A workstation with a sitting posture detection function

Abstract

The present invention discloses a sitting posture detection method based on human key points, which relates to the technical field of human sitting posture detection. The method includes the following steps: obtaining a first image and a first pressure value every first time period; identifying the human key point information in the first image; judging the state of the user according to the first pressure value; in a leisure state, judging whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group; in a working state, judging whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group, and judging whether there is a front-back tilt of the sitting posture according to the position of the human key part and the corresponding height threshold; when it is judged that the sitting posture is tilted left and right and / or front and back, an alarm is issued. The present invention can eliminate the interference of the backward posture relying on the backrest and improve the accuracy of user sitting posture detection.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is a divisional application based on the Chinese patent application with application number 2023107419068, application date: June 21, 2023, and invention name: "A sitting posture detection method, medium and device based on human body key points". Technical Field

[0003] The present invention relates to the technical field of human sitting posture detection, and in particular to a sitting posture detection method, medium and equipment. Background Art

[0004] With the progress of the times, more and more industries are being moved into office buildings. The problem that follows is that sitting at work for a long time has a great impact on the health of workers. Especially in the case of bad sitting posture, it will cause great damage to the workers' neck, shoulders and lumbar spine, leading to the occurrence of various diseases.

[0005] In this case, workers are immersed in their work and often find it difficult to check their sitting posture. Therefore, it is necessary to detect the sitting posture of the workers through external detection and send an alarm to the workers when a bad sitting posture is found. The invention patent with publication number CN 113052097 A discloses a real-time monitoring system and monitoring method for human sitting posture. The detected skeleton image is compared with the skeleton image of a standard normal sitting posture. When the deviation between the two is greater than a predetermined deviation for more than a predetermined time, the non-standard sitting posture is identified and a warning trigger signal is sent.

[0006] However, during the implementation process, the technicians found that when a worker rests on the back of a chair in a standard reclining posture, the skeleton diagram will also be identified as a reclining posture and a warning will be issued. However, in this state, the back of the chair supports the human body and has basically no effect on the human spine. The warning at this time is an interference and will give the worker a wrong instruction.

[0007] Therefore, how to further improve the accuracy of sitting posture detection and eliminate the influence of leaning back in a leisure state is one of the problems that needs to be solved urgently. Summary of the invention

[0008] In order to solve at least one of the technical problems mentioned in the background technology, the purpose of the present invention is to provide a sitting posture detection method, medium and device based on human body key points, which can eliminate the interference caused by leaning back on the chair back and improve the accuracy of user sitting posture detection.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A sitting posture detection method based on human key points, comprising the following steps:

[0011] S1. Obtain a first image and a first pressure value every first time period. The first image contains upper body image information of a user sitting on a seat, and the first pressure value is the pressure borne by the inner side of the backrest of the seat;

[0012] S2. Identify the human key point information in the first image. The human key point information includes the positions of one or more groups of symmetric human key parts on the left and right;

[0013] S3. Judge the state of the user according to the first pressure value. The state includes a leisure state and a working state;

[0014] S4. In the leisure state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group; in the working state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group, and judge whether the sitting posture is tilted forward and backward according to the position of the human key part and the corresponding height threshold;

[0015] S5. When it is judged that the sitting posture is tilted left and right and / or forward and backward, an alarm is issued.

[0016] In some embodiments of the present invention, before S5 issues an alarm, it further includes a verification process:

[0017] S51. When it is judged that the sitting posture is tilted left and right and / or forward and backward, record a first video of the same area as the first image and detect the first pressure value within the corresponding time period, and transmit it to the background;

[0018] S52. The background extracts video frame images every second time period from the first video. The second time period is less than the first time period;

[0019] S53. Perform the judgment process as in S2 to S4 on the video frame images. If the proportion of video frame images with a sitting posture tilted left and right and / or forward and backward in the judgment results exceeds a set proportion threshold, the verification is passed.

