CN117690166B

CN117690166B - Security monitoring method and system for electric control cabinet

Info

Publication number: CN117690166B
Application number: CN202410148544.6A
Authority: CN
Inventors: 童欣; 袁呈呈
Original assignee: Hubei Senyuan Century Electric Group Co ltd
Current assignee: Hubei Senyuan Century Electric Group Co ltd
Priority date: 2024-02-02
Filing date: 2024-02-02
Publication date: 2024-04-16
Anticipated expiration: 2044-02-02
Also published as: CN117690166A

Abstract

The invention provides a security monitoring method and a security monitoring system for an electric control cabinet, wherein the method comprises the following steps: acquiring an operation work order from a management center and extracting a work order number; generating a verification code according to the work order number; acquiring first image information containing the surrounding environment of the power control cabinet, and analyzing whether personnel activities exist in the first image information by utilizing an image recognition algorithm; if personnel activities exist, activating a second image acquisition device; collecting second image information of a target person beside the electric control cabinet; analyzing whether the dressing configuration of the target personnel meets the safety requirement; if the safety requirements are met, activating a cabinet door control device arranged outside the electric control cabinet body, and acquiring biological information of a target person and a request code input by the target person; when the personnel passes the verification and the verification code is consistent with the request code, the cabinet door of the electric control cabinet is opened through the cabinet door control device. The invention has the effect of improving the operation safety of the power control cabinet.

Description

Security monitoring method and system for electric control cabinet

Technical Field

The invention belongs to the technical field of control cabinet safety management, and particularly relates to a security monitoring method and system of an electric control cabinet.

Background

The power control cabinet is a key device in the power grid system, and the arrangement of the power control cabinet meets the requirement of normal operation of the power grid system. Along with the development of science and technology, the electric control cabinet has been upgraded from a traditional control cabinet to an intelligent electric control cabinet capable of carrying out data monitoring in real time. The intelligent power control cabinet can acquire the operation data of the control cabinet and related circuits in real time through various sensors installed in the cabinet, and can upload the operation data to the power management center. However, for safety reasons, various operations such as power-off, power-on, and operation suspension of the intelligent power control cabinet are performed manually. However, the existing intelligent power control cabinet generally has only password limitation on the cabinet door, and any person can open the cabinet door to operate as long as knowing the password, so that great security and protection potential safety hazards exist.

Disclosure of Invention

The invention provides a security monitoring method and a security monitoring system for an electric control cabinet, which aim to solve the problem that the electric control cabinet has great security hidden danger.

In a first aspect, the present invention provides a security monitoring method for an electric control cabinet, the method comprising the steps of:

acquiring a newly generated operation work order aiming at the power control cabinet from a management center, and extracting a work order number in the operation work order;

Generating a verification code according to the work order number;

acquiring first image information containing the surrounding environment of the power control cabinet in real time through a first image acquisition device arranged around the power control cabinet, and analyzing whether personnel activities exist in the first image information by utilizing an image recognition algorithm;

if personnel activities exist, activating a second image acquisition device arranged on the electric control cabinet;

acquiring second image information of a target person positioned beside the electric control cabinet in real time through the second image acquisition device;

analyzing whether the dressing configuration of the target person meets the safety requirement or not by combining the first image information and the second image information;

if the safety requirements are met, activating a cabinet door control device arranged outside the electric control cabinet body, and acquiring biological information of the target personnel and a request code input by the target personnel through the cabinet door control device;

performing personnel verification by combining the second image information and the biological information, and comparing the verification code with the request code;

and when the personnel passes the verification and the verification code is consistent with the request code, opening a cabinet door of the electric control cabinet through the cabinet door control device.

Optionally, after the cabinet door of the power control cabinet is opened by the cabinet door control device, the method further comprises the following steps:

activating a control cabinet operation device arranged in the power control cabinet;

identifying the personnel identity of the target personnel through the second image information and/or the biological information, and confirming the personnel operation authority of the target personnel based on the personnel identity;

opening all target operation functions in the control cabinet operation device, wherein the function use authority of the target operation functions is smaller than or equal to the personnel operation authority;

extracting operation content information in the operation work order, and determining operation requirement permission by analyzing the operation content information;

and if the personnel operation authority is smaller than the operation requirement authority, generating a temporary authority application form by combining the personnel identity and the operation content information, and feeding back the temporary authority application form to the management center.

Optionally, the extracting the operation content information in the operation worksheet, and determining the operation requirement authority by analyzing the operation content information includes the following steps:

extracting operation content information from the operation worksheet by adopting a keyword extraction technology;

Performing word segmentation processing on the operation content information to obtain a plurality of content word segments;

performing content matching on a plurality of content fragments and a preset operation content list in a fragment vector matching mode, and marking the content fragments successfully matched with the content as target content fragments;

acquiring operation authorities associated with all target content segmentation words in the operation content list;

and taking the highest authority in all the operation authorities as an operation requirement authority.

