A method and system for job safety portrait management based on big model

Abstract

The present invention discloses a method and system for job safety portrait management based on a large model, and relates to the technical field of job safety portrait management. The present invention includes: step one: selectively scheduling a number of video acquisition devices installed in the working area; step two: constructing a job safety assessment model for verification staff in multiple dimensions; step three: predicting the real-time risk factor of the experiment; step four: drawing a job safety portrait of the verification staff; step five: managing the working status of the verification staff. The present invention draws the real-time job portrait of the verification staff through the job safety assessment model constructed in multiple dimensions and the constructed risk assessment model. Compared with the prior art, the job safety portrait drawn by the present invention is real-time and has a higher matching degree. The real-time working status of the verification staff can be effectively managed through the real-time job safety portrait, thereby improving the use effect of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of job safety portrait management, and specifically to a job safety portrait management method and system based on a large model. Background Art

[0002] Operational safety profiling refers to the representation of hazardous factors and safety risks faced by workers during their operations in the form of images, which facilitates intuitive analysis of the workers' work safety situation.

[0003] In the prior art, a worker's operation safety portrait is drawn based on the probability of various illegal operations that occur during the worker's operation. Based on the drawing result, the worker's operation safety situation is analyzed. Through the existing method, it is impossible to predict the potential dangers of the worker in the operation process. At the same time, it is also impossible to analyze the worker's operation safety situation based on the worker's working status during the operation.

[0004] Therefore, people urgently need a large-model-based work safety portrait management method and system to solve the above technical problems. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for managing job safety portraits based on a large model to solve the problems raised in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for managing job safety portraits based on a large model, the method comprising:

[0007] Step 1: Use the camera installed at the entrance of the work area to collect the portrait of the staff, match the collected portrait with the portrait stored in the database, verify the identity of the staff, and based on the verification result, choose whether to retrieve the work content of the verified staff. Combined with the experiments that can be performed on each operating table in the work area, selectively dispatch several video acquisition devices installed in the work area;

[0008] Step 2: When executing each experimental step, the reaction video of the experiment is extracted through the video acquisition device. According to the video frame set corresponding to each reaction video and the odor sensor installed on each operating table in the working area, the real-time reaction feature set of the experiment is obtained. Based on the obtained results, a work safety assessment model for verification workers is constructed in multiple dimensions;

[0009] Step 3: Use the video acquisition device installed in the work area to obtain the real-time status characteristics of the verification staff in each reaction time period, build a risk assessment model based on the acquired information, and predict the real-time risk factor of the experiment;

[0010] Step 4: Draw a safety portrait of the verification staff based on the knowledge graph;

[0011] Step 5: Predict the real-time operation safety assessment index of the verification staff, and manage the work status of the verification staff based on the prediction results.

[0012] Furthermore, the step 1 includes: when the collected personnel portrait is the same as a random personnel portrait stored in the database, the identity verification of the staff member is successful, and the work content of the verification staff member is retrieved at this time;

[0013] According to the retrieved work content, the experiments that the verification staff need to perform are determined. Combined with the experiments that can be performed on each operating table in the work area, the experimental operating table of the verification staff is determined. It is judged whether the line connecting the center of the experimental operating table surface and the center of the lens of each video acquisition device is the central axis of the video acquisition device. If so, the corresponding video acquisition device is scheduled. If not, the corresponding video acquisition device is not scheduled.

[0014] Furthermore, the step 2 includes:

[0015] S201: when executing each experimental step, extracting the reaction video of the experiment through the video acquisition device, determining the start time of each reaction video, and in each reaction video, using the video frame acquired at the start time point as the key frame, and using the video frame set acquired except the start time point as the comparison frame set;

[0016] S202: According to the execution order of each step in the experiment, each reaction video is numbered, and the numbering result is: p = 1, 2, ..., q; q represents the total number of reaction videos, and the reaction time f of the pth step in the experiment is calculated according to the target object dose taken by the verification staff. p Determine the time T versus the reaction time f p Divide and number each dividing point. The numbering result is: h=1,2,…, f p / T;f p / T indicates the total number of partition points;