[0020] In some embodiments of the present invention, the judgment methods for the leisure state and the working state are as follows:

[0021] In some embodiments of the present invention, the judgment method for whether the sitting posture is tilted left and right is as follows:

[0022] S401. Extract the heights of the positions of two symmetric human key parts on the left and right in the same group;

[0023] S402. Calculate the height difference between the positions of the two human key parts;

[0024] S403. Compare the height difference with the corresponding height difference threshold. If the height difference is greater than the corresponding height difference threshold, it is determined that the sitting posture has a left - right tilt.

[0025] In some embodiments of the present invention, the same group of left - right symmetric human key parts include the left eye and the right eye, and the height difference threshold for the corresponding left eye and right eye is 1 to 2 centimeters; and / or, the same group of left - right symmetric human key parts include the left shoulder and the right shoulder, and the height difference threshold for the corresponding left shoulder and right shoulder is 3 to 6 centimeters.

[0026] In some embodiments of the present invention, the method for judging whether the sitting posture is tilted forward or backward is as follows:

[0027] Compare the first pressure value with the pressure threshold. When the first pressure value is greater than the pressure threshold, it is determined that the user is in a leisure state, otherwise the user is in a working state.

[0028] S411. Extract the heights of one or more groups of positions of human key parts.

[0029] S412. For the two heights of each group of human key parts, take the higher value as the height calibration value of this group of human key parts.

[0030] S413. Compare the height calibration value with the height threshold of the corresponding group of human key parts. If the height calibration value is less than the height threshold, it is determined that the sitting posture has a forward - backward tilt.

[0031] In some embodiments of the present invention, the AlphaPose algorithm is used to identify the human key point information.

[0032] In some embodiments of the present invention, when judging whether the sitting posture has a forward - backward tilt, the height threshold is corrected according to the seat height. The method for correcting the seat height is as follows:

[0033] Adjust the seat to a reference height, and subtract a first value from the height of the human key points where the user sits in the standard sitting posture at this reference height as the height threshold.

[0034] When judging whether the sitting posture has a forward - backward tilt, upload the current height of the seat, calculate the difference between the current height and the reference height, and use the result of adding the difference to the height threshold as the corrected height threshold.

[0035] A computer storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the sitting posture detection method based on human key points as described above.

[0036] A terminal device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned sitting posture detection method based on human key points is implemented.

[0037] A sitting posture detection system based on human key points includes the following modules:

[0038] A first image acquisition module that acquires a first image every first time period, where the first image contains upper body image information of a user sitting on a seat.

[0039] A first pressure value acquisition module that acquires a first pressure value every first time period, where the first pressure value is the pressure borne by the inner side of the backrest of the seat.

[0040] A human key point recognition module that recognizes human key point information in the first image, where the human key point information includes the positions of one or more groups of symmetric human key parts on the left and right.

[0041] A user state judgment module that judges the state of the user according to the first pressure value, where the state includes a leisure state and a working state.

[0042] A sitting posture judgment module that judges whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group; in the working state, judges whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group, and judges whether the sitting posture has a front-back tilt according to the position of the human key part and the corresponding height threshold.

[0043] An alarm module that issues an alarm when it is judged that the sitting posture is tilted left and right and / or front and back.

[0044] In some embodiments of the present invention, it further includes a seat height correction module that corrects the height threshold according to the seat height when judging whether the sitting posture has a front-back tilt.

[0045] A work station with a sitting posture detection function, the work station includes:

[0046] A desk, on which a camera is provided for acquiring a first image every first time period, and the first image contains upper body image information of a user sitting on a seat.

[0047] A lifting seat, a pressure sensor is provided on the inner side of the backrest of the lifting seat for acquiring a first pressure value every first time period, and the first pressure value is the pressure borne by the inner side of the backrest of the seat.

[0048] A controller that combines the first image and the first pressure value and executes the above-mentioned sitting posture detection method.

