In a first aspect, the present invention provides a method for counting squat tests, comprising:

[0006] Step 1: Acquire the video stream of the tester in real time, input the image of each frame of the video stream into the yoloE human object detection model for prediction, and acquire the coordinates of human key points through the Hrnet human pose estimation model; obtain human body state data; the human body state data includes: the angle between the left elbow-shoulder and the ground is recorded as lElowAngel, the angle between the right elbow-shoulder and the ground is recorded as rElowAngel, the angle between the left hip-knee and the ground is recorded as lHipAngel, and the angle between the right hip-knee and the ground is recorded as rHipAngel; when the human body state data meets the set conditions, prompt that the assessment starts; the set thresholds are defined as follows: straight both hands threshold is recorded as Acc1, elbow-shoulder parallel to ground threshold is recorded as Acc2, knee-bent state threshold is recorded as Acc3, effective knee-bent threshold is recorded as Acc3V, knee-extended state threshold is recorded as Acc4, effective knee-extended threshold is recorded as Acc4V, knee-extended state is recorded as A, knee-bent state is recorded as B, and the last effective counting time is recorded as Tlast;

[0007] Step 2: When the assessment starts, enter the knee-extended state A; detect whether (lArmAngel>Acc1 and rArmAngel>Acc1) and (lElowAngel<Acc2 and rElowAngel<Acc2) are satisfied; if yes, proceed to step 3; if not, prompt the violation that both hands are not straightened, and proceed to step 1;

[0008] Step 3: Detect whether (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, if yes, enter the knee-bent state B;

[0009] Step 4: In state B, detect whether (lHipAngel+rHipAngel) / 2<Acc3V is satisfied; if this condition is satisfied, and (lHipAngel+rHipAngel) / 2>Acc4 is also satisfied, enter the knee-extended state A and proceed to step 5;

[0010] If the above condition is not satisfied, but (lHipAngel+rHipAngel) / 2>Acc4 is satisfied, prompt the violation that the thigh is not parallel to the ground, and proceed to step 1;

[0011] Step 5: In state A, detect whether (lHipAngel+rHipAngel) / 2>Acc4V is satisfied; if yes, increment the effective count by 1, and update Tlast to the current time;

[0012] If the above condition is not satisfied, but (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, prompt the violation that the thigh is not straightened, and proceed to step 1;

[0013] Step 6: Repeat the above steps until the test is completed.

[0014] In a second aspect, the present invention provides a squat test counting device, comprising:

[0015] a video stream acquisition module, configured to acquire a video stream of a tester in real time, input an image of each frame of the video stream into a yoloE human object detection model for prediction, and acquire coordinates of human key points through a Hrnet human pose estimation model; acquire human body state data; the human body state data comprises: an included angle between the left elbow-shoulder and the ground is denoted as lElowAngel, an included angle between the right elbow-shoulder and the ground is denoted as rElowAngel, an included angle between the left hip-knee and the ground is denoted as lHipAngel, and an included angle between the right hip-knee and the ground is denoted as rHipAngel; when the human body state data meets a set condition, a prompt that the assessment starts is given; a straight arm threshold is set and denoted as Acc1, an elbow-shoulder parallel to the ground threshold is set and denoted as Acc2, a knee flexion state threshold is set and denoted as Acc3, an effective knee flexion threshold is set and denoted as Acc3V, a knee extension state threshold is set and denoted as Acc4, an effective knee extension threshold is set and denoted as Acc4V, a knee extension state is denoted as A, a knee flexion state is denoted as B, and the last effective counting time is denoted as Tlast;

[0016] a squatting process module, configured to enter the knee extension state A when the assessment starts; detect whether (lArmAngel>Acc1 and rArmAngel>Acc1) and (lElowAngel<Acc2 and rElowAngel<Acc2) are satisfied; if yes, proceed to step 3; if not, a prompt of violation for unstraightened double hands is given, and the process returns to the video stream acquisition module;

[0017] a squat arrival module, configured to detect whether (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, and enter the knee flexion state B if the condition is satisfied;