Optionally, the biological information is any one or more of fingerprint information, palmprint information, voiceprint information, retina information and iris information.

Optionally, the analyzing whether the dressing configuration of the target person meets the safety requirement by combining the first image information and the second image information includes the following steps:

constructing a safety helmet recognition model based on a convolutional neural network, recognizing the second image information through the safety helmet recognition model, and judging whether the target person in the second image information wears a safety helmet or not;

if the identification result shows that the target person does not wear the safety helmet, analyzing to obtain that the dressing configuration of the target person does not meet the safety requirement;

If the identification result shows that the target personnel wear the safety helmet, extracting target dressing characteristics of the target personnel from the first image information;

performing feature matching on the target dressing feature and a preset standard dressing feature;

if the feature matching is successful, analyzing to obtain that the dressing configuration of the target personnel meets the safety requirement;

and if the feature matching fails, analyzing to obtain that the dressing configuration of the target personnel does not meet the safety requirement.

Optionally, the construction of the safety helmet recognition model based on the convolutional neural network, the recognition of the second image information by the safety helmet recognition model, and the judgment of whether the target person wearing the safety helmet in the second image information includes the following steps:

constructing an initial recognition model based on a convolutional neural network and acquiring a training image data set;

dividing the image data set into a training set and a testing set, and combining the training set and the testing set to train the initial recognition model into a safety helmet recognition model;

inputting the second image information into the safety helmet recognition model, and extracting the face characteristics of the target person by utilizing a convolution layer of the safety helmet recognition model;

Selecting a region of interest in the second image information based on the face features and using a target detection algorithm;

classifying the region of interest through the safety helmet recognition model, and judging whether the target person in the second image information wears the safety helmet or not according to the classification result.

Optionally, the extracting the target dressing feature of the target person from the first image information includes the following steps:

acquiring standard image information of the standard dressing around the power control cabinet, wherein the standard image information is acquired in advance by the first image acquisition device and stored in a preset image database, and the standard image information also comprises a plurality of positioning identification objects arranged around the power control cabinet;

identifying the positioning identification object in the first image information and a target person area where the target person is located by adopting a preset image identification model;

screening a first target image frame of the positioning identification object most conforming to the position characteristics from the first image information based on the position characteristics of all the positioning identification objects in the standard image information;

Preprocessing the first target image frame according to the picture attribute of the standard image information;

selecting a local monochromatic area in the standard image information by using a clustering algorithm frame, and selecting a target area with the same position in the first target image frame according to the area position of the local monochromatic area in the standard image information;

extracting standard region image features of the local monochromatic region, and processing the first target image frame according to the standard region image features so that the target region image features of the target region are the same as the standard region image features;

and extracting target dressing characteristics of the target personnel from the target personnel area through a characteristic extraction algorithm.

Optionally, the step of acquiring, in real time, first image information including the surrounding environment of the power control cabinet through a first image acquisition device disposed around the power control cabinet, and analyzing whether personnel activities exist in the first image information by using an image recognition algorithm includes the following steps:

when the operation work order is acquired, switching the working state of a first image acquisition device arranged around the electric control cabinet into a work preparation state, and adjusting the direction of the first image acquisition device until the electric control cabinet is in a central area of an acquired image in the work preparation state;

An image recognition algorithm based on character recognition is operated, and whether a region of interest appears in the first image information is judged;

if the region of interest does not appear in the first image information, analyzing to obtain that no personnel activities exist in the first image information;

and if the region of interest appears in the first image information, analyzing to obtain that personnel activities exist in the first image information.

Optionally, the method further comprises the steps of:

after detecting that the cabinet door of the electric control cabinet is closed through the cabinet door control device, locking the cabinet door of the electric control cabinet through the cabinet door control device;

switching the working state of the first image acquisition device into an alert state after a preset waiting time, and starting an infrared heat induction mode by the first image acquisition device in the alert state, wherein the first image acquisition device reciprocally rotates within a preset rotation angle range and acquires third image information comprising the electric control cabinet in real time;

maintaining the image recognition algorithm, and judging whether an interesting region appears in the third image information;

if the region of interest appears in the third image information, the first image acquisition device is controlled to stop rotating, and personnel behavior information of common personnel in the region of interest in the third image information is acquired, wherein the personnel behavior information comprises personnel position, personnel longitudinal movement speed, personnel transverse movement speed and personnel movement direction of the common personnel;

Combining the personnel behavior information including personnel position, the personnel longitudinal movement speed, the personnel transverse movement speed and the personnel movement direction, and analyzing by a Kalman filtering method to obtain the predicted transverse speed, the predicted longitudinal speed and the predicted movement direction of the common personnel;

judging whether the common person has a movement intention of continuing to approach the power control cabinet or not according to the predicted transverse speed, the predicted longitudinal speed and the predicted movement direction;

if the common person has a movement intention of continuing to approach the electric control cabinet, sending out preset warning voice by utilizing sound playing equipment attached to the first image acquisition device, and generating safety warning information;

and uploading the safety warning information to the management center.

