A security management and on-duty performance management system for non-staff members and its management method

Abstract

The present invention relates to the field of access management technology, and discloses a management system for the safety management and on-the-job performance of outsiders and a management method thereof, comprising the following steps: by integrating the Internet of Things, artificial intelligence and big data analysis technology, real-time monitoring, identity authentication, on-the-job status tracking and performance behavior analysis of outsiders entering the work area are realized. The identity information, functional information and visited area information of the visiting outsiders are obtained, and the activity trajectory of the outsiders during their on-the-job period is monitored in real time to ensure that they perform their duties within the specified area, and the performance risk of outsiders in the specified area is estimated by constructing a risk estimation model, and an early warning is automatically triggered if there is a high risk. The present invention can solve the problems of low efficiency, incomplete supervision, inaccurate data and the like in the prior art in the on-the-job performance management of outsiders, and provide data support for optimizing the management process.

Description

Technical Field

[0001] The present invention relates to the field of access management technology, and in particular to a management system and a management method for security management of outsiders and on-the-job performance of duties. Background Art

[0002] Safety management and on-the-job performance evaluation of outsiders are crucial links in the operation of modern enterprises and institutions. Its necessity stems from multiple needs such as safety risk prevention and control, responsibility tracing and efficiency improvement. For example, outsiders (such as visitors, contractors, temporary workers) are unfamiliar with the on-site environment and may mistakenly enter dangerous areas, such as high-voltage power rooms and chemical warehouses. Untrained personnel in these places may cause accidents due to improper operation. If an accident or dispute occurs, it is necessary to clarify whether it is caused by illegal operations by outsiders. There are many advantages to implementing safety management and on-the-job performance evaluation of outsiders, including complete recording of personnel trajectory and behavior data, meeting regulatory requirements, preventing accidents, etc. At the same time, it can also replace manual inspections and reduce the investment of security personnel. Therefore, a safety management and on-the-job performance management system for outsiders and its management method are proposed. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a system and method for managing the safety and performance of duties of outsiders.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is:

[0005] The first aspect of the present invention provides a method for safety management and on-the-job performance management of outsiders, comprising the following steps:

[0006] Collect personal information and conduct facial recognition analysis on outsiders, and determine the identity level of outsiders based on the analysis results;

[0007] Based on the visitor's identity level, analyze the visitor's access rights to different work visit sub-areas and determine the access time window, while monitoring the visitor's real-time location in the work visit area in real time;

[0008] Generate the visitor's real-time path information based on the visitor's real-time location in the work visit area, and generate a visitor access risk estimation model based on the visitor's real-time path information;

[0009] Combined with the visitor access risk assessment model, risk assessment and risk warning are carried out for visitors in the work visit area;

[0010] When a visitor is working in the work visit area, the visitor's business function is determined based on the visitor's identity level, and the visitor's on-the-job performance indicators are evaluated based on the visitor's business function and the visitor access risk assessment model.

[0011] Furthermore, in a preferred embodiment of the present invention, personal information collection and face recognition analysis are performed on the external personnel, and the identity level of the external personnel is determined according to the analysis results, specifically as follows:

[0012] Determine the work visit area, collect the personal information of the external personnel at the entrance of the work visit area, and install face verification devices at the work visit area;

[0013] After recording the personal information of the external personnel, upload it to the management terminal of the work visit area, mark it as the target management terminal, and obtain the visit blacklist at the target management terminal;

[0014] Import the personal information of the external personnel into the visit blacklist for identity comparison, and determine whether the personal information of the external personnel exists in the visit blacklist. If so, mark the external personnel as a person refused entry at the target management terminal. If not, mark the external personnel as a visitor;

[0015] At the target management terminal, conduct reservation registration monitoring on the visitor, and determine whether there is a reservation registration when the visitor arrives. If so, generate the function information, target visit area, and target visit purpose of the visitor according to the information filled in when the visitor makes a reservation registration;

[0016] If not, collect the function information, target visit area, and target visit purpose of the visitor in real time and store them in the target management terminal. At the same time, connect the face verification device to the target management terminal so that the face verification device can verify the face of the visitor;

[0017] Based on the function information, target visit area, and target visit purpose of the visitor, divide the identity of the visitor, and obtain the identity level template in the target management terminal. Among them, different identity levels correspond to different function information, target visit areas, and target visit purposes;

[0018] Based on the identity level template, determine the identity level of the visitor.