[0017] When the verification staff performs step p, the real-time comparison frame collected in the reaction video numbered p is compared with the key frame to verify the real-time reaction feature set M of the experiment. p To obtain, , where H 1p =t 1p +T; indicates the time value corresponding to the start time point in the reaction video numbered p. Indicates that the target object is at t 1p + The jth type of physical change occurs at time h*T, Indicates that the target object is at t 1p +1*T time when the jth type of physical change occurs, Indicates that when the target object is at t 1p + When the jth type of physical change occurs at time h*T, the degree of physical change of the target object, Indicates that when the target object is at t 1p +1*T, the degree of physical change of the target object when the jth type of physical change occurs at time T, where T represents the time interval value, and H 1p Indicates the time value corresponding to the first division point in the reaction video numbered p, H hp Indicates the time value corresponding to the hth division point in the reaction video numbered p;

[0018] S203: According to the type of physical change that occurs to the target object during each reaction time, the reaction feature set M is p After classification, we get M p ={M p1 ,M p2 ,M p3 ,M p4}, where M p1 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 1, M p2 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 2, M p3 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 3, M p4 It means that when the physical change type of the target object is 4, when the verification staff performs the p-th step, the real-time reaction feature set of the experiment takes each type of physical change as a dimension, and based on the classification processing results, a work safety assessment model for the verification staff is constructed in multiple dimensions.

[0019] By comparing the comparison frame set with the key frame, the real-time response feature set of the experiment is obtained, and the operation safety assessment model for the verification staff is constructed in multiple dimensions, which weakens the correlation between the dimensions and is conducive to improving the accuracy of the operation safety assessment.

[0020] Furthermore, the specific method of obtaining the real-time response feature set of the verification staff when executing step p in step S202 is:

[0021] Will be in t 1p + The comparison frame collected at time h*T is compared with the frame collected at time t 1p The key frames collected at time t are compared to 1p + The physical changes that occur at time h*T are taken as the experimental 1p + The reaction features at time h*T are used as the set of real-time reaction features for the experiment. 1p + h*Response characteristics at time T = , where j = 1, 2, 3, 4, representing the number corresponding to each physical change type, H hp =t 1p + h*T;

[0022] The types of physical changes include Type 1: the shape and size of the target object changes, Type 2: new substances adhere to the surface of the target object, Type 3: the color of the target object changes, and Type 4: the target object produces gas.

[0023] Furthermore, the specific method of constructing a multi-dimensional operation safety assessment model for verification staff in S203 is:

[0024] The jth type of physical change is taken as the jth dimension;

[0025] In [t 1p +T,t 1p +f p ] time period, based on the acquired standard comparison frame, when the target object is in H hp The standard gas concentration generated on the target object when the jth type of physical change occurs at time Acquire, and build the j-th dimension operation safety assessment model based on the acquired information. The specific formula is:

[0026] ;

[0027] in, Indicates that H (h+1)p At this moment, the safety assessment coefficient of the verification staff in the jth dimension is predicted. Indicates that when the target object is in H (h+1)p When the jth type of physical change occurs at time instant, the degree of physical change of the target object, Indicates that when the target object is in H (h+1)p The standard gas concentration produced by the target object when the jth type of physical change occurs at time instant;

[0028] When the experiment does not proceed smoothly, the above situation can also be identified through the operation safety assessment model of each dimension, which is conducive to improving the work efficiency of the verification staff;

[0029] The operation safety assessment model of each dimension can also be used to evaluate the unperformed status.

[0030] In [t 1p +T, t 1p +f p ] time period, when the target object is in H hp The standard gas concentration generated on the target object when the fourth type of physical change occurs at any time Acquire and build the fourth dimension of the operation safety assessment model based on the acquired information. The specific formula is:

[0031] ;

[0032] in, Indicates that H (h+1)p At this moment, the safety assessment coefficient of the verification staff in the fourth dimension is predicted. Indicates that when the target object is in H (h+1)p The degree of physical change of the target object when the fourth type of physical change occurs at the moment, Indicates that when the target object is in H (h+1)p The standard gas concentration produced by the target object when the fourth type of physical change occurs at any time.

[0033] Furthermore, the step three includes:

[0034] S301: Verification staff [t 1p +T,t 1p +f p ] to obtain the real-time status characteristics within the time period, the status characteristics including the fatigue value D of the verification staff and the distance value L between the verification staff and the experimental operation table;

[0035] S302: Construct a risk assessment model to predict the real-time risk factor of the experiment.

[0036] By analyzing the real-time status characteristics of the verification staff, it is possible to analyze the reaction status of the verification staff at each moment. When evaluating the operation safety, the impact of the verification staff's working status on the evaluation results is taken into consideration to ensure that the drawn operation safety portrait is more in line with the actual situation.