[0018] a standing up process module, configured to, in the state B, detect whether (lHipAngel+rHipAngel) / 2<Acc3V is satisfied; if the condition is satisfied, and (lHipAngel+rHipAngel) / 2>Acc4 is satisfied, enter the knee extension state A, and enter a standing completion counting module;

[0019] if the condition is not satisfied but (lHipAngel+rHipAngel) / 2>Acc4 is satisfied, a violation prompt that the thigh is not parallel to the ground is given, and the process enters the video stream acquisition module;

[0020] A get-up completion counting module, when in state A, detects whether (lHipAngel+rHipAngel) / 2>Acc4V is satisfied; if the condition is satisfied, add 1 to the valid count and update Tlast to the current time; if the condition is not satisfied but (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, a violation prompt for unstraightened thighs is given, and the process enters the video stream acquisition module;

[0021] A counting completion module, which repeats the above modules until the test is completed.

[0022] In a third aspect, the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the method according to the first aspect when executing the program.

[0023] In a fourth aspect, the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method according to the first aspect.

[0024] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0025] No manual intervention is required during the test, which improves the test efficiency; the detection of whether the movement process is standard is realized, and corresponding prompts are given, which improves the efficiency and the exercise effect.

[0026] The above description is only an overview of the technical solutions of the present invention. In order to more clearly understand the technical means of the present invention and implement them in accordance with the content of the description, and in order to make the above and other objects, features and advantages of the present invention more obviously understandable, the specific embodiments of the present invention are specifically described below. Description of Drawings

[0027] The present invention will be further described below with reference to the accompanying drawings in combination with embodiments.

[0028] Figure 1 is a flow chart of the initial detection process of the present invention;

[0029] Figure 2 is a flow chart of the violation detection and counting process of the present invention;

[0030] Figure 3 is a flow chart of the end detection process of the present invention. Detailed Description of the Embodiments

[0031] The idea of the embodiments of the present application:

[0032] 1. As Figure 1 shown, the initial detection process:

[0033] 1. The camera is placed directly in front of the tester, and the tester is positioned to the side to ensure that the entire movement process of the tester can be captured.

[0034] 2. Real-time acquisition of video streams, input of images into the YOLOE human target detection model for prediction, and acquisition of the coordinates of human key points through the Hrnet human pose estimation model.

[0035] 3. The coordinates of the relevant key points are as follows: left wrist point lWrist, right wrist point rWrist, left elbow point lElbow, right elbow point rElbow, left shoulder point lShoulder, right shoulder point rShoulder, left hip point lHip, right hip point rHip, left knee point lKnee, and right knee point rKnee. The following angles are formed by connecting these points; for example, the left wrist-elbow-shoulder angle is the angle between the lines connecting the left wrist point and the left elbow point and the lines connecting the left elbow point and the left shoulder point.

[0036] The left wrist-elbow-shoulder angle is denoted as lArmAngel, and the right wrist-elbow-shoulder angle is denoted as rArmAngel;

[0037] The angle between the left elbow / shoulder and the ground is denoted as lElowAngel, and the angle between the right elbow / shoulder and the ground is denoted as rElowAngel.

[0038] The angle between the left hip and knee and the ground is denoted as lHipAngel, and the angle between the right hip and knee and the ground is denoted as rHipAngel.

[0039] The threshold for fully extended arms is recorded as Acc1 = 150;

[0040] The threshold for elbow and shoulder to be parallel to the ground is denoted as Acc2 = 30.

[0041] The knee flexion threshold is denoted as Acc3 = 30, and the effective knee flexion threshold is denoted as Acc3V = 10.

[0042] The knee extension threshold is denoted as Acc4 = 60, and the effective knee extension threshold is denoted as Acc4V = 80.