[0049] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0050] 1. The present invention distinguishes the user's state based on the pressure on the backrest, and in two states, different methods are respectively used to determine whether the user's sitting posture has left-right inclination and / or front-back inclination, thus avoiding misjudgment results caused by normal backward leaning in the leisure state and improving the accuracy of sitting posture recognition.

[0051] 2. On the one hand, the present invention uses low-frequency preliminary recognition at the front end to reduce the computing load at the front end; on the other hand, after the preliminary recognition at the front end passes, high-frequency verification is performed in the background to avoid misjudgment caused by the user's short-term body movements and improve the recognition accuracy; thus achieving a balance between the load and recognition accuracy of the overall method. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall method flowchart of the first embodiment of the present invention.

[0053] Figure 2 It is the verification flowchart of the first embodiment of the present invention.

[0054] Figure 3 It is the flowchart for judging the left-right inclination of the sitting posture in the first embodiment of the present invention.

[0055] Figure 4 It is the flowchart for judging the front-back inclination of the sitting posture in the first embodiment of the present invention.

[0056] Figure 5 It is a side schematic view of the human body sitting posture with forward inclination.

[0057] Figure 6 It is a side schematic view of the human body sitting posture with backward inclination.

[0058] Figure 7 It is a front schematic view of the human body sitting posture with left body inclination.

[0059] Figure 8 It is a front schematic view of the human body sitting posture with right head inclination.

[0060] Figure 9 It is a schematic view of the office workstation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] Example 1:

[0063] Please refer to Figure 1 , this embodiment provides a sitting posture detection method based on human key points, including the following steps:

[0064] S1. Obtain the first image and the first pressure value F every 1 minute. The interval time can also be 2 minutes, 3 minutes, etc.

[0065] The first image contains the upper body image information of the user sitting on the seat, as Figure 9 shown, and is taken by the camera P set on the office desk in front of the user.

[0066] The first pressure value is the pressure borne by the inner side of the seat back, as Figure 9 shown, and is detected by the pressure sensor K set on the inner side of the seat back. The pressure sensor K is set on the upper part of the seat back. Only when the user leans on the seat back as a whole, the pressure sensor K detects the corresponding first pressure value.

[0067] S2. Identify the human key point information in the first image. The human key point information includes the positions of one or more groups of human key parts that are symmetric left and right.

[0068] The left-right symmetry mentioned in this embodiment is the left-right symmetry about the human center line in the standard sitting posture, as Figure 7 shown, including the left eye LE and the right eye RE that are symmetric left and right, and the left shoulder LS and the right shoulder RS that are symmetric left and right.

[0069] In this embodiment, the left eye LE and the right eye RE can be identified separately, and the left shoulder LS and the right shoulder RS can be identified separately. It is preferred to identify the two groups together.

[0070] The identification of human key point information uses the AlphaPose algorithm, which is trained with an image data set in which the left eye LE and the right eye RE, and the left shoulder LS and the right shoulder RS are manually marked.

[0071] S3. Judge the state of the user according to the first pressure value F. The state includes a leisure state and a working state.

[0072] As Figure 5 and Figure 6 shown, when the user is in the standard sitting posture or the forward leaning state, the pressure sensor K does not detect a pressure signal.

[0073] When the user leans backward until the whole body leans on the seat back, the user's back touches the pressure sensor K, and the first pressure value F is detected. It is worth mentioning that when the backward leaning angle of the user does not reach the state of leaning on the seat back as a whole, the pressure sensor K also does not detect a pressure signal, and this backward leaning state lacks the support of the seat back, which will also cause damage to the user's spine.

[0074] Therefore, the method for judging the leisure state and the working state is as follows:

[0075] Set a pressure threshold, such as 40 N. The pressure threshold can be adjusted according to the actual situation. The smaller the threshold, the more sensitive it is, but the easier it is to misjudge.

[0076] Compare the first pressure value F with the pressure threshold. When the first pressure value is greater than the pressure threshold, it is determined that the user is in the leisure state; otherwise, the user is in the working state.

[0077] S4. When the user is in the leisure state and the working state, different methods are used to judge the sitting posture inclination, specifically as follows:

[0078] In the leisure state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two key human body parts in the same group.