[0037] Furthermore, the specific formula of the risk assessment model constructed in S302 is:

[0038] ;

[0039] in, Indicates that the verification staff is in H (h+1)p Fatigue value at the moment, Indicates that the verification staff is in H (h+1)p The distance from the experimental operation table at the moment, Indicates that the experiment was conducted in H (h+1)p The risk coefficient at a given moment, exp represents an exponential function with the natural constant e as the base and e=2.72.

[0040] Furthermore, the step four includes: using the personnel number of the verification staff as the entity of the job safety portrait, using the job safety assessment model of the verification staff in each dimension and the risk assessment model of the experiment as entity attributes, using the safety assessment coefficient of each job safety assessment model and the assessment coefficient of the risk assessment model as the attribute values ​​of the entity attributes, and obtaining the job safety portrait of the verification staff according to the entity-attribute-attribute value drawing method, and drawing the job safety portrait of the verified staff through the knowledge graph drawing method to ensure that the job safety portrait is more intuitive.

[0041] Furthermore, the step five includes: according to E μ =a 1 ×U 1μ +a 2 ×U 2μ +a 3 ×U 3μ +a 4 ×U 4μ +a 5 ×F μ For verification staff in H (h+1)p The operation safety assessment index at the moment is predicted, where a 1 +a 2 +a 3 +a 4 +a 5 =1 and a 1 、a 2 、a 3 、a 4 、a 5 Both represent proportionality coefficient, μ=H (h+1)p， U 1μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient on the first dimension, U 2μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the second dimension, U 3μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the third dimension, U4μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the fourth dimension, F μ Indicates that the experiment was conducted in H (h+1)p The risk factor at the moment;

[0042] When 0.5≤E μ When ≤1, the verification staff's work safety portrait is transmitted to the experimental terminal. When the experimental terminal receives the verification staff's work safety portrait, the verification staff's work is terminated;

[0043] When 0≤E μ When the value is less than 0.5, the operation safety portrait of the verification staff will be transmitted to the experimental terminal. When the experimental terminal receives the operation safety portrait of the verification staff, it will send an early warning message to the intelligent terminal of the verification staff.

[0044] By conducting a comprehensive analysis of the drawn work safety portrait, the work status of the verification staff can be managed in advance before the danger comes. The comprehensive analysis results are timely and can reduce the occurrence of safety accidents to a certain extent.

[0045] A large-model-based operation safety portrait management system, the system comprising a video acquisition device scheduling module, an operation safety assessment model building module, a risk assessment model building module, an operation safety portrait drawing module and an operation safety management module;

[0046] The video acquisition device scheduling module is used to selectively schedule a number of video acquisition devices installed in the working area;

[0047] The job safety assessment model building module is used to build a job safety assessment model for verification workers in multiple dimensions;

[0048] The risk assessment model construction module is used to predict the real-time risk factor of the experiment through the constructed risk assessment model;

[0049] The operation safety portrait drawing module is used to draw the operation safety portrait of the verification staff;

[0050] The operation safety management module is used to perform safety management on the working status of the verification staff.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The present invention analyzes the real-time reaction state of the experiment by comparing the collected comparison frame set with the key frame. Based on the analysis results, the real-time operation safety of the experiment is evaluated in multiple dimensions, which is conducive to improving the evaluation accuracy and the evaluation results are more in line with the actual situation.

[0053] 2. The present invention analyzes the real-time reaction of the verification staff when an accident occurs in the experiment by verifying the real-time status characteristics of the verification staff during the operation, thereby improving the operation safety analysis effect of the verification staff.

[0054] 3. The present invention draws a real-time work portrait of the verification staff through a work safety assessment model constructed in multiple dimensions and a constructed risk assessment model. Compared with the prior art, the work safety portrait drawn by the present invention is real-time and has a higher matching degree, which is conducive to the effective management of the real-time working status of the verification staff through the real-time work safety portrait, thereby improving the use effect of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The present invention is a schematic diagram of the workflow of a large-model-based operation safety portrait management method and system.

[0056] Figure 2 A schematic diagram of a job safety portrait of a job safety portrait management method and system based on a large model of the present invention. DETAILED DESCRIPTION

[0057] Next, the technical solution of the present invention will be clearly and completely explained through specific case analysis to ensure that those skilled in the art can implement the present invention according to this embodiment.