[0043] The extended knee position is denoted as A, and the flexed knee position is denoted as B;

[0044] The last valid count time is Tlast;

[0045] 4. The assessment will begin when all of the following conditions are met simultaneously;

[0046] 1) lArmAngel>Acc1&&rArmAngel>Acc1 (both arms extended);

[0047] 2) lElowAngel<Acc2 && rElowAngel<Acc2 (both hands stretch out horizontally);

[0048] 3) lHipAngel>Acc4V && rHipAngel>Acc4V (knee extension is greater than the effective threshold);

[0049] 2. As Figure 2 shown, the violation detection and counting process is:

[0050] 1. After waiting for the completion of initialization, enter knee extension state A;

[0051] 2. Detect throughout the whole process whether (lArmAngel>Acc1 && rArmAngel>Acc1) &&

[0052] (lElowAngel<Acc2 && rElowAngel<Acc2) is satisfied;

[0053] 2.1 If satisfied: proceed to step 3;

[0054] 2.2 If not satisfied: give a violation prompt that both hands are not straightened, and proceed to step 1;

[0055] 3. Detect whether (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, if yes, enter knee flexion state B;

[0056] 4. In state B, detect whether (lHipAngel+rHipAngel) / 2<Acc3V is satisfied;

[0057] 4.1 If satisfied, and (lHipAngel+rHipAngel) / 2>Acc4 is satisfied, enter knee extension state A, and proceed to step 5;

[0058] 4.2 If not satisfied, and (lHipAngel+rHipAngel) / 2>Acc4 is satisfied, give a violation prompt that the thigh is not parallel to the ground, and proceed to step 1;

[0059] 5. In state A, detect whether (lHipAngel+rHipAngel) / 2>Acc4V is satisfied;

[0060] 5.1 If the condition is satisfied: add 1 to the effective count, and update Tlast to the current time;

[0061] 5.2 If the condition is not satisfied, and (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, give a violation prompt that the thigh is not straightened, and proceed to step 1;

[0062] 3. As Figure 3 As shown, end the detection process:

[0063] When any of the following conditions is met, it is prompted that the assessment is ended.

[0064] 1) The tester leaves the lens;

[0065] 2) Current time - Tlast > 10S.

[0066] It can monitor the standardization of the movement process in real time, such as whether both hands are kept straight throughout the whole process, whether the thigh is parallel to the ground during knee bending, and whether the thigh is straight during knee extension; by detecting irregular movements and sending reminders, the tester can exercise more effectively and protect themselves.

[0067] First Embodiment

[0068] This embodiment provides a test counting method for deep squats, comprising:

[0069] Step 1: Acquire the video stream of the tester in real time, input the image of each frame of the video stream into the yoloE human object detection model for prediction, and acquire the coordinates of human key points through the Hrnet human pose estimation model; acquire human state data; the human state data comprises: the included angle between the left elbow and shoulder and the ground is recorded as lElowAngel, the included angle between the right elbow and shoulder and the ground is recorded as rElowAngel, the included angle between the left hip and knee and the ground is recorded as lHipAngel, and the included angle between the right hip and knee and the ground is recorded as rHipAngel; a straight both hands threshold is set as Acc1, an elbow-shoulder parallel to ground threshold is set as Acc2, a knee flexion state threshold is set as Acc3, an effective knee flexion threshold is set as Acc3V, a knee extension state threshold is set as Acc4, an effective knee extension threshold is set as Acc4V, the knee extension state is recorded as A, the knee flexion state is recorded as B, and the last effective counting time is recorded as Tlast; the acquiring of human key points through the Hrnet human pose estimation model comprises left wrist point lWrist, right wrist point rWrist, left elbow point lElbow, right elbow point rElbow, left shoulder point lShoulder, right shoulder point rShoulder, left hip point lHip, right hip point rHip, left knee point lKnee and right knee point rKnee,

[0070] When the human state data simultaneously satisfy the following conditions, it is prompted that the assessment starts;

[0071] 1) lArmAngel>Acc1 and rArmAngel>Acc1;

[0072] 2) lElowAngel<Acc2 and rElowAngel<Acc2;

[0073] 3) lHipAngel > Acc4V and rHipAngel > Acc4V;

[0074] Step 2, when the assessment starts, enter the knee extension state A; detect whether (lArmAngel > Acc1 and rArmAngel > Acc1) and (lElowAngel < Acc2 and rElowAngel < Acc2) are satisfied; if satisfied, proceed to step 3; if not satisfied, prompt that the unstraightened both hands is a violation, and proceed to step 1;

[0075] Step 3, detect whether (lHipAngel + rHipAngel) / 2 < Acc3 is satisfied, if satisfied, enter the knee flexion state B;