In a second aspect, the present invention further provides a security monitoring system of a power control cabinet, the system comprising:

the first image acquisition device is arranged around the power control cabinet and is used for acquiring first image information containing the surrounding environment of the power control cabinet;

the second image acquisition device is arranged on the electric control cabinet and is used for acquiring second image information of a target person beside the electric control cabinet;

The cabinet door control device is arranged outside the cabinet body of the electric control cabinet and is used for collecting biological information of target personnel and request codes input by the target personnel and controlling the cabinet door of the electric control cabinet to be opened and closed;

the control cabinet security management device is arranged inside the cabinet body of the electric control cabinet and is in communication connection with the first image acquisition device, the second image acquisition device and the cabinet door control device, the control cabinet security management device comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, and the security monitoring method of the electric control cabinet in the first aspect is realized when the processor executes the computer program.

The beneficial effects of the invention are as follows:

through the technical means of image recognition, personnel verification and code comparison, only the verified personnel can access and operate the electric control cabinet, compared with the prior art, the potential safety hazard that the cabinet door is easy to be opened by anyone is eliminated, and the safety of the electric control cabinet is greatly improved. In addition, through acquiring the first image information in real time and analyzing personnel activities, personnel activities around the electric power control cabinet can be found and recorded in time, and the safe and stable operation of equipment is ensured. Through the operation worksheet and verification code provided by the management center, the operation of the power control cabinet can be recorded and traced every time, and subsequent audit and analysis are facilitated.

Drawings

Fig. 1 is a device configuration diagram of a security monitoring system of a power control cabinet according to one embodiment of the present application.

Fig. 2 is a device configuration diagram of the power control cabinet in one embodiment of the present application.

Fig. 3 is a schematic flow chart of a security monitoring method of a power control cabinet in one embodiment of the application.

Reference numerals illustrate:

1. an electric control cabinet; 2. a first image acquisition device; 3. positioning and identifying an object; 4. a second image acquisition device; 5. a cabinet door control device; 6. and a control cabinet operating device.

Detailed Description

Technical solutions in the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application are within the scope of the protection of the present application.

The terms first, second and the like in the description and in the claims, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged, as appropriate, such that embodiments of the present application may be implemented in sequences other than those illustrated or described herein, and that the objects identified by "first," "second," etc. are generally of a type and not limited to the number of objects, e.g., the first object may be one or more. Furthermore, in the description and claims, "and/or" means at least one of the connected objects, and the character "/", generally means that the associated object is an "or" relationship.

The application discloses a security protection monitoring system of power control cabinet. Referring to fig. 1 and 2, the system includes:

the cabinet door control device is arranged outside the cabinet body of the electric control cabinet and is used for collecting biological information of a target person, a request code input by the target person and controlling the cabinet door of the electric control cabinet to open and close;

the control cabinet security management device is arranged in the cabinet body of the electric control cabinet and is in communication connection with the first image acquisition device, the second image acquisition device and the cabinet door control device, and comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and the processor realizes the security monitoring method of the electric control cabinet when executing the computer program. The control cabinet security management device is in wireless communication connection with a management center of the power system.

The processor may be a Central Processing Unit (CPU), or of course, according to actual use, other general purpose processors, digital Signal Processors (DSP), application Specific Integrated Circuits (ASIC), ready-made programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., and the general purpose processor may be a microprocessor or any conventional processor, etc., which is not limited in this application.

The memory may be an internal storage unit of the computer device, for example, a hard disk or a memory of the computer device, or an external storage device of the computer device, for example, a plug-in hard disk, a Smart Memory Card (SMC), a secure digital card (SD), or a flash memory card (FC) provided on the computer device, or the like, and may be a combination of the internal storage unit of the computer device and the external storage device, where the memory is used to store a computer program and other programs and data required by the computer device, and the memory may also be used to temporarily store data that has been output or is to be output, which is not limited in the present application.

The application also discloses a security monitoring method of the power control cabinet. Fig. 3 is a flow chart of a security monitoring method of the power control cabinet in one embodiment. It should be understood that, although the steps in the flowchart of fig. 3 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in fig. 3 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or other steps. As shown in fig. 3, the security monitoring method of the electric control cabinet specifically includes the following steps:

S101, acquiring a newly generated operation work order aiming at the power control cabinet from a management center, and extracting the work order number in the operation work order.