[0058] like Figure 1 and Figure 2 As shown, the present invention provides the following technical solution, a method for managing job safety portraits based on a large model, the method comprising:

[0059] Step 1: Use the camera installed at the entrance of the work area to collect the portrait of the staff, match the collected portrait with the portrait stored in the database, verify the identity of the staff, and based on the verification result, choose whether to retrieve the work content of the verified staff. Combined with the experiments that can be performed on each operating table in the work area, selectively dispatch several video acquisition devices installed in the work area;

[0060] Step 1 includes: when the collected person portrait is the same as a random person portrait stored in the database, the identity verification of the staff member is successful, and the work content of the verification staff member is retrieved; when the collected person portrait is different from any person portrait stored in the database, the identity verification of the staff member is unsuccessful, and the alarm terminal connected to the camera sends an alarm signal;

[0061] According to the retrieved work content, the experiments that the verification staff need to conduct are determined. In combination with the experiments that can be performed on each operating table in the work area, the experimental operating table of the verification staff is determined, and it is judged whether the line connecting the center of the experimental operating table surface and the center of the lens of each video acquisition device is the central axis of the video acquisition device. The central axis of the video acquisition device refers to an invisible line connecting the lens of the video acquisition device and the photographed object. If so, the corresponding video acquisition device is controlled to be in the on state to realize the scheduling of the video acquisition device. If not, the corresponding video acquisition device is controlled to be in the off state.

[0062] The corresponding video acquisition device is scheduled, if not, the corresponding video acquisition device is not scheduled;

[0063] Step 2: When executing each experimental step, the reaction video of the experiment is extracted through the video acquisition device. According to the video frame set corresponding to each reaction video and the odor sensor installed on each operating table in the working area, the real-time reaction feature set of the experiment is obtained. Based on the obtained results, a work safety assessment model for verification workers is constructed in multiple dimensions;

[0064] Step 2 includes:

[0065] S201: When executing each experimental step, a reaction video of the experiment is extracted through a video acquisition device. The reaction video refers to a video recording a physical change of a target object, and the target object refers to a hazardous chemical. The start time of each reaction video is determined. In each reaction video, a video frame collected at the start time point is used as a key frame, and a video frame set collected except the start time point is used as a comparison frame set.

[0066] S202: According to the execution order of each step in the experiment, each reaction video is numbered, and the numbering result is: p = 1, 2, ..., q; q represents the total number of reaction videos, and the reaction time f of the pth step in the experiment is calculated according to the target object dose taken by the verification staff. p Determine the time T versus the reaction time f p Divide and number each dividing point. The numbering result is: h=1,2,…, f p / T;f p / T indicates the total number of partition points;

[0067] When the verification staff performs step p, the real-time comparison frame collected in the reaction video numbered p is compared with the key frame to verify the real-time reaction feature set M of the experiment. p To obtain, , where H 1p =t 1p +T,t 1pIndicates the time value corresponding to the start time point in the reaction video numbered p. Indicates that the target object is at t 1p + The jth type of physical change occurs at time h*T, Indicates that the target object is at t 1p +1*T time when the jth type of physical change occurs, Indicates that when the target object is at t 1p + When the jth type of physical change occurs at time h*T, the degree of physical change of the target object, Indicates that when the target object is at t 1p +1*T, the degree of physical change of the target object when the jth type of physical change occurs at time T, where T represents the time interval value, and H 1p Indicates the time value corresponding to the first division point in the reaction video numbered p, H hp Indicates the time value corresponding to the hth division point in the reaction video numbered p. The specific method is:

[0068] Will be in t 1p + The comparison frame collected at time h*T is compared with the frame collected at time t 1p The key frames collected at time t are compared to 1p + The physical changes that occur at time h*T are taken as the experimental 1p + The reaction features at time h*T are used as the set of real-time reaction features for the experiment. 1p + h*Response characteristics at time T = , where j = 1, 2, 3, 4, representing the number corresponding to each physical change type, H hp =t 1p + h*T;

[0069] The types of physical changes include type 1: the shape and size of the target object changes, type 2: new substances are attached to the surface of the target object, type 3: the color of the target object changes, and type 4: the target object produces gas;

[0070] In the reaction video numbered p, when j=1, the target object is in H hp The degree of physical change occurring at the moment = the target object at H hp The volume / target object in the comparison frame acquired at time t 1p The volume in the keyframe obtained at the moment;

