Artificial intelligence monitoring motion system for high-altitude work training and assessment

Abstract

The present invention discloses an artificial intelligence monitoring action system for high-altitude operation training and assessment, which relates to the field of image processing technology and includes a visual monitoring module, a danger determination module, and an early warning module. The visual monitoring module is used to monitor and identify the actions of assessors by using a method of multiple cameras looking around. The danger determination module is used to determine the body center of gravity position of the current assessor and whether the current state of both hands being away is in a dangerous situation. The early warning module is used to warn the assessor when he is in a dangerous action and posture. The visual monitoring module includes a camera, an image splicing module, an object recognition module, a personnel detection module, an action recognition module, and a data storage module. The camera is electrically connected to the image splicing module. A plurality of the cameras cover various heights of the assessment area and shoot the assessment area. The present invention has the characteristic of being able to be specifically analyzed.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to an artificial intelligence monitoring action system for high-altitude operation training and assessment. Background Art

[0002] Working at height refers to working at a height that is more than 2 meters (including 2 meters) above the falling height reference plane, where there is a possibility of falling. In order to ensure the personal safety of workers working at height, it is stipulated that workers engaged in working at height must undergo special safety training and pass the assessment to obtain the corresponding qualification certificate. During the training and assessment, the artificial intelligence motion monitoring system can replace the human eye for monitoring and issue early warnings for illegal actions.

[0003] When working at heights, hands are important support points for maintaining body balance. When both hands are separated from the support, the body's balance will decrease, and it is easy to shake or even fall. Therefore, the principle of not vacating the body unless necessary should be followed. However, in some high-altitude installation operations, it is necessary to free up the hands briefly to operate the tools. The existing monitoring system will trigger an early warning as long as the hands are released, or the early warning will be triggered after the set time is released. It cannot be judged based on the specific content of the operation, and whether the center of gravity of the human body is in a safe position is also related to whether the hands can be released. The existing technology cannot specifically analyze whether it is safe. Therefore, it is necessary to design an artificial intelligence monitoring action system for high-altitude operation training and assessment that can be specifically analyzed. Summary of the invention

[0004] The purpose of the present invention is to provide an artificial intelligence monitoring action system for high-altitude work training and assessment to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: an artificial intelligence monitoring action system for high-altitude work training and assessment, including a visual monitoring module, a danger judgment module, and an early warning module. The visual monitoring module is used to monitor and identify the actions of the assessors using a multi-camera surround view method. The danger judgment module is used to determine the center of gravity position of the current assessor's body, and to determine whether the current state of both hands being away is in a dangerous situation. The early warning module is used to warn the assessor when he is in a dangerous action or posture.

[0006] According to the above technical solution, the visual monitoring module includes a camera, an image stitching module, an object recognition module, a personnel detection module, an action recognition module, and a data storage module. The camera is electrically connected to the image stitching module, and the personnel detection module is electrically connected to the action recognition module. A plurality of the cameras cover various heights of the assessment area and shoot the assessment area. The image stitching module is used to stitch images shot by different cameras into a complete panoramic two-dimensional image. The object recognition module is used to identify objects in contact with the assessor within the monitoring range. The personnel detection module uses a target detection algorithm to detect the assessor target in the image. The action recognition module is used to extract posture features from the tracked assessor action trajectory. The data storage module is used to store and manage the collected video data.

[0007] The danger determination module includes a center of gravity analysis module, a support determination module, a work type distinction module, a flight timing module, and a danger degree calculation module. The center of gravity analysis module is electrically connected to the action recognition module, the support determination module is electrically connected to the center of gravity analysis module, the work type distinction module and the flight timing module are electrically connected to the danger degree calculation module, and the flight timing module is electrically connected to the action recognition module. The center of gravity analysis module is used to analyze the center of gravity position of the assessee's body in real time according to the assessee's body posture. The support determination module is used to determine whether the objects in the assessment area can support the assessee's body and determine the deviation between the body's center of gravity and the support. The work type distinction module is used to set the time allowed for both hands to be in the air according to the current assessee's work type. The flight timing module is used to calculate the time when the assessee's hands are in the air during the operation and make statistics. The danger degree calculation module is used to calculate the danger degree of the current assessee in combination with the degree of deviation between the body's center of gravity and the support and the time when both hands are in the air;

[0008] The early warning module includes a voice reminder module and a safety rope traction control unit. The voice reminder module is used to give a voice reminder when it is determined that the examinee is in a generally dangerous state, so that the examinee can correct the action immediately. The safety rope traction control unit is used to control the safety rope to pull the examinee when the examinee is in a highly dangerous state, correct the behavior and prevent accidents.