Among them, a power control cabinet is a device for controlling and protecting a power system, and generally includes a switch, a protection device, a metering device, and the like. All power control cabinets in the power system are uniformly managed through the management center, the power control cabinets can upload real-time data information to the management center, and the management center can issue control instructions, distribute operation worksheets and other operations to all power control cabinets. The control cabinet security management device inside the electric power control cabinet can acquire the latest operation work order information in real time through a communication interface between the control cabinet security management device and a management center. The operator list refers to a file for recording and guiding the operation of the power control cabinet, and comprises information such as an operation instruction, an operation target, operation contents, a work list number and the like. The work order numbers of all the operation work orders are unique numbers, and text processing technology such as regular expressions, character string matching and the like can be used for extracting the work order numbers from the operation work orders.

S102, generating verification codes according to the work order numbers.

The work order number may be a unique identifier, and the verification code may be generated using a hash algorithm or other encoding means. The authentication code may be a string of numbers or characters for authenticating identity or rights. After generating the verification code, the verification code is sent to a mobile terminal held by an operator corresponding to the operator sheet.

S103, acquiring first image information containing the surrounding environment of the power control cabinet in real time through a first image acquisition device arranged around the power control cabinet, analyzing whether personnel activities exist in the first image information by utilizing an image recognition algorithm, and executing the step S104 if the personnel activities exist.

Wherein if there is no personnel activity, no other steps are performed. The first image capture device may be a camera or other image capture device disposed about the power control cabinet. The image recognition algorithm may use computer vision techniques such as object detection, pedestrian detection, etc. to analyze the first image information for the presence of human activity.

S104, activating a second image acquisition device arranged on the power control cabinet.

The second image acquisition device can be a camera or other image acquisition equipment arranged on the power control cabinet. And after the personnel activities are detected in the first image information, activating a second image acquisition device to start to acquire the second image information of the target personnel in real time.

S105, acquiring second image information of a target person positioned beside the power control cabinet in real time through a second image acquisition device.

S106, analyzing whether the dressing configuration of the target personnel meets the safety requirement by combining the first image information and the second image information, and executing step S107 if the dressing configuration meets the safety requirement.

Wherein the first image information and the second image information are analyzed and processed using computer vision techniques such as image segmentation, feature extraction, classifier, etc. And judging whether the safety requirements are met by comparing the dressing configuration of the target personnel with the safety requirements according to the preset safety requirements. If the operation incomplete information does not meet the safety requirement, a cabinet door control device arranged outside the electric control cabinet body is not activated, operation incomplete information is generated, and the operation incomplete information is uploaded to a management center.

S107, activating a cabinet door control device arranged outside the electric control cabinet body, and acquiring biological information of a target person and a request code input by the target person through the cabinet door control device.

Wherein, after meeting the safety requirement, activate the cabinet door controlling means that sets up in the outside of electric control cabinet body. Biological information of the target person, such as fingerprints, facial recognition and the like, is acquired through the cabinet door control device, and a request code input by the target person is acquired. The target personnel can acquire the verification code in advance by checking the carried mobile terminal, and the target personnel can acquire the request code input by the target personnel according to the verification code and perform code input through a display screen or an input keyboard on the cabinet door control device.

S108, carrying out personnel verification by combining the second image information and the biological information, and comparing the verification code with the request code.

The biological information identification technology is utilized to identify biological characteristics in the biological information, such as fingerprint identification, voiceprint identification and the like, and corresponding personnel identities are retrieved from a preset personnel information database according to the identified biological characteristics. Extracting face features from the second image information by utilizing a computer vision technology, retrieving corresponding personnel identities from a personnel information database according to the extracted face features, comparing whether the personnel identities retrieved in the two modes are consistent or not to finish personnel verification, and if so, passing the personnel verification; otherwise, the verification is not passed. Comparing the verification code with the request code, and judging whether the verification code is consistent with the request code or not so as to ensure the validity of verification.

S109, when the personnel passes the verification and the verification code is consistent with the request code, opening a cabinet door of the electric control cabinet through the cabinet door control device.

When the personnel passes the verification and the verification code is consistent with the request code, a door opening instruction is sent through the door control device, and the door of the electric control cabinet is opened. Therefore, only the verified target personnel can open the power control cabinet to perform operation control.

In one embodiment, the method further comprises the following steps after step S109:

extracting operation content information in an operation work order, and determining operation requirement authority by analyzing the operation content information;

if the personnel operation authority is smaller than the operation requirement authority, a temporary authority application form is generated by combining the personnel identity and the operation content information, and the temporary authority application form is fed back to the management center.