[0019] Furthermore, in a preferred embodiment of the present invention, in combination with the identity level of the visitor, perform access permission analysis and access time window determination for different work visit sub-areas of the visitor, and at the same time, monitor the real-time position of the visitor in the work visit area in real time, specifically as follows:

[0020] In the work visit area, determine all work visit sub-areas, where the work visit sub-areas include the target visit area of the visitor;

[0021] All work visit sub - regions that visitors with different functional information can freely access, where all work visit sub - regions that can be freely accessed are all work visit sub - regions that do not require approval for access;

[0022] Analyze the functional information in the identity level of the visitor in the target management terminal, and determine all work visit sub - regions that the visitor can freely access, which are marked as the first - class work visit sub - regions. At the same time, determine the access time window for different work visit sub - regions;

[0023] Combine the access time windows of different work visit sub - regions to determine the access time window for the first - class work visit sub - regions;

[0024] Obtain all cameras in the work visit area, mark them as positioning cameras, and connect face verification devices in the work visit area to the positioning cameras. Through the face verification devices, perform real - time face verification and recognition processing on all personnel in the work visit area, and locate the real - time position of the visitor in the work visit area according to the face verification and recognition processing results.

[0025] Furthermore, in a preferred embodiment of the present invention, generating the real - time path information of the visitor according to the real - time position of the visitor in the work visit area, and generating a visitor access risk prediction model in combination with the real - time path information of the visitor, specifically:

[0026] Combine the real - time position of the visitor in the work visit area to generate the real - time path information of the visitor in the target management terminal, where the real - time path information of the visitor is obtained by connecting all the real - time positions that the visitor has passed through in the work visit area;

[0027] Calculate the residence time of the visitor at different real - time positions, so as to generate a complete timestamp of the visitor at different real - time positions, and import the complete timestamp of the visitor at different real - time positions into the real - time path information of the visitor to update the real - time path information, and obtain the updated real - time path information of the visitor;

[0028] Introduce the mRMR algorithm, an auto - encoder, and a multi - layer perceptron in the target management terminal. Based on the mRMR algorithm, first construct an mRMR model, and import the updated real - time path information of the visitor into the blank mRMR model. At the same time, import the identity level of the visitor, the first - class work visit sub - regions corresponding to the visitor, and the access time window of the corresponding first - class work visit sub - regions;

[0029] In the mRMR blank model, all the imported data is converted into feature data for encoding through an autoencoder and a multi-layer perceptron. The mRMR blank model after importing the data is trained, and during the training process, a particle swarm algorithm is introduced to optimize the network structure. The evaluation accuracy of the mRMR blank model after importing the data is used as the particle fitness.

[0030] Obtain and continuously update the particle position and particle velocity according to the particle fitness. When the termination condition is met, obtain the number of autoencoder layers and the number of hidden layers of the multi-layer perceptron according to the optimal particle position, and construct a visitor access risk prediction model.

[0031] Among them, the visitor access risk prediction model predicts the travel route of the visitor and evaluates the risk level of the visitor in the work visit area according to the real-time position and stay time of the visitor in the work visit area.

[0032] Furthermore, in a preferred embodiment of the present invention, in combination with the visitor access risk prediction model, risk prediction and risk warning are performed on the visitor in the work visit area, specifically as follows:

[0033] In the work visit area, if the face verification device detects that the real-time position of the visitor is not in the corresponding first-class work visit sub-area, or the stay time of the visitor in the corresponding first-class work visit sub-area is not within the corresponding access time window, a visitor crisis alarm is generated in the target management terminal.

[0034] When a visitor crisis alarm is generated in the target management terminal, the buzzer in the work visit area is connected through the target management terminal and a buzzer sound is emitted.

[0035] A preset target prediction time is set. If the real-time position of the visitor remains in the corresponding first-class work visit sub-area, then based on the visitor access risk prediction model, predict the travel route of the visitor within the target prediction time, and evaluate the risk level of the visitor in the work visit area according to the travel route of the visitor within the target prediction time.

[0036] A preset dangerous risk level is set. If the risk level of the visitor in the work visit area is less than the dangerous risk level, no visitor crisis alarm is generated.