[0009] According to the above technical solution, the working method of the monitoring action system is:

[0010] S1. When conducting training assessment work, identify the body of the assessor and the objects stepped on by the assessor in the monitoring image, and stitch the images taken by different cameras into a complete panoramic two-dimensional image, and model the body contour of the assessor in a two-dimensional coordinate system, where the coordinate system plane is the work surface where the assessment is located, the abscissa is the coordinate on the horizontal line, and the ordinate is the coordinate on the vertical line;

[0011] S2. Analyze the body posture characteristics of the examinee, analyze the center of gravity of the examinee in real time, and determine whether the examinee is on an object to be stepped on, whether it is a supporting object, and whether the examinee is allowed to keep both hands in the air for a long time;

[0012] S3, reading the current job type of the assessor, setting the time allowed for both hands to be free according to the current job type of the assessor, and giving a voice reminder when the time allowed for both hands to be free is exceeded;

[0013] S4. When the examinee's body center of gravity leaves the support, calculate the vertical distance between the body center of gravity and the support, adjust the time allowed for one-handed vacancy, and give a voice reminder when the time exceeds the time allowed for one-handed vacancy.

[0014] According to the above technical solution, in S2, the body posture characteristics are analyzed only when the examinee's body movements are in a static state, which is divided into the following two situations:

[0015] S2-1. When the examinee steps on the same object with both feet and the body's center of gravity is on the stepped object, the stepped object is marked as a supporting object, and the maximum allowed time for both hands to be free is ;

[0016] S2-2. When the examinee steps on different objects with both feet and the body center of gravity is located between the two stepped objects at a vertical distance, both stepped objects are marked as supporting objects. At this time, the allowed time for both hands to be free is Dynamic adjustment between

[0017] S2-3. When the examinee's body center of gravity is outside the vertical distance of two stepped objects, and when the body center of gravity is not on the stepped objects, both hands are not allowed to be in the air.

[0018] According to the above technical solution, in S2-2, The specific method for dynamic adjustment is as follows: the plane rectangular coordinate system coordinates of the two supports are and , the distance between the two is , the horizontal coordinate of the body center of gravity is , the vertical deviation between the body center of gravity and the two supports is The greater the distance between the two supports and the higher the vertical deviation, the easier it is for the body to get tired. ,in The conversion factor between the distance between the two supports and the time it takes for the body to enter fatigue is obtained from the experiment, and the time allowed for both hands to be free , It is the conversion coefficient between the deviation of body center of gravity and the time it takes for the body to enter fatigue obtained from the experiment.

[0019] According to the above technical solution, in S3, the job type of the assessor is divided into two-handed installation job, one-handed installation job, and mobile job. When the job type of the assessor is two-handed installation job, the allowed free time of both hands is determined according to the steps in S2; when the job type of the assessor is one-handed installation job, only the minimum time is allowed. The following is done with both hands free, and ; When the assessee’s work type is active work, it is not allowed to have both hands free.

[0020] According to the above technical solution, the specific method for adjusting the time allowed for one-handed vacancy in S4 is: adjusting the horizontal coordinate of the position of the examinee's one-handed grip There are two situations for recording:

[0021] S4-1, when the horizontal coordinate of the body's center of gravity The horizontal coordinate of the position when holding with one hand With support and The middle position The longer the hand is held, the more effort it will take. ,in It is the time coefficient of body fatigue under one-handed pressure grip;

[0022] S4-2, when the horizontal coordinate of the body's center of gravity The horizontal coordinate of the position when holding with one hand With support and The middle position When the hand is in the state of being pulled and held with one hand, the farther the hand is held, the less effort it will take. ,in It is the time coefficient of body fatigue under one-handed tensile grip.