In this embodiment, after the cabinet door is opened, the control cabinet operation device provided inside the power control cabinet is activated, that is, the control cabinet operation device is switched from the sleep state to the operation state. In the step of identifying the personnel identity of the target personnel through the second image information and/or the biological information and confirming the personnel operation authority of the target personnel based on the personnel identity, a preset personnel information data set is required to be extracted from a preset personnel information database, wherein the personnel information data set contains personnel detailed information of all operators meeting the operation standard of the power control cabinet in advance, and the personnel detailed information specifically comprises personnel names, personnel working ages, personnel positions, personnel authorities, personnel face features, fingerprint information, voiceprint information and other biological information of the personnel. In the personnel information data set, according to the information type of the personnel detailed information, a plurality of information index lists of the same type of information are respectively constructed.

The biological information identification technology is utilized to identify biological characteristics in the biological information, the second image information can be utilized to extract human face characteristics from the second image information by utilizing the computer vision technology, the human face characteristics and/or the biological characteristics are respectively subjected to information retrieval in a corresponding information index list, when any one or two of the same characteristics are retrieved in the information index list, the retrieval is stopped, the personnel identity information corresponding to the retrieved characteristics is marked, and personnel permission in the marked personnel identity information is inquired, so that personnel operation permission of target personnel is obtained.

According to the acquired personnel operation authority, opening the authority of all the target operation functions of which the function use authority is smaller than or equal to the personnel operation authority in the control cabinet operation device, so that the target personnel can use all the target operation functions with the opened authority, wherein the target operation functions comprise related operations of power systems such as regional power-off, regional power-on and the like.

Finally, extracting the operation content information in the operation work order, and determining the operation requirement authority by analyzing the operation content information; if the personnel operation authority is smaller than the operation requirement authority, a temporary authority application form is generated by combining personnel identity and operation content information, the temporary authority application form is fed back to the management center, and if a temporary authority authorization instruction sent by the management center is received at the moment, the current operation of the target personnel can operate all target operation functions related to the operation content information; if the personnel operation authority is greater than or equal to the operation requirement authority, no step is executed.

In this embodiment, the step of extracting the operation content information in the operation worksheet and determining the operation requirement authority by analyzing the operation content information specifically includes the steps of:

extracting operation content information from an operation work order by adopting a keyword extraction technology;

performing content matching on a plurality of content participles and a preset operation content list in a participle vector matching mode, and marking the content participles successfully matched with the content as target content participles;

acquiring operation authorities associated with all target content segmentation words in an operation content list;

In this embodiment, keywords related to the operation may be extracted from the operator sheet text using natural language processing techniques, such as keyword extraction algorithms. The keyword extraction algorithm can extract keywords based on word frequency, TF-IDF and other methods. And performing word segmentation processing on the operation content information by using a word segmentation technology, and segmenting the text into a plurality of words or phrases. The word segmentation technique may use rule-based methods or statistical-based methods, such as maximum matching methods, hidden markov models, etc. And performing word segmentation processing on the preset operation content list to obtain a word segmentation list of each operation content. And matching the plurality of content word fragments with a word fragment list of the operation content by using a word fragment vector matching technology such as cosine similarity and the like, and calculating a similarity score. If the similarity score exceeds a set threshold, the content word is marked as a target content word.

And according to the target content word segmentation, the operation authority associated with the target content word segmentation is found in the operation content list. The list of operation contents may be a predefined list, each operation content being associated with a corresponding operation right. Traversing all the associated operation authorities and finding out the highest authority. The highest rights may be determined according to rules for the setting of rights, such as a larger number indicating a higher right, or using other rules.

For example, assume that the operation content information in the operation work order is "open cabinet door and shut off power". First, keywords "open cabinet door" and "cut off power" are extracted from a work order using a keyword extraction technique. Then, the keywords are subjected to word segmentation processing to obtain word segmentation of opening, cabinet door, cutting off and power supply. And matching the segmented words with a preset operation content list by using a segmented word vector matching mode, and marking the segmented words as target content segmented words if the similarity score exceeds a threshold value. It is assumed that the "opening cabinet door" is successfully matched with the "opening door" in the operation content list, and the "cutting power supply" is successfully matched with the "cutting power supply" in the operation content list. Then, the operation authority associated with the target content word is found according to the target content word, and the operation authority associated with the door opening is assumed to be an operation cabinet door, and the operation authority associated with the power cutting-off is assumed to be a power cutting-off. And finally, taking the highest authority of the two operation authorities as an operation requirement authority, namely a power-off authority.