[0037] If the risk level of the visitor in the work visit area is greater than the dangerous risk level, the warning broadcast in the work visit area is connected through the target management terminal, and the visitor is reminded based on the warning broadcast in the work visit area.

[0038] Further, in a preferred embodiment of the present invention, when a visitor is working in the work visit area, based on the identity level of the visitor, the business function of the visitor is determined, and combined with the business function of the visitor and the visitor access risk prediction model, an on-the-job performance index evaluation is carried out on the visitor, specifically as follows:

[0039] Determine the business function - visit purpose template in the target management terminal, and based on the identity level of the visitor, determine the business function of the visitor in the work visit area in the business function - visit purpose template, and label it as the target business function;

[0040] Determine the work visit sub - area corresponding to the target business function in the work visit area, label it as the secondary work visit sub - area, and at the same time determine the corresponding on - the - job performance time range of the target business function in different secondary work visit sub - areas, and label it as the target on - the - job performance time range;

[0041] If the on - the - job performance area of the visitor in the work visit area completely coincides with the secondary work visit sub - area, and the on - the - job performance time is completely maintained within the corresponding target on - the - job performance time range, then the on - the - job performance index of the visitor is evaluated as excellent in the target management terminal;

[0042] If the on - the - job performance area of the visitor in the work visit area does not completely coincide with the secondary work visit sub - area, or the on - the - job performance time is not completely maintained within the corresponding target on - the - job performance time range, then the on - the - job performance index of the visitor is evaluated as to be excellent in the target management terminal.

[0043] The second aspect of the present invention also provides a security management and on - the - job performance management system for external personnel. The management system includes a memory and a processor. The memory stores a management method. When the management method is executed by the processor, the following steps are realized:

[0044] Collect personal information and perform face recognition analysis on external personnel, and determine the identity level of external personnel according to the analysis results;

[0045] Combined with the identity level of the visitor, analyze the access permission of the visitor to different work visit sub - areas and determine the access time window, and at the same time monitor the real - time location of the visitor in the work visit area in real - time;

[0046] Generate real - time path information of the visitor according to the real - time location of the visitor in the work visit area, and generate a visitor access risk prediction model in combination with the real - time path information of the visitor;

[0047] Combined with the visitor access risk prediction model, conduct risk prediction and risk warning on the visitor in the work visit area;

[0048] When a visitor is working in the work arrival area, based on the visitor's identity level, determine the visitor's business function, and combine the visitor's business function and the visitor access risk prediction model to evaluate the visitor's on-duty performance indicators.

[0049] The technical defects existing in the background art solved by the present invention have the following beneficial effects: By integrating Internet of Things, artificial intelligence, and big data analysis technologies, it realizes real-time monitoring, identity verification, on-duty status tracking, and performance behavior analysis of external personnel entering the work area. Obtain the identity information, function information, and arrival area information of the visiting external personnel, and implement real-time monitoring of the activity trajectories of external personnel during their on-duty period to ensure the purpose of their performing duties within the specified area, and predict the performance risks of external personnel within the specified area by constructing a risk prediction model. If the risk is too large, an alarm will be automatically triggered. The present invention can solve the problems in the prior art such as low management efficiency, incomplete supervision, and inaccurate data in the on-duty performance management of external personnel, and provide data support for optimizing the management process. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0051] Figure 1 Shows a flowchart of a method for the safety management and on-duty performance management of external personnel;

[0052] Figure 2 Shows a flowchart of a method for constructing a visitor access risk prediction model;

[0053] Figure 3 Shows a program view of a system for the safety management and on-duty performance management of external personnel. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] In order to more clearly understand the above-mentioned objects, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0055] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0056] Figure 1 A flowchart showing a method for the safety management of external personnel and the management of on-the-job performance is provided, including the following steps:

[0057] S102: Collect personal information of external personnel and perform face recognition analysis, and determine the identity level of external personnel according to the analysis results;

[0058] S104: Combine the identity level of the visitor to analyze the access permission of different work arrival sub-areas for the visitor and determine the access time window, and at the same time, monitor the real-time location of the visitor in the work arrival area in real time;

[0059] S106: Generate real-time path information of the visitor according to the real-time location of the visitor in the work arrival area, and generate a visitor access risk prediction model in combination with the real-time path information of the visitor;

[0060] S108: Combine the visitor access risk prediction model to perform risk prediction and risk warning on the visitor in the work arrival area;

[0061] S110: When the visitor is working in the work arrival area, based on the identity level of the visitor, determine the business function of the visitor, and combine the business function of the visitor and the visitor access risk prediction model to evaluate the on-the-job performance indicators of the visitor.