[0023] According to the above technical solution, in S3, when both hands are in the air when both hands are not allowed to be in the air, it is judged that the situation is highly dangerous and the hands are immediately pulled by controlling the safety rope; in S4, when both hands are in the air, the hands are immediately pulled by controlling the safety rope.

[0024] According to the above technical solution, in S3, the type of work of the assessee also includes sitting work. When the type of work of the assessee is sitting work, the action recognition module only detects the action of the upper body, and only uses the upper body to participate in the calculation of the body center of gravity, and marks the object being sat on at this time as a support. If the horizontal coordinate of the body center of gravity does not exceed the left and right horizontal coordinate range of the support, the maximum allowed time for both hands to be free is 100%. If the horizontal coordinate of the body's center of gravity exceeds the left and right horizontal coordinate range of the support, both hands are not allowed to be lifted off the ground.

[0025] According to the above technical solution, when judging whether both hands are vacant, the air timing module first identifies the objects next to the two hands of the examinee, and uses the action recognition module to judge whether the objects are touched and whether the fingers are in a bending motion. If it is judged that there is no contact or that the hands are in contact but not in a bending motion, the hand is determined to be in an air state.

[0026] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention, by analyzing the body posture of the assessee, can determine whether the center of gravity of the human body is on the support and the vertical distance between the center of gravity of the body and the support, and in combination with the type of job currently being assessed, can determine the time allowed for both hands to be free for the current job, calculate the safety index of the duration of releasing both hands, and conduct specific analysis and early warning on the current behavior of releasing both hands, which is more in line with reality. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0028] Figure 1 It is a schematic diagram of the overall module structure of the present invention. DETAILED DESCRIPTION

[0029] 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.

[0030] See also Figure 1The present invention provides a technical solution: an artificial intelligence monitoring action system for high-altitude work training and assessment, including a visual monitoring module, a danger determination module, and an early warning module. The visual monitoring module is used to monitor and identify the actions of the assessor by using a method of multiple cameras looking around. The danger determination module is used to determine the body center of gravity position of the current assessor and determine whether the current state of leaving both hands is in a dangerous situation. The early warning module is used to warn the assessor when he is in a dangerous action and posture;

[0031] The visual monitoring module includes a camera, an image stitching module, an object recognition module, a personnel detection module, an action recognition module, and a data storage module. The camera is electrically connected to the image stitching module, and the personnel detection module is electrically connected to the action recognition module. Multiple cameras cover various heights of the assessment area and shoot the assessment area. The image stitching module is used to stitch images shot by different cameras into a complete panoramic two-dimensional image. The object recognition module is used to identify objects in contact with the assessor within the monitoring range. The personnel detection module uses a target detection algorithm to detect the assessor target in the image. The action recognition module is used to extract posture features from the tracked assessor action trajectory. The data storage module is used to store and manage the collected video data.

[0032] The danger determination module includes a center of gravity analysis module, a support determination module, an operation type distinction module, a flight timing module, and a danger degree calculation module. The center of gravity analysis module is electrically connected to the action recognition module, the support determination module is electrically connected to the center of gravity analysis module, the operation type distinction module and the flight timing module are electrically connected to the danger degree calculation module, and the flight timing module is electrically connected to the action recognition module. The center of gravity analysis module is used to analyze the center of gravity position of the assessee's body in real time according to the assessee's body posture. The support determination module is used to determine whether the objects in the assessment area can support the assessee's body and determine the deviation between the body's center of gravity and the support. The operation type distinction module is used to set the time allowed for both hands to be vacated according to the current assessee's operation type. The flight timing module is used to calculate the time when the assessee's hands are vacated during the operation and make statistics. The danger degree calculation module is used to calculate the danger degree of the current assessee in combination with the degree of deviation between the body's center of gravity and the support and the time when both hands are vacated;

[0033] The early warning module includes a voice reminder module and a safety rope traction control unit. The voice reminder module is used to give a voice reminder when it is determined that the examinee is in a general dangerous state, so that the examinee can correct the action immediately. The safety rope traction control unit is used to control the safety rope to pull the examinee when the examinee is in a highly dangerous state, correct the behavior and prevent accidents.