[0079] Taking the left shoulder LS and the right shoulder RS as an example for illustration, as Figure 7 shown, after the left shoulder LS and the right shoulder RS are tilted to the left, the left shoulder LS' and the right shoulder RS' are obtained. At the same time, after the left eye LE and the right eye RE are tilted, the left eye LE' and the right eye RE' are obtained.

[0080] As Figure 3 shown, the method for judging whether the sitting posture is tilted left and right is as follows:

[0081] S401. Extract the heights of the positions of the left shoulder LS' and the right shoulder RS' respectively;

[0082] S402. Calculate the height difference ΔS between the positions of the left shoulder LS' and the right shoulder RS';

[0083] S403. Compare the height difference ΔS with the corresponding height difference threshold. If the height difference ΔS is greater than the corresponding height difference threshold, it is determined that there is a left - right tilt in the sitting posture.

[0084] It is worth mentioning that the height difference threshold corresponding to the left shoulder and the right shoulder of this group of human key points is 3 cm, and it can also be 4 cm, 5 cm, or 6 cm.

[0085] In this case, the left eye LE and the right eye RE also tilt left and right synchronously. It is also possible to use the left eye LE' and the right eye RE' instead of the left shoulder LS' and the right shoulder RS' for judgment. Correspondingly, the height difference threshold needs to be adjusted. The height difference threshold corresponding to the left eye and the right eye of this group of human key points is 1 cm, and it can also be 2 cm.

[0086] In another case, as Figure 8As shown, only the left eye LE and the right eye RE are tilted left and right. When judging the left and right tilt with the left eye LE and the right eye, after the left eye LE and the right eye RE are tilted, the left eye LE’’ and the right eye RE’’ are obtained. Using the left eye LE’’ and the right eye RE’’ to replace the left shoulder LS’ and the right shoulder RS’ in S401 to S403 can also detect whether the sitting posture is tilted left and right.

[0087] In the working state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two key human body parts in the same group. The method is the same as that in the leisure state and will not be elaborated here.

[0088] Judge whether the sitting posture has a front-back tilt according to the position of the key human body part and the corresponding height threshold.

[0089] Taking the left shoulder and the right shoulder as an example for illustration, as Figure 4 shown, the method for judging whether the sitting posture has a front-back tilt is as follows:

[0090] S411, extract the heights of the positions of the left shoulder and the right shoulder;

[0091] S412, as Figure 5 shown, for the two heights of the left shoulder and the right shoulder, take the higher one as the current shoulder S’ of the user, and the corresponding height is the height calibration value; Figure 5 In

[0092] S413, compare the height calibration value with the height threshold of the corresponding shoulder. If the height calibration value is less than the height threshold, it is determined that the sitting posture has a front-back tilt.

[0093] The judgment method for the eyes is the same. Figure 5 In

[0094] The judgment of the backward tilt state is the same as that of the forward tilt state. As Figure 6 shown, E’’ and S’’ represent the eyes and shoulders of the user in the backward tilt state respectively.

[0095] In actual application, the lifting height of the seat will also affect the value of the height threshold. Therefore, when judging whether the sitting posture has a front-back tilt, the height threshold needs to be corrected for the seat height. The method for correcting the seat height is as follows:

[0096] Adjust the seat to a reference height, and subtract the first value from the height of the key human body point where the user sits in the standard sitting posture at this reference height as the height threshold;

[0097] When determining whether there is a forward or backward tilt in the sitting posture, upload the current height of the seat, calculate the difference between the current height and the reference height, and use the result of adding the height threshold to the difference as the corrected height threshold.

[0098] Through the above settings, the influence of different seat heights on the judgment method can be adapted.

[0099] S5. When it is determined that there is a left - right tilt and / or a forward - backward tilt in the sitting posture, send a corresponding alarm to remind the user to correct the sitting posture accordingly.