[0062] Further, in a preferred embodiment of the present invention, the step of collecting personal information of external personnel and performing face recognition analysis, and determining the identity level of external personnel according to the analysis results is specifically as follows:

[0063] Determine the work arrival area, collect personal information of external personnel at the entrance of the work arrival area, and install a face verification device at the work arrival area;

[0064] After recording the personal information of the external personnel, upload it to the management terminal of the work arrival area, mark it as the target management terminal, and obtain the blacklist of visitors at the target management terminal;

[0065] Import the personal information of the external personnel into the blacklist of visitors for identity comparison, and determine whether the personal information of the external personnel exists in the blacklist of visitors. If so, mark the external personnel as a person refused entry at the target management terminal. If not, mark the external personnel as a visitor;

[0066] At the target management terminal, monitor the reservation registration of the visitor, and determine whether there is a reservation registration when the visitor arrives. If so, generate the function information, target arrival area, and target arrival purpose of the visitor according to the information filled in by the visitor during reservation registration;

[0067] If not, the functional information, target visiting area and target visiting purpose of the visitor are collected in real time and stored in the target management terminal, and at the same time, the face verification device is connected to the target management terminal so that the face verification device can perform face verification on the visitor;

[0068] Based on the functional information, target visiting area and target visiting purpose of the visitor, the visitor is classified into identities, and an identity level template is obtained in the target management terminal, wherein different identity levels correspond to different functional information, target visiting area and target visiting purpose;

[0069] Determine the visitor's identity level based on the identity level template.

[0070] It should be noted that the purpose of collecting the personal information of outsiders is to determine whether the outsiders are members of the blacklist in the work visit area. The blacklist is a list of people who are prohibited from entering. If the outsider has been registered and blacklisted, the outsider has no right to enter the work visit area. If not blacklisted, he / she is marked as a visitor. Before the visitor visits, it can be checked whether he / she has made an appointment. When making an appointment, the visitor's visit status will be recorded, so as to facilitate the on-the-job performance and security management of the visitor in the following period. At the same time, the visitor's information will be transmitted to the face verification device to facilitate the location of the visitor. The purpose of determining the visitor's identity level is to determine the visitor's authority in the work visit area.

[0071] Furthermore, in a preferred embodiment of the present invention, the visitor access risk prediction model is combined to perform risk prediction and risk warning on visitors in the work visit area, specifically:

[0072] In the work visit area, if the face verification device detects that the visitor's real-time location is not in the corresponding work visit sub-area, or the visitor's stay time in the corresponding work visit sub-area is not within the corresponding admission time window, a visitor crisis alarm is generated in the target management terminal;

[0073] When a visitor crisis alarm is generated in the target management terminal, the buzzer in the work visit area is connected through the target management terminal and sounds;

[0074] A target prediction time is preset. If the real-time location of the visitor is maintained in a corresponding type of work visit sub-area, the visitor's travel route within the target prediction time is predicted based on the visitor access risk estimation model, and the risk level of the visitor in the work visit area is evaluated based on the visitor's travel route within the target prediction time.

[0075] The danger risk level is preset. If the risk level of the visitor in the work visit area is less than the danger risk level, no visitor crisis alarm is generated;

[0076] If the risk level of a visitor in the work visit area is greater than the dangerous risk level, connect to the warning broadcast of the work visit area through the target management terminal and remind the visitor based on the warning broadcast of the work visit area.