[0034] The working method of the monitoring action system is:

[0035] S1. When conducting training assessment work, identify the body of the assessor and the objects stepped on by the assessor in the monitoring image, and stitch the images taken by different cameras into a complete panoramic two-dimensional image, and model the body contour of the assessor in a two-dimensional coordinate system, where the coordinate system plane is the work surface where the assessment is located, the abscissa is the coordinate on the horizontal line, and the ordinate is the coordinate on the vertical line;

[0036] S2. Analyze the body posture characteristics of the examinee, analyze the center of gravity of the examinee in real time, and determine whether the examinee is on an object to be stepped on, whether it is a supporting object, and whether the examinee is allowed to keep both hands in the air for a long time;

[0037] S3, reading the current job type of the assessor, setting the time allowed for both hands to be free according to the current job type of the assessor, and giving a voice reminder when the time allowed for both hands to be free is exceeded;

[0038] S4. When the examinee's body center of gravity leaves the support, the vertical distance between the body center of gravity and the support is calculated, and the time allowed for one-handed vacancy is adjusted. When the time allowed for one-handed vacancy is exceeded, a voice reminder is given;

[0039] In S2, the body posture characteristics are analyzed only when the examinee's body movements are in a static state, which can be divided into the following two situations:

[0040] S2-1. When the examinee steps on the same object with both feet and the body's center of gravity is on the stepped object, the stepped object is marked as a supporting object, and the maximum allowed time for both hands to be free is ;

[0041] S2-2. When the examinee steps on different objects with both feet and the body center of gravity is located between the two stepped objects at a vertical distance, both stepped objects are marked as supporting objects. At this time, the allowed time for both hands to be free is Dynamic adjustment between

[0042] S2-3: When the examinee's body center of gravity is outside the vertical distance of two stepped objects, or when the body center of gravity is not on the stepped objects, both hands are not allowed to be in the air;

[0043] In S2-2, The specific method for dynamic adjustment is as follows: the plane rectangular coordinate system coordinates of the two supports are and , the distance between the two is , the horizontal coordinate of the body center of gravity is , the vertical deviation between the body center of gravity and the two supports is The greater the distance between the two supports and the higher the vertical deviation, the easier it is for the body to get tired. ,in The conversion factor between the distance between the two supports and the time it takes for the body to enter fatigue is obtained from the experiment, and the time allowed for both hands to be free , It is the conversion coefficient between the deviation of body center of gravity and the time it takes for the body to enter fatigue obtained from the experiment;

[0044] In S3, the job type of the assessee is divided into two-handed installation job, one-handed installation job, and mobile job. When the job type of the assessee is two-handed installation job, the allowed free time of both hands is determined according to the steps in S2; when the job type of the assessee is one-handed installation job, only the minimum time is allowed. The following is done with both hands free, and ; When the assessment personnel's work type is active work, it is not allowed to have both hands free;

[0045] In S4, the specific method for adjusting the time allowed for one-handed vacancy is as follows: There are two situations for recording:

[0046] S4-1, when the horizontal coordinate of the body's center of gravity The horizontal coordinate of the position when holding with one hand With support and The middle position The longer the hand is held, the more effort it will take. ,in It is the time coefficient of body fatigue under one-handed pressure grip;

[0047] S4-2, when the horizontal coordinate of the body's center of gravity The horizontal coordinate of the position when holding with one hand With support and The middle position When the hand is in the state of being pulled and held with one hand, the farther the hand is held, the less effort it will take. ,in is the time coefficient of body fatigue under one-handed tensile grip;

[0048] In S3, when both hands are in the air when both hands are not allowed to be in the air, it is determined that the person is in a highly dangerous state and is immediately pulled by controlling the safety rope; in S4, when both hands are in the air, they are immediately pulled by controlling the safety rope.