[0100] In the above method, judging the interval time only based on the first image and the first pressure value may cause the user's movement within an extremely short time to exactly hit the time node when the first image is obtained, resulting in misjudgment.

[0101] Therefore, in this embodiment, verification is also performed before the alarm is sent in S5, as Figure 2 shown, and the verification process is as follows:

[0102] S51. When it is determined that there is a left - right tilt and / or a forward - backward tilt in the sitting posture, record the first video of the same area as the first image and detect the first pressure value within the corresponding time period, and transmit them to the background;

[0103] S52. The background extracts video frame images at every second time period from the first video. The second time period is less than the first time period, and the second time period can be 10s or 5s to achieve high - frequency recognition.

[0104] S53. Perform the judgment process from S2 to S4 on the video frame images. If the proportion of the video frame images with left - right tilt and / or forward - backward tilt in the sitting posture in the judgment results exceeds the set proportion threshold, the verification passes.

[0105] For example, if the length of the first video is 2 minutes and video frame images are extracted every 10s, a total of 13 video frame images can be extracted.

[0106] Set the proportion threshold to 80% (13 * 80% = 10.4). When the proportion of the video frame images with left - right tilt and / or forward - backward tilt in the sitting posture respectively exceeds 10, the verification of left - right tilt and / or forward - backward tilt in the sitting posture passes, and a corresponding alarm is sent.

[0107] Through the low - frequency recognition at the front - end and the high - frequency recognition at the background, the load and recognition accuracy of the overall method can be balanced.

[0108] Embodiment 2:

[0109] This embodiment provides a sitting posture detection system based on human key points, including the following modules:

[0110] A first image acquisition module that acquires a first image every first time period, where the first image contains upper body image information of a user sitting on a seat;

[0111] A first pressure value acquisition module that acquires a first pressure value every first time period, where the first pressure value is the pressure borne by the inner side of the seat back;

[0112] A human body key point recognition module that recognizes human body key point information in the first image, where the human body key point information includes the positions of one or more groups of symmetric human body key parts on the left and right;

[0113] A user state judgment module that judges the state of the user according to the first pressure value, where the state includes a leisure state and a working state;

[0114] A sitting posture judgment module that judges whether the sitting posture is tilted left and right according to the height difference between the positions of two human body key parts in the same group; in the working state, it judges whether the sitting posture is tilted left and right according to the height difference between the positions of two human body key parts in the same group, and judges whether there is a front - back tilt of the sitting posture according to the position of the human body key part and the corresponding height threshold

[0115] An alarm module that issues an alarm when it judges that the sitting posture is tilted left and right and / or tilted front and back.

[0116] In another embodiment, the sitting posture detection system based on human body key points further includes:

[0117] A seat height correction module that corrects the height threshold for the seat height when judging whether there is a front - back tilt of the sitting posture.

[0118] The judgment methods for the above - mentioned leisure state and working state, whether the sitting posture is tilted left and right, whether the sitting posture is tilted front and back, and the seat height correction method are the same as those in Embodiment 1, and will not be elaborated here.

[0119] Embodiment 3:

[0120] This embodiment provides a computer storage medium, on which a computer program is stored, and is characterized in that when the program is executed by a processor, it implements the sitting posture detection method based on human body key points as described in Embodiment 1.

[0121] Embodiment 4:

[0122] This embodiment provides a terminal device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and is characterized in that when the processor executes the computer program, it implements the sitting posture detection method based on human body key points as described in Embodiment 1.

[0123] Embodiment 5:

[0124] This embodiment provides a work station with a sitting posture detection function, as Figure 9 shown, the work station includes:

[0125] A desk, on which a camera P is provided for obtaining a first image every first time period, and the first image includes upper body image information of a user sitting on a seat.

[0126] A liftable seat, and a pressure sensor K is provided inside the backrest of the liftable seat for obtaining a first pressure value every first time period, and the first pressure value is the pressure borne by the inside of the backrest of the seat.

[0127] A controller, which combines the first image and the first pressure value and executes the sitting posture detection method as described in Embodiment 1.