In the present embodiment, the biological information is any one or more of fingerprint information, palm print information, voiceprint information, retina information, and iris information.

In one embodiment, the step S106 specifically includes the following steps:

constructing a safety helmet recognition model based on the convolutional neural network, recognizing the second image information through the safety helmet recognition model, and judging whether a target person in the second image information wears the safety helmet or not;

if the identification result is that the target person does not wear the safety helmet, analyzing to obtain that the dressing configuration of the target person does not meet the safety requirement;

if the feature matching fails, analyzing to obtain that the dressing configuration of the target personnel does not meet the safety requirement.

In this embodiment, a helmet recognition model is constructed using deep learning techniques, such as Convolutional Neural Networks (CNNs). The training model uses a annotated image dataset containing image samples of a helmet worn and a helmet not worn. And identifying the second image information by using the trained model, and judging whether the target person wears the safety helmet. If the safety helmet recognition model judges that the target personnel does not wear the safety helmet, the result can be used as a basis for analyzing the dressing configuration of the target personnel. Analysis may include checking the wear of other safety equipment, such as safety shoes, safety glasses, etc., to determine whether the target person's wear meets safety requirements.

Target dressing features of the target person are extracted from the first image information using computer vision techniques such as human body posture estimation, target detection, and the like. Target apparel characteristics may include garment color, garment style, shoe type, and the like. The predetermined standard dressing characteristics may be predefined dressing requirements, such as clothing of a specific color, shoes of a specific style, etc. And comparing the target dressing characteristics with preset standard dressing characteristics, and judging whether the dressing of the target personnel meets the preset standard dressing. If the target dressing characteristics of the target personnel are successfully matched with the preset standard dressing characteristics, the dressing configuration of the target personnel can be judged to meet the safety requirements. If the target dressing characteristics of the target personnel are failed to be matched with the preset standard dressing characteristics, the dressing configuration of the target personnel can be judged to be not in accordance with the safety requirements.

For example, it is assumed that the second image information is recognized by the helmet recognition model, and it is determined that the target person does not wear the helmet. Then, target dressing characteristics of the target person, such as red clothing and sports shoes, are extracted from the first image information. Next, the target dressing characteristics were feature-matched with the preset standard dressing characteristics, and it was found that the preset standard dressing requirements were to wear blue clothes and safety shoes. Because the target dressing characteristics of the target personnel are not matched with the preset standard dressing characteristics, analysis shows that the dressing configuration of the target personnel does not meet the safety requirements. In this case, corresponding measures can be taken.

In one embodiment, a safety helmet recognition model is constructed based on a convolutional neural network, the second image information is recognized through the safety helmet recognition model, and the step of judging whether the target person in the second image information wears the safety helmet specifically comprises the following steps:

dividing the image data set into a training set and a testing set, and training the initial recognition model into a safety helmet recognition model by combining the training set and the testing set;

inputting the second image information into a safety helmet recognition model, and extracting face features of a target person by using a convolution layer of the safety helmet recognition model;

classifying the region of interest through the safety helmet recognition model, and judging whether the target person in the second image information wears the safety helmet according to the classification result.

In this embodiment, an initial image recognition model is constructed using deep learning techniques, such as Convolutional Neural Networks (CNNs). The collection of image datasets, including both worn and unworn helmets, may be obtained by live shooting or from public datasets. The collected image dataset is divided into a training set and a test set, typically using 80% of the data as the training set and 20% of the data as the test set. The training set is used for training the initial recognition model, and the weight and bias of the model are updated through a back propagation algorithm. And evaluating the performance of the trained model by using the test set, and calculating indexes such as accuracy, recall rate and the like.

And inputting the second image information into a trained safety helmet recognition model, and obtaining the output of the model through a forward propagation algorithm. Face features of the target person are extracted from the convolution layer, and the features can capture detailed information of the face. The position of the target person is detected in the second image information using a target detection algorithm, such as a convolutional neural network based target detection algorithm (e.g., YOLO, fast R-CNN, etc.). And marking the interested area of the target person in a frame selection mode according to the position and the size of the face features. The region of interest is input into the helmet recognition model, and the output of the model is obtained through a forward propagation algorithm. And judging whether the target personnel wear the safety helmet according to the output result of the model. If the output result is the type of wearing the safety helmet, judging that the target personnel wear the safety helmet; and if the output result is the type of the unworn safety helmet, judging that the target personnel does not wear the safety helmet.