[0076] Step 4, in state B, detect whether (lHipAngel + rHipAngel) / 2 < Acc3V is satisfied; if satisfied, and (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, enter the knee extension state A, and proceed to step 5;

[0077] if not satisfied, but (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, give a violation prompt that the thigh is not parallel to the ground, and proceed to step 1;

[0078] Step 5, in state A, detect whether (lHipAngel + rHipAngel) / 2 > Acc4V is satisfied; if satisfied, increase the valid count by 1, and update Tlast to the current time;

[0079] if not satisfied, but (lHipAngel + rHipAngel) / 2 < Acc3 is satisfied, give a violation prompt that the thigh is not straightened, and proceed to step 1;

[0080] Step 6, repeat the above steps, when the tester leaves the camera or the current time minus Tlast is greater than the set time threshold, prompt that the assessment is ended.

[0081] Based on the same inventive concept, the present application further provides an apparatus corresponding to the method in the first embodiment, refer to the second embodiment for details.

[0082] Second Embodiment

[0083] In this embodiment, a squat test counting apparatus is provided, comprising:

[0084] A video stream acquisition module, which acquires a video stream of a tester in real time, inputs an image of each frame of the video stream into a yoloE human body target detection model for prediction, and acquires coordinates of human body key points through a Hrnet human body pose estimation model; acquires human body state data; the human body state data includes: an angle between a left elbow-shoulder and the ground is recorded as lElowAngel, an angle between a right elbow-shoulder and the ground is recorded as rElowAngel, an angle between a left hip-knee and the ground is recorded as lHipAngel, and an angle between a right hip-knee and the ground is recorded as rHipAngel; a threshold for straight both hands is set as Acc1, a threshold for elbow-shoulder parallel to the ground is set as Acc2, a threshold for knee flexion state is set as Acc3, an effective threshold for knee flexion is set as Acc3V, a threshold for knee extension state is set as Acc4, an effective threshold for knee extension is set as Acc4V, a knee extension state is recorded as A, a knee flexion state is recorded as B, and a last effective counting time is recorded as Tlast; the human body key points acquired through the Hrnet human body pose estimation model include a left wrist point lWrist, a right wrist point rWrist, a left elbow point lElbow, a right elbow point rElbow, a left shoulder point lShoulder, a right shoulder point rShoulder, a left hip point lHip, a right hip point rHip, a left knee point lKnee and a right knee point rKnee,

[0085] When the human body state data simultaneously meets the following conditions, a prompt that the assessment starts is given;

[0086] 1) lArmAngel>Acc1 and rArmAngel>Acc1;

[0087] 2) lElowAngel<Acc2 and rElowAngel<Acc2;

[0088] 3) lHipAngel>Acc4V and rHipAngel>Acc4V;

[0089] A squatting process module, when the assessment starts, enters the knee extension state A; detects whether (lArmAngel>Acc1 and rArmAngel>Acc1) and (lElowAngel<Acc2 and rElowAngel<Acc2) are satisfied; if the conditions are satisfied, the process proceeds to step 3; if the conditions are not satisfied, a prompt that both hands are not straightened and the operation is violated is given, and the process returns to the video stream acquisition module;

[0090] A squat arrival module, detects whether (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, if the condition is satisfied, the process enters the knee flexion state B;

[0091] Standing up process module: in state B, detecting whether (lHipAngel + rHipAngel) / 2 < Acc3V is satisfied; if the condition is satisfied and (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, enter the knee extension state A and the standing completion counting module;

[0092] If the condition is not satisfied, but (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, give an illegal prompt that the thigh is not parallel to the ground, and enter the video stream acquisition module;

[0093] Standing completion counting module: in state A, detecting whether (lHipAngel + rHipAngel) / 2 > Acc4V is satisfied; if the condition is satisfied, increment the valid count by 1 and update Tlast to the current time; if the condition is not satisfied, but (lHipAngel + rHipAngel) / 2 < Acc3 is satisfied, give an illegal prompt that the thigh is not fully extended, and enter the video stream acquisition module;

[0094] Counting completion module: repeating the above modules, and when the tester leaves the camera or the current time minus Tlast is greater than the set time threshold, prompting that the assessment is completed.