[0071] When j=2, the target object is in H hp The degree of physical change occurring at time H hp The volume of the new substance attached to the surface of the target object in the comparison frame obtained at time t / the volume of the target object at time t 1p The volume in the keyframe obtained at the moment;

[0072] When j=3, the target object is in H hp The degree of physical change occurring at the moment = |The target object is at H hp The corresponding grayscale value in the contrast frame obtained at time t 1p The corresponding grayscale value in the key frame obtained at time | / the target object at t 1p The corresponding grayscale value in the key frame obtained at each moment;

[0073] When j=4, the odor sensor installed on each operating table in the working area is used. The gas sensor is a converter that converts the volume fraction of a certain gas into a corresponding electrical signal. The concentration of the above-mentioned gas is obtained through the proportional relationship between the change in the electrical signal and the gas concentration. The gas concentration generated by the target object is obtained. The target object is in H hp The degree of physical change occurring at the moment = the target object at H hp The concentration of the gas produced at any moment;

[0074] S203: According to the type of physical change that occurs to the target object during each reaction time, the reaction feature set M is p Perform classification processing and obtain M p ={M p1 ,M p2 ,M p3 ,M p4}, where M p1 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 1, M p2 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 2, M p3 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 3, M p4 It means that when the physical change type of the target object is 4, when the verification staff performs the pth step, the real-time reaction feature set of the experiment takes each type of physical change as a dimension, and based on the classification processing results, the verification staff's operation safety assessment model is constructed in multiple dimensions. The specific method is as follows:

[0075] The jth type of physical change is taken as the jth dimension;

[0076] In [t 1p +T,t 1p +f p ] time period, based on the acquired standard comparison frame, when the target object is in H hp The standard gas concentration generated on the target object when the jth type of physical change occurs at time Acquire, and build the j-th dimension operation safety assessment model based on the acquired information. The specific formula is:

[0077] ;

[0078] in, Indicates that H (h+1)p At this moment, the safety assessment coefficient of the verification staff in the jth dimension is predicted. Indicates that when the target object is in H (h+1)p When the jth type of physical change occurs at time instant, the degree of physical change of the target object, Indicates that when the target object is in H (h+1)p When the jth type of physical change occurs at time, the standard gas concentration generated by the target object is extracted according to the target object dose taken by the verification staff. In the standard reaction video, the comparison frame collected except the starting time point is called the standard comparison frame, and the video frame collected at the starting time point is used as the standard key frame. When j=1, = Target object at H (h+1)p The volume / target object in the standard comparison frame acquired at time t 1p The volume in the standard keyframe obtained at the moment, when j=2, = in H (h+1)p The volume of the new substance attached to the surface of the target object in the comparison frame obtained at time t / the volume of the target object at time t 1p The volume of the key frame obtained at the moment, when j=3, = Target object at H (h+1)p The corresponding grayscale value in the standard comparison frame obtained at time t / the target object at time t 1p The corresponding grayscale value in the key frame obtained at each moment;

[0079] In [t 1p +T, t 1p +f p ] time period, when the target object is in H hp The standard gas concentration generated on the target object when the fourth type of physical change occurs at any time Acquire and build the fourth dimension of the operation safety assessment model based on the acquired information. The specific formula is:

[0080] ;

[0081] in, Indicates that H (h+1)p At this moment, the safety assessment coefficient of the verification staff in the fourth dimension is predicted. Indicates that when the target object is in H (h+1)p The degree of physical change of the target object when the fourth type of physical change occurs at the moment, Indicates that when the target object is in H (h+1)p The standard gas concentration produced by the target object when the fourth type of physical change occurs at any time;

[0082] Step 3: Use the video acquisition device installed in the work area to obtain the real-time status characteristics of the verification staff in each reaction time period, build a risk assessment model based on the acquired information, and predict the real-time risk factor of the experiment;

[0083] Step three includes:

[0084] S301: Verification staff [t 1p +T,t 1p +f p ] time period to obtain the real-time state characteristics, the state characteristics include the verification staff fatigue value D and the verification staff from the experimental operation table distance value L, fatigue value is used to reflect the verification staff fatigue degree, the fatigue value is determined by: through the video acquisition device installed in the working area, the verification staff in the fatigue state of the facial expression state characteristics to determine according to (standard facial expression state characteristic value - facial expression state characteristic value) / standard facial expression state characteristic value, the verification staff real-time fatigue is evaluated, the standard facial expression state characteristic value refers to the verification staff in the awake state of the corresponding facial expression state characteristic value, the distance value is calculated by the experimental operation table and the verification work The size of the personnel in the video frame is used to calculate the distance value of the video acquisition device for scheduling the distance between the experimental operation table and the verification staff. Based on the triangle formed by the three points of the experimental operation table center, the verification staff center and the center of the video acquisition device lens, the real-time distance between the verification staff and the experimental operation table is calculated based on the cosine theorem. The method for obtaining the verification staff center is as follows: the intersection position of the plane where the experimental operation table is located and the verification staff is obtained, and the distance value between the center of the experimental operation table and each intersection position is calculated using the distance calculation formula between two points. A plane coordinate system is constructed with the center of the experimental operation table as the coordinate origin, and the intersection position corresponding to the calculated minimum value is used as the verification staff center;

[0085] S302: Construct a risk assessment model to predict the real-time risk factor of the experiment. The specific formula of the risk assessment model is:

[0086] ;

[0087] in, Indicates that the verification staff is in H (h+1)p Fatigue value at the moment, Indicates that the verification staff is in H (h+1)p The distance from the experimental operation table at the moment, Indicates that the experiment was conducted in H (h+1)p The risk coefficient at the moment, exp represents an exponential function with the natural constant e as the base and e=2.72;

[0088] Step 4: Draw a safety portrait of the verification staff based on the knowledge graph;

[0089] Step 4 includes: using the personnel number of the verification staff as the entity of the operation safety portrait, using the operation safety assessment model of the verification staff in each dimension and the risk assessment model of the experiment as the entity attributes, using the safety assessment coefficient of each operation safety assessment model and the assessment coefficient of the risk assessment model as the attribute value of the entity attribute, and obtaining the operation safety portrait of the verification staff according to the entity-attribute-attribute value drawing method;

[0090] Step 5: Predict the real-time operation safety assessment index of the verification staff, and manage the work status of the verification staff based on the prediction results;

[0091] Step five includes: μ =a 1 ×U 1μ +a 2 ×U 2μ +a 3 ×U 3μ +a 4 ×U 4μ +a 5 ×F μ For verification staff in H (h+1)p The operation safety assessment index at the moment is predicted, where a 1 +a 2 +a 3 +a 4 +a 5 =1 and a 1 、a 2 、a 3 、a 4 、a 5 Both represent proportionality coefficient, μ=H (h+1)p， U 1μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient on the first dimension, U 2μ Indicates that H(h+1)p At this moment, verify the worker's work safety assessment coefficient in the second dimension, U 3μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the third dimension, U 4μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the fourth dimension, F μ Indicates that the experiment was conducted in H (h+1)p The risk factor at the moment;

[0092] When 0.5≤E μ When ≤1, the verification staff's work safety portrait is transmitted to the experimental terminal. When the experimental terminal receives the verification staff's work safety portrait, the verification staff's work is terminated;

[0093] When 0≤E μ When the value is less than 0.5, the operation safety portrait of the verification staff will be transmitted to the experimental terminal. When the experimental terminal receives the operation safety portrait of the verification staff, it will send an early warning message to the intelligent terminal of the verification staff.

[0094] A large-model-based operation safety portrait management system, the system includes a video acquisition device scheduling module, an operation safety assessment model building module, a risk assessment model building module, an operation safety portrait drawing module and an operation safety management module;

[0095] The video acquisition device scheduling module is used to selectively schedule a number of video acquisition devices installed in the working area;

[0096] The job safety assessment model building module is used to build a job safety assessment model for verification workers in multiple dimensions;

[0097] The risk assessment model building module is used to predict the real-time risk factor of the experiment through the constructed risk assessment model;

[0098] The operation safety portrait drawing module is used to draw the operation safety portrait of the verification staff;

[0099] The job safety management module is used to perform safety management on the working status of verification staff.

[0100] Example 1: Assume that the verification staff is at H (h+1)p The distance from the experimental operation table at the moment , unit: meter, verification staff in H (h+1)p Fatigue value at the moment , then the experiment is in H (h+1)p The risk factor at this moment is:

[0101] ;

[0102] The experiment is in H (h+1)p The risk factor at this moment is 0.8.

[0103] 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.