[0077] It should be noted that if it is detected that the real-time position of the visitor is not in the corresponding first-class work visit sub-area, that is, the visitor steps into the area where their access is prohibited, or the stay time of the visitor in the first-class work visit sub-area is not within the specified time, an alarm needs to be issued to prevent the visitor from being in danger or performing dangerous behaviors in the specified area. Therefore, connect the buzzer in the work visit area and sound the buzzer. On the contrary, if the real-time position of the visitor remains in the corresponding first-class work visit sub-area, it is necessary to predict whether the visitor will pass through the area where they are prohibited from stepping into after a period of time. By accessing the risk prediction model, the movement route of the visitor can be judged because there is a connection between tasks. For example, after working in area a, the visitor may go to area b, and area b is the area where the visitor is prohibited from stepping into. Then, through the access risk prediction model, it can be predicted that the visitor may step into area b after working in area a. At this time, the probability of the visitor stepping into area b will be calculated, that is, the risk level of the visitor in the work visit area will be evaluated. If the risk level of the visitor in the work visit area is less than the dangerous risk level, it proves that the visitor has a relatively low probability of going to the prohibited area. On the contrary, if the risk level is higher, the visitor needs to be reminded not to step into the area.

[0078] Further, in a preferred embodiment of the present invention, when the visitor is working in the work visit area, based on the identity level of the visitor, determine the business function of the visitor, and combine the business function of the visitor and the visitor access risk prediction model to evaluate the on-duty performance index of the visitor. Specifically:

[0079] Determine the business function - visit purpose template in the target management terminal, and based on the identity level of the visitor, determine the business function of the visitor in the work visit area in the business function - visit purpose template, and mark it as the target business function;

[0080] Determine the work visit sub-area corresponding to the target business function in the work visit area, and mark it as the second-class work visit sub-area. At the same time, determine the on-duty performance time range corresponding to the target business function in different second-class work visit sub-areas, and mark it as the target on-duty performance time range;

[0081] If the on-duty performance area of the visitor in the work visit area completely coincides with the second-class work visit sub-area, and the on-duty performance time completely remains within the corresponding target on-duty performance time range, the on-duty performance index of the visitor will be evaluated as excellent in the target management terminal;

[0082] If there is an incomplete overlap between the on-duty performance area and the secondary work visit sub-area in the work visit area of the visitor, or the on-duty performance time does not completely maintain within the corresponding target on-duty performance time range, the on-duty performance index of the visitor will be evaluated as to be optimized in the target management terminal.

[0083] It should be noted that evaluating the on-duty performance index of the visitor is to judge the work attitude and completion situation of the visitor in the work visit area. Under normal circumstances, the visitor works in the work visit area according to the rules and regulations, and his on-duty performance index is evaluated as excellent, that is, the visitor will set foot in the secondary work visit sub-area in the work visit area, and the on-duty performance time range in the secondary work visit sub-area maintains within the standard range. On the contrary, if the visitor sets foot in the area outside the secondary work visit sub-area, it proves that the visitor's work may be careless, and if the on-duty performance time range does not maintain within the standard range, it proves that the visitor's work efficiency is too low, or the visitor's work attitude is not serious. Therefore, the on-duty performance index of the visitor is evaluated as to be optimized. Different work visit areas correspond to different work tasks that can be completed. Therefore, after determining the business function of the visitor, the secondary work visit sub-area can be obtained according to the business function.

[0084] Figure 2 The method flow chart for constructing a visitor access risk prediction model is shown, including the following steps:

[0085] S202: Combining the identity level of the visitor, analyze the access permission of the visitor to different work visit sub-areas and determine the access time window, and at the same time, monitor the real-time location of the visitor in the work visit area in real time;

[0086] S204: Generate the real-time path information of the visitor according to the real-time location of the visitor in the work visit area, and generate a visitor access risk prediction model in combination with the real-time path information of the visitor.

[0087] Further, in a preferred embodiment of the present invention, the combining the identity level of the visitor, analyzing the access permission of the visitor to different work visit sub-areas and determining the access time window, and at the same time, monitoring the real-time location of the visitor in the work visit area in real time, specifically:

[0088] In the work visit area, determine all work visit sub-areas, where the work visit sub-areas include the target visit area of the visitor;

[0089] Obtain all work visit sub-areas that visitors with different function information can freely access, where all work visit sub-areas that can be freely accessed are all work visit sub-areas that do not require approval for access;

[0090] Analyze the functional information in the identity level of the visitor in the target management terminal, determine all the work visit sub-areas that the visitor can freely access, label them as the first-class work visit sub-areas, and at the same time determine the access time windows of different work visit sub-areas;

[0091] Combine the access time windows of different work visit sub-areas to determine the access time window of the first-class work visit sub-areas;

[0092] Obtain all the cameras in the work visit area, label them as positioning cameras, connect the face verification devices in the work visit area to the positioning cameras, and through the affiliated face verification devices, perform real-time face verification and recognition processing on all personnel in the work visit area, and locate the real-time position of the visitor in the work visit area according to the face verification and recognition processing results.