[0095] Since the device introduced in the second embodiment of the present invention is a device adopted to implement the method of the first embodiment of the present invention, those skilled in the art can understand the specific structure and variations of the device based on the method introduced in the first embodiment of the present invention, thus detailed description is omitted here. All devices adopted for implementing the method of the first embodiment of the present invention fall within the protection scope intended by the present invention.

[0096] Based on the same inventive concept, the present application provides an electronic device embodiment corresponding to the first embodiment, refer to the third embodiment for details.

[0097] Embodiment 3

[0098] This embodiment provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the computer program, any implementation manner in the first embodiment can be implemented.

[0099] Since the electronic device introduced in this embodiment is the device adopted for implementing the method in the first embodiment of the present application, based on the method introduced in the first embodiment of the present application, those skilled in the art can understand the specific implementation manner of the electronic device in this embodiment and various variations thereof, so detailed description of how the electronic device implements the method in the embodiments of the present application is omitted herein. All devices adopted by those skilled in the art for implementing the method in the embodiments of the present application fall within the protection scope intended by the present application.

[0100] Based on the same inventive concept, this application provides a storage medium corresponding to Embodiment 1, as detailed in Embodiment 4.

[0101] Example 4

[0102] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it can implement any of the implementation methods in Embodiment 1.

[0103] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the present invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method of testing squat count, characterized by, Comprising: Step 1: Acquiring the video stream of a tester in real time, inputting the image of each frame of the video stream into a yoloE human object detection model for prediction, and acquiring coordinates of human key points through an Hrnet human pose estimation model; Acquiring human body state data; The human body state data comprises: the included angle between the left elbow-shoulder and the ground is recorded as lElowAngel, the included angle between the right elbow-shoulder and the ground is recorded as rElowAngel, the included angle between the left hip-knee and the ground is recorded as lHipAngel, and the included angle between the right hip-knee and the ground is recorded as rHipAngel; prompting the start of assessment when the human body state data meets a set condition; a set straight both hands threshold is recorded as Acc1, an elbow-shoulder parallel to ground threshold is recorded as Acc2, a knee bending state threshold is recorded as Acc3, an effective knee bending threshold is recorded as Acc3V, a knee stretching state threshold is recorded as Acc4, an effective knee stretching threshold is recorded as Acc4V, a knee stretching state is recorded as A, a knee bending state is recorded as B, and a last effective counting time is recorded as Tlast; Step 2: Entering the knee stretching state A when the assessment starts; detecting whether (lArmAngel>Acc1 and rArmAngel>Acc1) and (lElowAngel<Acc2 and rElowAngel<Acc2) are satisfied; if yes, entering step 3; if no, giving a violation prompt that both hands are not straightened, and entering step 1; Step 3: detecting whether (lHipAngel + rHipAngel) / 2<Acc3 is satisfied, and entering the knee bending state B if satisfied; Step 4: in the state B, detecting whether (lHipAngel + rHipAngel) / 2<Acc3V is satisfied; if satisfied and (lHipAngel + rHipAngel) / 2>Acc4 is satisfied, entering the knee stretching state A, and proceeding to step 5; If the above condition is not satisfied but (lHipAngel + rHipAngel) / 2>Acc4 is satisfied, giving a violation prompt that the thigh is not parallel to the ground, and entering step 1; Step 5: in the state A, detecting whether (lHipAngel + rHipAngel) / 2>Acc4V is satisfied; if satisfied, adding 1 to the effective count and updating Tlast to the current time; If the above condition is not satisfied but (lHipAngel + rHipAngel) / 2<Acc3 is satisfied, giving a violation prompt that the thigh is not straightened, and entering step 1; Step 6: repeating the above steps until the test is completed.

2. The squat test counting method according to claim 1, characterized in that, in step 1, prompting the start of assessment when the human body state data meets the set condition is further specifically: prompting the start of assessment when the human body state data simultaneously satisfies the following conditions; 1) lArmAngel>Acc1 and rArmAngel>Acc1; 2) lElowAngel<Acc2 and rElowAngel<Acc2; 3) lHipAngel>Acc4V and rHipAngel>Acc4V.