In one embodiment, the step of extracting the target dressing feature of the target person from the first image information specifically includes the steps of:

acquiring standard image information which is arranged around the power control cabinet in a standard mode, acquiring the standard image information in advance by a first image acquisition device, and storing the standard image information in a preset image database, wherein the standard image information also comprises a plurality of positioning identification objects arranged around the power control cabinet;

Identifying a positioning identification object in the first image information and a target person area where a target person is located by adopting a preset image identification model;

screening a first target image frame of which the positioning identification object most accords with the position characteristic from the first image information based on the position characteristics of all the positioning identification objects in the standard image information;

preprocessing a first target image frame according to the picture attribute of the standard image information;

selecting a local monochromatic area in the standard image information by using a clustering algorithm frame, and selecting a target area with the same position in a first target image frame according to the area position of the local monochromatic area in the standard image information;

extracting standard region image features of a local monochromatic region, and processing a first target image frame according to the standard region image features so that the target region image features of the target region are the same as the standard region image features;

In this embodiment, as shown in fig. 1, a plurality of positioning recognition objects may be disposed around the power control cabinet, and the positioning recognition objects may be a sign or a marker of a specific shape. And shooting the standard dressing around the power control cabinet by using the first image acquisition device, and shooting, positioning and identifying the object. The standard image information is stored in a preset image database. And processing the first image information by using a preset image recognition model, and recognizing the positioning recognition object and the target personnel area in the first image information. Target detection and target segmentation may be performed using deep learning techniques, such as Convolutional Neural Networks (CNNs), to obtain the region in which the target person is located.

The most similar positioning recognition object in the position features is found by comparing the positioning recognition object in the first image information with the positioning recognition object in the standard image information. The first target image frame containing the positioning recognition object is selected as a target image for subsequent processing. The first target image frame is preprocessed according to picture properties of the standard image information, such as brightness, contrast, etc., so as to have similar picture properties as the standard image information. And processing the standard image information by using a clustering algorithm, such as a K-means clustering algorithm, and dividing and identifying the local monochromatic area in the image. And according to the region position of the local monochromatic region in the standard image information, finding a target region at the same position in the first target image frame, and performing frame selection.

Image features of the local monochromatic area are extracted from the standard image information, and feature extraction algorithms such as Local Binary Pattern (LBP), direction gradient Histogram (HOG), etc. may be used. The extracted standard region image features are applied to the target region of the first target image frame such that the image features of the target region are the same as the standard region image features. Target dressing features of the target person are extracted from the target person region using a feature extraction algorithm, such as a Convolutional Neural Network (CNN) in deep learning. These characteristics may include color, texture, shape, etc. characteristics for determining whether the target person's clothing meets the preset standard clothing.

In one embodiment, the step S103 specifically includes the following steps:

when an operation work order is acquired, switching the working state of a first image acquisition device arranged around the electric control cabinet into a work preparation state, and adjusting the direction of the first image acquisition device until the electric control cabinet is positioned in the central area of the acquired image in the work preparation state;

In the present embodiment, when the operation order is received, the operation state of the first image pickup device is switched to the work preparation state. The direction of the first image acquisition device is adjusted to be aligned with the power control cabinet, so that the power control cabinet is located in the central area of an acquired image. The first image information is processed using an image recognition algorithm based on person recognition to determine whether a region of interest, such as a person's active region, is present therein. The images may be analyzed and identified using deep learning techniques, such as Convolutional Neural Networks (CNNs) or object detection algorithms. If the image recognition algorithm based on person recognition judges that the region of interest does not appear in the first image information, namely, the moving region of the person is not detected, the conclusion can be drawn that no person moves in the first image information. If the image recognition algorithm based on person recognition judges that the region of interest appears in the first image information, namely the moving region of the person is detected, the condition that the person moves in the first image information can be concluded.

In one embodiment, the security monitoring method of the power control cabinet further comprises the following steps:

maintaining the image recognition algorithm, and judging whether a region of interest appears in the third image information;

if the region of interest appears in the third image information, the first image acquisition device is controlled to stop rotating, and personnel behavior information of common personnel in the region of interest in the third image information is acquired, wherein the personnel behavior information comprises the personnel position, the personnel longitudinal movement speed, the personnel transverse movement speed and the personnel movement direction of the common personnel;

the method comprises the steps of combining personnel behavior information including personnel position, personnel longitudinal movement speed, personnel transverse movement speed and personnel movement direction, and analyzing by a Kalman filtering method to obtain predicted transverse speed, predicted longitudinal speed and predicted movement direction of common personnel;

Judging whether the common person has a movement intention of continuing to approach the power control cabinet or not by combining the predicted transverse speed, the predicted longitudinal speed and the predicted movement direction;

if the common person has the movement intention of continuing to approach the electric control cabinet, sending out preset warning voice by utilizing sound playing equipment attached to the first image acquisition device, and generating safety warning information;

and uploading the safety warning information to a management center.