[0104] 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. A method for managing job safety portraits based on a large model, characterized in that: The method comprises: Step 1: Use the camera installed at the entrance of the work area to collect the portrait of the staff, match the collected portrait with the portrait stored in the database, verify the identity of the staff, and based on the verification result, choose whether to retrieve the work content of the verified staff. Combined with the experiments that can be performed on each operating table in the work area, selectively dispatch several video acquisition devices installed in the work area; Step 2: When executing each experimental step, the reaction video of the experiment is extracted through the video acquisition device. According to the video frame set corresponding to each reaction video and the odor sensor installed on each operating table in the working area, the real-time reaction feature set of the experiment is obtained. Based on the obtained results, a work safety assessment model for verification workers is constructed in multiple dimensions; The second step comprises: S201: when executing each experimental step, extracting the reaction video of the experiment through the video acquisition device, determining the start time of each reaction video, and in each reaction video, using the video frame acquired at the start time point as the key frame, and using the video frame set acquired except the start time point as the comparison frame set; S202: numbering each reaction video according to the execution order of each step in the experiment, and the numbering result is: p=1,2,…,q; q represents the total number of reaction videos; According to the dose of the target object taken by the verification staff, the reaction time f to the pth step in the experiment p Determine the time T versus the reaction time f p Divide and number each dividing point. The numbering result is: h=1,2,…,f p / T;f p / T indicates the total number of partition points; When the verification staff performs step p, the real-time comparison frame collected in the reaction video numbered p is compared with the key frame to verify the real-time reaction feature set M of the experiment. p To obtain, , where H 1p =t 1p +T,t 1p Indicates the time value corresponding to the start time point in the reaction video numbered p. Indicates that the target object is at t 1p +h*T time when the jth type of physical change occurs, Indicates that the target object is at t 1p +1*T time when the jth type of physical change occurs, Indicates that when the target object is at t 1p +h*T, when the jth type of physical change occurs at time, the degree of physical change of the target object, It indicates the degree of physical change of the target object when the target object undergoes the jth type of physical change at time t1p+1*T. T represents the time interval value, and H 1p Indicates the time value corresponding to the first division point in the reaction video numbered p, H hp Indicates the time value corresponding to the hth division point in the reaction video numbered p; S203: According to the type of physical change that occurs to the target object during each reaction time, the reaction feature set M is p Perform classification processing and obtain M p ={M p1 ,M p2 ,M p3 ,M p4 }, where M p1 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 1, M p2 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 2, M p3 It represents the real-time reaction feature set of the experiment when the verification staff performs the pth step when the physical change type of the target object is 3, M p4 It means that when the physical change type of the target object is 4, when the verification staff performs the pth step, the real-time reaction feature set of the experiment takes each type of physical change as a dimension, and based on the classification processing results, the verification staff's operation safety assessment model is constructed in multiple dimensions. The specific method is as follows: The jth type of physical change is taken as the jth dimension; In the time period [t1p+T, t1p+fp], based on the acquired standard comparison frame, when the target object undergoes the jth type of physical change at the moment Hhp, the standard gas concentration generated for the target object Acquire, and build the j-th dimension operation safety assessment model based on the acquired information. The specific formula is: ； in, It indicates that at the time H(h+1)p, the safety assessment coefficient of the verification staff in the jth dimension is predicted. It indicates the degree of physical change of the target object when the target object undergoes the jth type of physical change at time H(h+1)p. It indicates the standard gas concentration generated by the target object when the target object undergoes the jth type of physical change at time H(h+1)p; In the time period [t1p+T, t1p+fp], when the target object undergoes the fourth type of physical change at the moment Hhp, the standard gas concentration generated on the target object Acquire and build the fourth dimension of the operation safety assessment model based on the acquired information. The specific formula is: ； in, It indicates that at the moment H(h+1)p, the safety assessment coefficient of the verification staff in the fourth dimension is predicted. It indicates the degree of physical change of the target object when the target object undergoes the fourth type of physical change at time H(h+1)p. It indicates the standard gas concentration generated by the target object when the target object undergoes the fourth type of physical change at the moment H(h+1)p; Step 3: Use the video acquisition device installed in the work area to obtain the real-time status characteristics of the verification staff in each reaction time period, build a risk assessment model based on the acquired information, and predict the real-time risk factor of the experiment; Step 4: Draw a safety portrait of the verification staff based on the knowledge graph; The step 4 includes: using the personnel number of the verification staff as the entity of the operation safety portrait, using the operation safety assessment model of the verification staff in each dimension and the risk assessment model of the experiment as the entity attributes, using the safety assessment coefficient of each operation safety assessment model and the assessment coefficient of the risk assessment model as the attribute value of the entity attribute, and obtaining the operation safety portrait of the verification staff according to the entity-attribute-attribute value drawing method; Step 5: Predict the real-time operation safety assessment index of the verification staff, and manage the work status of the verification staff based on the prediction results; The step five comprises: according to E μ =a1×U 1μ +a2×U 2μ +a3×U 3μ +a4×U 4μ +a5×F μ For verification staff in H (h+1)p The operation safety assessment index at the time is predicted, where a1+a2+a3+a4+a5=1 and a1, a2, a3, a4, a5 all represent proportional coefficients, μ=H (h+1)p , U 1μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient on the first dimension, U 2μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the second dimension, U 3μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the third dimension, U 4μ Indicates that H (h+1)p At this moment, verify the worker's work safety assessment coefficient in the fourth dimension, F μ Indicates that the experiment was conducted in H (h+1)p The risk factor at the moment; When 0.5≤E μ When ≤1, the verification staff's work safety portrait is transmitted to the experimental terminal. When the experimental terminal receives the verification staff's work safety portrait, the verification staff's work is terminated; When 0≤E μ When the value is less than 0.5, the operation safety portrait of the verification staff will be transmitted to the experimental terminal. When the experimental terminal receives the operation safety portrait of the verification staff, it will send an early warning message to the intelligent terminal of the verification staff.