[0093] It should be noted that the work visit sub-areas that visitors with different functional information can freely access are different, that is, the first-class work visit sub-areas are different. After determining the first-class work visit sub-areas of the visitor, the corresponding access time window needs to be determined at the same time, because some areas are not allowed to enter within the specified time, and entering will cause risks. The purpose of obtaining the positioning cameras is to connect the face verification devices to verify the real-time position of the visitor in the work visit area.

[0094] Further, in a preferred embodiment of the present invention, generating the real-time path information of the visitor according to the real-time position of the visitor in the work visit area, and generating a visitor access risk prediction model in combination with the real-time path information of the visitor, specifically:

[0095] Combine the real-time position of the visitor in the work visit area to generate the real-time path information of the visitor in the target management terminal, where the real-time path information of the visitor is obtained by connecting all the real-time positions that the visitor has passed through in the work visit area;

[0096] Calculate the residence time of the visitor at different real-time positions, thereby generating the complete time stamps of the visitor at different real-time positions, and import the complete time stamps of the visitor at different real-time positions into the real-time path information of the visitor to update the real-time path information, and obtain the updated real-time path information of the visitor;

[0097] Introduce the mRMR algorithm, the autoencoder and the multi-layer perceptron in the target management terminal. Based on the mRMR algorithm, first construct the mRMR model, and import the updated real-time path information of the visitor into the mRMR blank model, and at the same time import the identity level of the visitor, the first-class work visit sub-areas corresponding to the visitor, and the access time windows of the corresponding first-class work visit sub-areas;

[0098] In the mRMR blank model, all the imported data is converted into feature data for encoding through an autoencoder and a multi-layer perceptron. The mRMR blank model after importing the data is trained, and during the training process, a particle swarm algorithm is introduced to optimize the network structure. The evaluation accuracy of the mRMR blank model after importing the data is used as the particle fitness;

[0099] Obtain and continuously update the particle position and particle velocity according to the particle fitness. When the termination condition is met, obtain the number of autoencoder layers and the number of hidden layers of the multi-layer perceptron according to the optimal particle position, and construct a visitor access risk prediction model;

[0100] Among them, the visitor access risk prediction model predicts the travel route of the visitor and evaluates the risk level of the visitor in the work visit area according to the real-time position and stay time of the visitor in the work visit area.

[0101] It should be noted that the different real-time positions of the visitor combined are his real-time path. After the real-time path is determined, the next travel route of the visitor can be inferred according to the path position, speed, stay time, etc. The mRMR algorithm is used to select variables with low redundancy between features and high correlation between features and the dependent variable, and has the advantages of fast calculation speed and good stability. Therefore, the mRMR algorithm is used to construct the prediction model, combined with an autoencoder and a multi-layer perceptron. The autoencoder and the multi-layer perceptron are devices for determining the number of layers and the number of hidden layers in the mRMR blank model. Different numbers of layers and hidden layers determine the prediction ability and prediction direction of the mRMR blank model. The updated real-time path information of the visitor will be imported into the mRMR blank model, and at the same time, the identity level of the visitor, a type of work visit sub-area corresponding to the visitor, and the access time window of the corresponding type of work visit sub-area will be imported to realize feature data extraction, and combined with the particle swarm algorithm, the training of the model can be realized. When the particle position and particle velocity meet the termination condition, that is, when the corresponding position and velocity are met, a visitor access risk prediction model can be constructed to evaluate the possible travel route of the visitor in the work visit area and evaluate the risk level of the visitor in the work visit area.

[0102] As Figure 3 shown, the second aspect of the present invention also provides a safety management and on-the-job performance management system for foreign personnel. The management system includes a memory 31 and a processor 32. The memory 31 stores a management method. When the management method is executed by the processor 32, the following steps are realized:

[0103] Collect personal information of foreign personnel and perform face recognition analysis, and determine the identity level of foreign personnel according to the analysis results;

[0104] Combined with the identity level of the visitor, analyze the access permissions of the visitor to different sub-regions for work visits and determine the access time window, and at the same time, monitor the real-time location of the visitor in the work visit area in real time;

[0105] Generate the real-time path information of the visitor according to the real-time location of the visitor in the work visit area, and generate a visitor access risk prediction model in combination with the real-time path information of the visitor;

[0106] Combined with the visitor access risk prediction model, conduct risk prediction and risk warning for the visitor in the work visit area;

[0107] When the visitor is working in the work visit area, based on the identity level of the visitor, determine the business functions of the visitor, and conduct an on-the-job performance index evaluation for the visitor in combination with the business functions of the visitor and the visitor access risk prediction model.