[0049] In S3, the type of work of the assessee also includes sitting work. When the type of work of the assessee is sitting work, the action recognition module only detects the action of the upper body, and only uses the upper body to calculate the center of gravity of the body, and marks the object being sat on as a support. If the horizontal coordinate of the center of gravity of the body does not exceed the left and right horizontal coordinate range of the support, the maximum allowed time for both hands to be free is 20 seconds. If the horizontal coordinate of the body's center of gravity exceeds the left and right horizontal coordinate range of the support, both hands are not allowed to be lifted off the ground.

[0050] When judging whether both hands are free, the air timing module first identifies the objects next to the examinee's two hands, and uses the motion recognition module to determine whether the objects are touched and whether the fingers are in a bent motion. If it is judged that there is no contact or that the hands are in contact but not in a bent motion, the hand is determined to be in a free state.

[0051] A loose safety rope will cause a longer fall distance and a greater impact force. A tightened safety rope can stop a fall earlier, reduce the impact force, and reduce the harm to the human body. Therefore, it is safer to pull the safety rope immediately when a high risk is determined.

[0052] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0053] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Artificial intelligence monitoring action system for high-altitude operation training and assessment, characterized by: It includes a visual monitoring module, a danger determination module, and an early warning module. The visual monitoring module is used to monitor and identify the movements of the examinee by using a method of multiple cameras looking around. The danger determination module is used to determine the center of gravity of the current examinee and whether the current state of leaving both hands is in a dangerous situation. The early warning module is used to warn the examinee when he is in a dangerous movement or posture. The visual monitoring module includes a camera, an image stitching module, an object recognition module, a personnel detection module, an action recognition module, and a data storage module. The camera is electrically connected to the image stitching module, and the personnel detection module is electrically connected to the action recognition module. A plurality of the cameras cover various heights of the assessment area and shoot the assessment area. The image stitching module is used to stitch images shot by different cameras into a complete panoramic two-dimensional image. The object recognition module is used to identify objects in contact with the assessor within the monitoring range. The personnel detection module uses a target detection algorithm to detect the assessor target in the image. The action recognition module is used to extract posture features from the tracked assessor action trajectory. The data storage module is used to store and manage the collected video data. The danger determination module includes a center of gravity analysis module, a support determination module, a work type distinction module, a flight timing module, and a danger degree calculation module. The center of gravity analysis module is electrically connected to the action recognition module, the support determination module is electrically connected to the center of gravity analysis module, the work type distinction module and the flight timing module are electrically connected to the danger degree calculation module, and the flight timing module is electrically connected to the action recognition module. The center of gravity analysis module is used to analyze the center of gravity position of the assessee's body in real time according to the assessee's body posture. The support determination module is used to determine whether the objects in the assessment area can support the assessee's body and determine the deviation between the body's center of gravity and the support. The work type distinction module is used to set the time allowed for both hands to be in the air according to the current assessee's work type. The flight timing module is used to calculate the time when the assessee's hands are in the air during the operation and make statistics. The danger degree calculation module is used to calculate the danger degree of the current assessee in combination with the degree of deviation between the body's center of gravity and the support and the time when both hands are in the air; The early warning module includes a voice reminder module and a safety rope traction control unit. The voice reminder module is used to give a voice reminder when it is determined that the examinee is in a generally dangerous state, so that the examinee can correct the action immediately. The safety rope traction control unit is used to control the safety rope to pull the examinee when the examinee is in a highly dangerous state, correct the behavior and prevent accidents.

2. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 1 is characterized by: The working method of the monitoring action system is: S1. When conducting training assessment work, identify the body of the assessor and the objects stepped on by the assessor in the monitoring image, and stitch the images taken by different cameras into a complete panoramic two-dimensional image, and model the body contour of the assessor in a two-dimensional coordinate system, where the coordinate system plane is the work surface where the assessment is located, the abscissa is the coordinate on the horizontal line, and the ordinate is the coordinate on the vertical line; S2. Analyze the body posture characteristics of the examinee, analyze the center of gravity of the examinee in real time, and determine whether the examinee is on an object to be stepped on, whether it is a supporting object, and whether the examinee is allowed to keep both hands in the air for a long time; S3, reading the current job type of the assessor, setting the time allowed for both hands to be free according to the current job type of the assessor, and giving a voice reminder when the time allowed for both hands to be free is exceeded; S4. When the examinee's body center of gravity leaves the support, calculate the vertical distance between the body center of gravity and the support, adjust the time allowed for one-handed vacancy, and give a voice reminder when the time exceeds the time allowed for one-handed vacancy.

3. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 2 is characterized by: In S2, the body posture characteristics are analyzed only when the examinee's body movements are in a static state, which is divided into the following two situations: S2-1. When the examinee steps on the same object with both feet and the center of gravity of the body is on the stepped object, the stepped object is marked as a supporting object. The maximum allowed time for both hands to be free is t max ; S2-2. When the examinee steps on different objects with both feet and the body center of gravity is located between the two stepped objects at a vertical distance, both stepped objects are marked as supporting objects. At this time, the allowed time for both hands to be free is [t a ~t max ] to make dynamic adjustments between them; S2-3. When the examinee's body center of gravity is outside the vertical distance of two stepped objects, and when the body center of gravity is not on the stepped objects, both hands are not allowed to be in the air.

4. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 3 is characterized by: In S2-2, at [t a ~t max The specific method of dynamic adjustment between ] is as follows: the plane rectangular coordinate system coordinates of the two supports are (x1, y1) and (x2, y2), and the distance between them is The horizontal coordinate of the body's center of gravity is x3, and the vertical deviation between the body's center of gravity and the two supports is The greater the distance between the two supports and the higher the vertical deviation, the easier it is for the body to get tired. Among them, α is the conversion coefficient between the distance between the two supports and the time it takes for the body to enter fatigue, and the time allowed for both hands to be free is β is the conversion coefficient between the deviation of the body's center of gravity and the time it takes for the body to enter fatigue, obtained from the experiment.

5. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 4 is characterized by: In S3, the job type of the assessee is divided into two-handed installation job, one-handed installation job, and mobile job. When the job type of the assessee is two-handed installation job, the allowed free time of both hands is determined according to the steps in S2; when the job type of the assessee is one-handed installation job, only the minimum time t is allowed. min The following is done with both hands free, and t min <t a ; When the assessee’s work type is active work, it is not allowed to have both hands free.

6. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 5 is characterized by: In S4, the specific method for adjusting the time allowed for one-handed vacancy is: recording the horizontal coordinate x4 of the position of the examinee's one-handed grip, which is divided into the following two situations: S4-1. When the horizontal coordinate x3 of the body's center of gravity is in the middle of the horizontal coordinate x4 of the position of holding with one hand and the support x1 and x2 The longer the hand is held, the more effort it will take. Where μ is the time coefficient of body fatigue under one-handed pressure grip; S4-2, when the horizontal coordinate x3 of the body's center of gravity is in the middle of the horizontal coordinate x4 of the position of holding with one hand and the support x1 and x2 When the hand is in the state of being pulled and held with one hand, the farther the hand is held, the less effort it will take. Where δ is the time coefficient of body fatigue under one-handed tensile grip.

7. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 6 is characterized by: In S3, when both hands are in the air when both hands are not allowed to be in the air, it is determined that the situation is highly dangerous and the hands are immediately pulled by controlling the safety rope; in S4, when both hands are in the air, the hands are immediately pulled by controlling the safety rope.

8. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 7 is characterized by: In S3, if the type of work of the assessee also includes sitting work, when the type of work of the assessee is sitting work, the action recognition module only detects the action of the upper body, and only uses the upper body to participate in the calculation of the body center of gravity, and marks the object being sat on as a support. If the horizontal coordinate of the body center of gravity does not exceed the left and right horizontal coordinate range of the support, the maximum allowed time t for both hands to be free is max If the horizontal coordinate of the body's center of gravity exceeds the left and right horizontal coordinate range of the support, both hands are not allowed to be lifted off the ground.

9. The artificial intelligence monitoring action system for high-altitude work training and assessment according to claim 8 is characterized by: When determining whether both hands are vacant, the air timing module first identifies the objects next to the two hands of the examinee, and uses the action recognition module to determine whether the fingers are in contact with the objects and whether they are in a bending motion. If it is determined that there is no contact or that the fingers are in contact but not in a bending motion, the hand is determined to be in the vacant state.

Images