In this embodiment, a sensor is provided in the cabinet door control device for detecting the cabinet door state of the power control cabinet. After the sensor detects that the cabinet door is closed, locking operation is performed through the cabinet door control device, so that the cabinet door of the electric control cabinet cannot be opened. And after the preset waiting time, switching the working state of the first image acquisition device into an alert state. In the alert state, the first image acquisition device can reciprocally rotate within a preset rotation angle range and acquire third image information comprising the power control cabinet in real time. An image recognition algorithm is run in the third image information to determine whether a region of interest, such as a person's active region, is present therein. The images may be analyzed and identified using deep learning techniques, such as Convolutional Neural Networks (CNNs) or object detection algorithms. The first image acquisition device starts an infrared heat induction mode, so that safety warning can be normally realized at night.

If the image recognition algorithm judges that the region of interest appears in the third image information, namely, the moving region of the person is detected, the first image acquisition device can be controlled to stop rotating, and the person behavior information of the common person in the region of interest is extracted. The personnel behavior information can be obtained through a target tracking algorithm and a motion analysis algorithm, and comprises information such as the position, the longitudinal moving speed, the transverse moving speed and the motion direction of personnel. And predicting and estimating the transverse speed, the longitudinal speed and the movement direction of the common personnel by using a Kalman filtering method and combining personnel behavior information. The Kalman filtering method is a commonly used state estimation algorithm, and state prediction and updating can be performed through an observation value and a system model. And judging whether the common person has the movement intention of continuing to approach the power control cabinet or not by using the predicted transverse speed, longitudinal speed and movement direction. A certain threshold or rule may be set, such as if the predicted lateral speed exceeds a certain value, or the predicted direction of movement is directed to the power control cabinet, determining that there is a continued approaching movement intention of the average person.

If the movement intention of the ordinary person, which is continuously close to the power control cabinet, is judged, a preset warning voice can be sent out through sound playing equipment attached to the first image acquisition device so as to remind the ordinary person of leaving a dangerous area. Meanwhile, safety warning information including time, place, related personnel information and the like of warning can be generated for subsequent recording and management. And uploading the generated safety warning information to a management center through a network or other communication modes. So that the management center takes corresponding measures in time, such as notifying relevant personnel or performing further processing.

Those of ordinary skill in the art will appreciate that: the discussion of any of the embodiments above is merely exemplary and is not intended to imply that the scope of the present application is limited to such examples; the technical features of the above embodiments or in the different embodiments may also be combined under the idea of the present application, the steps may be implemented in any order, and there are many other variations of the different aspects of one or more embodiments in the present application as above, which are not provided in details for the sake of brevity.

One or more embodiments herein are intended to embrace all such alternatives, modifications and variations that fall within the broad scope of the present application. Any omissions, modifications, equivalents, improvements, and the like, which are within the spirit and principles of the one or more embodiments in the present application, are therefore intended to be included within the scope of the present application.

Claims

1. The security monitoring method of the power control cabinet is characterized by comprising the following steps of:

generating a verification code according to the work order number;

when the personnel passes the verification and the verification code is consistent with the request code, opening a cabinet door of the electric control cabinet through the cabinet door control device;

taking the highest authority of all the operation authorities as an operation requirement authority;

2. The security monitoring method of an electric control cabinet according to claim 1, wherein the biological information is any one or more of fingerprint information, palm print information, voiceprint information, retina information, and iris information.

3. The security monitoring method of a power control cabinet according to claim 1, wherein the analyzing whether the dressing configuration of the target person meets the security requirement by combining the first image information and the second image information comprises the steps of:

4. The security monitoring method of the power control cabinet according to claim 3, wherein the construction of a safety helmet recognition model based on a convolutional neural network, the recognition of the second image information by the safety helmet recognition model, and the judgment of whether the target person in the second image information wears a safety helmet, comprises the following steps:

5. The security monitoring method of the power control cabinet according to claim 3, wherein the extracting the target dressing characteristic of the target person from the first image information comprises the steps of:

6. The security monitoring method of a power control cabinet according to claim 1, wherein the step of acquiring, in real time, first image information including the surrounding environment of the power control cabinet through a first image acquisition device arranged around the power control cabinet, and analyzing whether personnel activities exist in the first image information by using an image recognition algorithm includes the steps of:

7. The method for security monitoring of a power control cabinet of claim 6, further comprising the steps of:

And uploading the safety warning information to the management center.

8. A security monitoring system for an electrical control cabinet, the system comprising:

the control cabinet security management device is arranged in the cabinet body of the electric control cabinet and is in communication connection with the first image acquisition device, the second image acquisition device and the cabinet door control device, the control cabinet security management device comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and the processor realizes the security monitoring method of the electric control cabinet according to any one of claims 1 to 7 when executing the computer program.