[0108] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for the safety management of external personnel and the management of on-duty performance, characterized in that, It includes the following steps: Collect personal information of external personnel and conduct face recognition analysis, and determine the identity level of external personnel according to the analysis results; Combined with the identity level of the visitor, analyze the access permission of the visitor to different sub-regions of the work visit and determine the access time window, and at the same time monitor the real-time location of the visitor in the work visit area in real time; Generate the real-time path information of the visitor according to the real-time location of the visitor in the work visit area, and generate a visitor access risk prediction model in combination with the real-time path information of the visitor; Combined with the visitor access risk prediction model, conduct risk prediction and risk warning for the visitor in the work visit area; When the visitor is working in the work visit area, based on the identity level of the visitor, determine the business function of the visitor, and combined with the business function of the visitor and the visitor access risk prediction model, evaluate the on-the-job performance index of the visitor; Among them, the step of generating the real-time path information of the visitor according to the real-time location of the visitor in the work visit area and generating a visitor access risk prediction model in combination with the real-time path information of the visitor is specifically as follows: Combined with the real-time location of the visitor in the work visit area, generate the real-time path information of the visitor in the target management terminal. Among them, the real-time path information of the visitor is obtained by connecting all the real-time locations that the visitor has passed through in the work visit area, so as to obtain the real-time path information of the visitor; Calculate the residence time of the visitor at different real-time locations, so as to generate the complete time stamp of the visitor at different real-time locations, and import the complete time stamp of the visitor at different real-time locations into the real-time path information of the visitor to update the real-time path information, and obtain the updated real-time path information of the visitor; Introduce the mRMR algorithm in the target management terminal, and introduce the autoencoder and the multi-layer perceptron. Based on the mRMR algorithm, first construct the mRMR model, and import the updated real-time path information of the visitor into the blank mRMR model, and at the same time import the identity level of the visitor, the first-class sub-region of the work visit corresponding to the visitor, and the access time window of the corresponding first-class sub-region of the work visit; In the blank mRMR model, convert all the imported data into feature data for encoding through the autoencoder and the multi-layer perceptron, train the blank mRMR model after importing the data, and introduce the particle swarm optimization algorithm for network structure optimization during the training process, and use the evaluation accuracy of the blank mRMR model after importing the data as the particle fitness; Continuously update the particle position and particle velocity according to the particle fitness. When the termination condition is met, obtain the number of layers of the autoencoder and the number of hidden layers of the multi-layer perceptron according to the optimal particle position, and construct a visitor access risk prediction model; Among them, the visitor access risk prediction model predicts the travel route of the visitor and evaluates the risk level of the visitor in the work visit area according to the real-time location and residence time of the visitor in the work visit area.

2. The method for the safety management of external personnel and the on-the-job performance management according to claim 1, characterized in that, The step of collecting personal information of external personnel and conducting face recognition analysis, and determining the identity level of external personnel according to the analysis results is specifically as follows: Determine the work visit area, collect the personal information of external personnel at the entrance of the work visit area, and install a face verification device at the work visit area; After recording the personal information of the external personnel, upload it to the management terminal of the work visit area, mark it as the target management terminal, and obtain the visit blacklist at the target management terminal; Import the personal information of the external personnel into the visit blacklist for identity comparison, and judge whether the personal information of the external personnel exists in the visit blacklist. If so, mark the external personnel as refused entry personnel at the target management terminal. If not, mark the external personnel as visitors; At the target management terminal, monitor the appointment registration of visitors, and judge whether there is an appointment registration when the visitors arrive. If so, generate the functional information, target visit area and target visit purpose of the visitors according to the information filled in when the visitors make an appointment registration; If not, collect the functional information, target visit area and target visit purpose of the visitors in real time and store them in the target management terminal. At the same time, connect the face verification device to the target management terminal so that the face verification device can verify the faces of the visitors; Based on the functional information, target visit area and target visit purpose of the visitors, divide the identities of the visitors, and obtain the identity level template in the target management terminal. Among them, different identity levels correspond to different functional information, target visit areas and target visit purposes; Based on the identity level template, determine the identity level of the visitors.