2. The method for managing job safety portraits based on a large model according to claim 1, characterized in that: The step 1 includes: when the collected personnel portrait is the same as a random personnel portrait stored in the database, the identity verification of the staff member is successful, and the work content of the verification staff member is retrieved at this time; According to the retrieved work content, the experiments that the verification staff need to perform are determined. Combined with the experiments that can be performed on each operating table in the work area, the experimental operating table of the verification staff is determined. It is judged whether the line connecting the center of the experimental operating table surface and the center of the lens of each video acquisition device is the central axis of the video acquisition device. If so, the corresponding video acquisition device is scheduled. If not, the corresponding video acquisition device is not scheduled.

3. The method for managing job safety portraits based on a large model according to claim 2, characterized in that: The specific method of obtaining the real-time response feature set of the verification staff when executing step p in step S202 is: Will be in t 1p +h*T time acquisition of the comparison frame and at t 1p The key frames collected at time t are compared to 1p The physical changes that occur at time +h*T are taken as the experimental 1p +h*T time, the set of real-time reaction features obtained is used as the real-time reaction feature set of the experiment. The experiment is performed at t 1p +h*Reaction characteristics at time T= , where j = 1, 2, 3, 4, representing the number corresponding to each physical change type, H hp =t 1p +h*T; The types of physical changes include Type 1: the shape and size of the target object changes, Type 2: new substances adhere to the surface of the target object, Type 3: the color of the target object changes, and Type 4: the target object produces gas.

4. The method for managing job safety portraits based on a large model according to claim 3 is characterized in that: The step three comprises: S301: Verification staff in [t 1p +T,t 1p +f p ] to obtain the real-time status characteristics within the time period, the status characteristics including the fatigue value D of the verification staff and the distance value L between the verification staff and the experimental operation table; S302: Construct a risk assessment model to predict the real-time risk factor of the experiment.

5. The method for managing job safety portraits based on a large model according to claim 4 is characterized in that: The specific formula of the risk assessment model constructed in S302 is: ； in, Indicates that the verification staff is in H (h+1)p Fatigue value at the moment, Indicates that the verification staff is in H (h+1)p The distance from the experimental operation table at the moment, Indicates that the experiment was conducted in H (h+1)p The risk coefficient at a given moment, exp represents an exponential function with the natural constant e as the base and e=2.

72.

6. A large-model-based job safety portrait management system applied to the large-model-based job safety portrait management method according to any one of claims 1 to 5, characterized in that: The system includes a video acquisition device scheduling module, an operation safety assessment model building module, a risk assessment model building module, an operation safety portrait drawing module and an operation safety management module; The video acquisition device scheduling module is used to selectively schedule a number of video acquisition devices installed in the working area; The job safety assessment model building module is used to build a job safety assessment model for verification workers in multiple dimensions; The risk assessment model construction module is used to predict the real-time risk factor of the experiment through the constructed risk assessment model; The operation safety portrait drawing module is used to draw the operation safety portrait of the verification staff; The operation safety management module is used to perform safety management on the working status of the verification staff.

Images