[0128] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced by the present invention.

Claims

1. A work station with a sitting posture detection function, characterized in that, The work station includes: A desk, on which a camera is provided for acquiring a first image every first time period, the first image including upper body image information of a user sitting on a seat; A liftable seat, on the inner side of the backrest of the liftable seat, a pressure sensor is provided for acquiring a first pressure value every first time period, the first pressure value being the pressure borne by the inner side of the backrest of the seat; A controller, which combines the first image and the first pressure value and executes a sitting posture detection method, including the following steps: S1. Acquire the first image and the first pressure value every first time period; S2. Identify the human key point information in the first image, the human key point information including the positions of one or more groups of symmetric human key parts on the left and right; S3. Judge the state of the user according to the first pressure value, the state including a leisure state and a working state; S4. In the leisure state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group; in the working state, judge whether the sitting posture is tilted left and right according to the height difference between the positions of two human key parts in the same group, and judge whether the sitting posture is tilted forward and backward according to the position of the human key part and the corresponding height threshold; S5. When it is judged that the sitting posture is tilted left and right and / or forward and backward, an alarm is issued.

2. The workstation with a sitting posture detection function according to claim 1, wherein Before S5 issues an alarm, it also includes a verification process: S51. When it is judged that the sitting posture is tilted left and right and / or forward and backward, record a first video of the same area as the first image and detect the first pressure value during the corresponding time period, and transmit it to the background; S52. The background extracts video frame images every second time period from the first video, the second time period being less than the first time period; S53. Perform the judgment process as in S2 to S4 on the video frame images. If the proportion of video frame images with a sitting posture tilted left and right and / or forward and backward in the judgment result exceeds the set proportion threshold, the verification is passed.

3. The workstation with a sitting posture detection function according to claim 1, characterized in that, The judgment methods for the leisure state and the working state are as follows: Compare the first pressure value with the pressure threshold. When the first pressure value is greater than the pressure threshold, it is determined that the user is in the leisure state, otherwise the user is in the working state.

4. A workstation with a sitting posture detection function according to claim 1, wherein The judgment method for whether the sitting posture is tilted left and right is as follows: S401. Extract the heights of the positions of two symmetric human key parts on the left and right in the same group; S402. Calculate the height difference between the positions of the two human key parts; S403. Compare the height difference with the corresponding height difference threshold. If the height difference is greater than the corresponding height difference threshold, it is determined that the sitting posture is tilted left and right.

5. A workstation with a sitting posture detection function according to claim 4, characterized in that, The two symmetric human key parts in the same group include the left eye and the right eye, and the corresponding height difference threshold for the left eye and the right eye is 1 to 2 cm; and / or, the two symmetric human key parts in the same group include the left shoulder and the right shoulder, and the corresponding height difference threshold for the left shoulder and the right shoulder is 3 to 6 cm.

6. A work station with a sitting posture detection function according to claim 1, characterized in that, The judgment method for whether the sitting posture is tilted forward and backward is as follows: S411. Extract the heights of the positions of one or more groups of human key parts; S412. For the two heights of each group of human key parts, take the higher one as the height calibration value of the group of human key parts; S413. Compare the height calibration value with the height threshold of the corresponding set of human key parts. If the height calibration value is less than the height threshold, it is determined that the sitting posture has a front-back tilt.

7. A work station with a sitting posture detection function according to claim 1, characterized in that, The AlphaPose algorithm is used to identify the human key point information.

8. A workstation with a sitting posture detection function according to claim 1, characterized in that, When judging whether the sitting posture has a front-back tilt, the height threshold is corrected according to the seat height. The method for correcting the seat height is as follows: Adjust the seat to a reference height. When the user sits in the standard sitting posture at this reference height, subtract the first value from the height of the human key points at this reference height as the height threshold. When judging whether the sitting posture has a front-back tilt, upload the current height of the seat, calculate the difference between the current height and the reference height, and use the result of adding the difference to the height threshold as the corrected height threshold.

Images