3. The method for the safety management of external personnel and the on-the-job performance management according to claim 1, wherein Combined with the identity level of the visitors, analyze the access permission of different work visit sub-areas for the visitors and determine the access time window, and at the same time, monitor the real-time position of the visitors in the work visit area in real time. Specifically: In the work visit area, determine all work visit sub-areas, where the work visit sub-areas include the target visit area of the visitors; Obtain all work visit sub-areas that visitors with different functional information can freely access, where all work visit sub-areas that can be freely accessed are all work visit sub-areas that do not require approval to access; Analyze the functional information in the identity level of the visitors in the target management terminal, and determine all work visit sub-areas that the visitors can freely access, mark them as the first-class work visit sub-areas, and at the same time determine the access time window of different work visit sub-areas; Combined with the access time window of different work visit sub-areas, determine the access time window of the first-class work visit sub-areas; Obtain all cameras in the work visit area, mark them as positioning cameras, and connect the face verification device in the work visit area to the positioning cameras. Through the affiliated face verification device, perform face verification and recognition processing on all personnel in the work visit area in real time, and locate the real-time position of the visitors in the work visit area according to the face verification and recognition processing results.

4. The method for the safety management of external personnel and the management of on-the-job performance according to claim 1, characterized in that Combined with the visitor access risk prediction model, conduct risk prediction and risk warning on the visitors in the work visit area. Specifically: In the work visit area, if the face verification device detects that the real-time location of the visitor is not in the corresponding first-class work visit sub-area, or the stay time of the visitor in the corresponding first-class work visit sub-area is not within the corresponding access time window, a visitor crisis alarm is generated in the target management terminal; When a visitor crisis alarm is generated in the target management terminal, the buzzer in the work visit area is connected through the target management terminal and a buzzer sound is emitted; A preset target prediction time is set. If the real-time location of the visitor remains in the corresponding first-class work visit sub-area, then based on the visitor access risk prediction model, the travel route of the visitor within the target prediction time is predicted, and according to the travel route of the visitor within the target prediction time, the risk level of the visitor in the work visit area is evaluated; A preset dangerous risk level is set. If the risk level of the visitor in the work visit area is less than the dangerous risk level, no visitor crisis alarm is generated; If the risk level of the visitor in the work visit area is greater than the dangerous risk level, the warning broadcast in the work visit area is connected through the target management terminal, and the visitor is reminded based on the warning broadcast in the work visit area.

5. The method for the safety management of external personnel and the management of on-the-job performance according to claim 1, characterized in that When the visitor is working in the work visit area, based on the identity level of the visitor, the business function of the visitor is determined, and combined with the business function of the visitor and the visitor access risk prediction model, an on-the-job performance index evaluation of the visitor is carried out. Specifically: The business function - visit purpose template is determined in the target management terminal, and based on the identity level of the visitor, the business function of the visitor in the work visit area is determined in the business function - visit purpose template and marked as the target business function; In the work visit area, the work visit sub-area corresponding to the target business function is determined and marked as the second-class work visit sub-area. At the same time, the on-the-job performance time range corresponding to the target business function in different second-class work visit sub-areas is determined and marked as the target on-the-job performance time range; If the on-the-job performance area of the visitor in the work visit area completely coincides with the second-class work visit sub-area and the on-the-job performance time completely remains within the corresponding target on-the-job performance time range, the on-the-job performance index of the visitor is evaluated as excellent in the target management terminal; If the on-the-job performance area of the visitor in the work visit area does not completely coincide with the second-class work visit sub-area, or the on-the-job performance time does not completely remain within the corresponding target on-the-job performance time range, the on-the-job performance index of the visitor is evaluated as to be optimized in the target management terminal.

6. An outsider safety management and on-duty performance management system, characterized in that, The management system includes a memory and a processor. The memory stores a management method program. When the management method program is executed by the processor, the steps of a safety management and on-the-job performance management method for external personnel as described in any one of claims 1-5 are implemented.

Images