3. The squat test counting method according to claim 1, characterized in that, Said step 6 is further specifically as follows: repeating the above steps, and when the tester leaves the lens or the current time minus Tlast is greater than a set time threshold, a prompt that the assessment is finished is given.

4. The squat test counting method according to claim 1, characterized in that, Said acquiring human body key points through the Hrnet human pose estimation model comprises a left wrist point lWrist, a right wrist point rWrist, a left elbow point lElbow, a right elbow point rElbow, a left shoulder point lShoulder, a right shoulder point rShoulder, a left hip point lHip, a right hip point rHip, a left knee point lKnee and a right knee point rKnee.

5. A squatting test counting device, characterized in that, comprises:[rough translation] a video stream acquisition module, configured to acquire a video stream of a tester in real time, input an image of each frame of the video stream into a yoloE human object detection model for prediction, and acquire coordinates of human body key points through the Hrnet human pose estimation model; acquiring human body state data; Said human body state data comprises: an included angle between the left elbow-shoulder and the ground is recorded as lElowAngel, an included angle between the right elbow-shoulder and the ground is recorded as rElowAngel, an included angle between the left hip-knee and the ground is recorded as lHipAngel, and an included angle between the right hip-knee and the ground is recorded as rHipAngel; when the human body state data conforms to a set condition, a prompt that the assessment starts is given; a straight double-hand threshold is set as Acc1, an elbow-shoulder parallel to ground threshold is set as Acc2, a knee-bent state threshold is set as Acc3, a valid knee-bent threshold is set as Acc3V, a knee-stretched state threshold is set as Acc4, a valid knee-stretched threshold is set as Acc4V, a knee-stretched state is recorded as A, a knee-bent state is recorded as B, and the last valid counting time is recorded as Tlast; a squatting process module, configured to enter the knee-stretched state A when the assessment starts; detect whether (lArmAngel > Acc1 and rArmAngel > Acc1) and (lElowAngel < Acc2 and rElowAngel < Acc2) are satisfied; if yes, proceed to step 3; if not, prompt that the violation of unstraightened double hands occurs, and return to the video stream acquisition module; a squat arrival module, configured to detect whether (lHipAngel + rHipAngel) / 2 < Acc3 is satisfied, and enter the knee-bent state B if the condition is satisfied; a standing-up process module, configured to detect whether (lHipAngel + rHipAngel) / 2 < Acc3V is satisfied when in state B; if the condition is satisfied and (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, enter the knee-stretched state A and proceed to a standing-up completion counting module; if the condition is not satisfied but (lHipAngel + rHipAngel) / 2 > Acc4 is satisfied, a violation prompt that the thigh is not parallel to the ground is given, and the process returns to the video stream acquisition module; Stand-up completion counting module, when in state A, detecting whether (lHipAngel+rHipAngel) / 2>Acc4V is satisfied; if yes, increment the valid count by 1 and update Tlast to the current time; if not, and (lHipAngel+rHipAngel) / 2<Acc3 is satisfied, giving an illegal prompt that the thighs are not straightened, and entering the video stream acquisition module; Counting completion module, repeating the above modules until the test is completed.

6. The squat test counting device according to claim 5, characterized in that, When the human body state data meets the set conditions in the video stream acquisition module, prompting that the assessment starts is further specifically: when the human body state data meets the following conditions at the same time, prompting the start of the assessment; 1) lArmAngel>Acc1 and rArmAngel>Acc1; 2) lElowAngel<Acc2 and rElowAngel<Acc2; 3) lHipAngel>Acc4V and rHipAngel>Acc4V.

7. The squat test counting device according to claim 5, characterized in that, The counting completion module is further specifically: repeating the above modules, when the tester leaves the lens or the current time minus Tlast is greater than the set time threshold, prompting that the assessment is ended.

8. A squat testing and counting device according to claim 5, characterized in that, Acquiring human body key points through the Hrnet human pose estimation model includes left wrist point lWrist, right wrist point rWrist, left elbow point lElbow, right elbow point rElbow, left shoulder point lShoulder, right shoulder point rShoulder, left hip point lHip, right hip point rHip, left knee point lKnee and right knee point rKnee.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, the method according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images