A big data security monitoring system

Abstract

The application relates to the field of regional security management, and particularly discloses a big data security monitoring system, which comprises an image acquisition module, an image application module, a supply data detection module and an emergency end control module. The image acquisition module is used for inserting an information reader into a camera network to acquire scene images under dynamic frequency by the information reader. The image application module is used for carrying out numerical conversion on the scene images, and determining target distribution in a region according to the converted values. The supply data detection module is used for calculating a demand value according to the target distribution, randomly detecting supply data according to the demand value, and outputting a risk value according to a detection result. The emergency end control module is used for generating a control instruction pointing to an emergency end, and sending the control instruction to the corresponding emergency end. The application acquires images through a monitoring system, carries out contour recognition on the images, further determines energy supply parameters required by the images, identifies actual energy supply parameters according to the energy supply parameters, judges the risk value, accurately and timely generates reminding information, and reports the reminding information to senior management personnel.

Description

Technical Field

[0001] This invention relates to the field of regional security management, specifically a big data security monitoring system. Background Technology

[0002] The need for surveillance in production and living areas has always existed. As the cost of random electronic devices continues to decrease and cameras become more widespread, the need for surveillance can be met.

[0003] However, existing camera-based area monitoring processes are mostly completed by monitoring personnel, which have a low level of intelligence and require dedicated personnel. In order to improve the monitoring effect, the monitoring personnel themselves must be very familiar with the production and living areas; otherwise, it is difficult to quickly detect abnormalities. This makes the labor cost of monitoring personnel extremely high. How to intelligently monitor areas and optimize the cost of monitoring personnel is the technical problem that this invention aims to solve. Summary of the Invention

[0004] The purpose of this invention is to provide a big data security monitoring system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A big data security monitoring system, the system comprising:

[0007] The image acquisition module is used to insert information readers into the camera network, and the information readers acquire scene images at dynamic frequencies.

[0008] The image application module is used to perform numerical conversion on scene images, determine the distribution of targets in the region based on the converted values, calculate outliers based on the target distribution, and update the dynamic frequency synchronously based on the outliers; the targets in the target distribution include animals and still objects.

[0009] The supply data detection module is used to calculate the demand value based on the target distribution, perform random detection on the supply data based on the demand value, and output the risk value based on the detection results.

[0010] An emergency control module is used to generate control commands for the emergency terminal based on abnormal values ​​and risk values ​​as reference parameters, and send them to the corresponding emergency terminal; wherein, the emergency terminal includes at least an alarm and a safety door.

[0011] As a further aspect of the present invention: the image acquisition module includes:

[0012] The camera parameter acquisition unit is used to query the installed camera terminals with image acquisition functions and obtain the camera's position and camera parameters; the camera parameters include camera direction, camera wide angle, and camera resolution;

[0013] The area calculation and assignment unit is used to calculate the camera area of ​​each camera end based on the camera direction and camera width, and to determine the additional value of the camera area based on the camera sharpness;

[0014] A region overlay unit is used to overlay all camera regions and their additional values ​​to obtain a camera range map; the camera range map is used to characterize the information exposure of each location in the region;

[0015] An execution unit is used to determine the number and location of information readers based on the camera range map, and the information readers acquire scene images at a dynamic frequency.

[0016] As a further aspect of the present invention: the acquisition execution unit includes:

[0017] The initial position determination subunit is used to determine the initial installation position in the camera range map based on the boundary dimensions of the camera range map and the pre-set fixed-length grid.

[0018] A circle creation subunit is used to create a circle based on the initial installation position, wherein the radius of the circle is related to the value of each pixel point within the circle;

[0019] The circle removal subunit is used to remove circles from the camera range map and repeatedly perform the circle creation process in the remaining area until the union of the circles contains the camera range map;

[0020] The circular application sub-unit is used to determine the number of all circles as the installation quantity and the center position of the circle as the installation position.

[0021] The rule for calculating the radius of a circle is as follows: In the formula, R is the radius, N and M are the dependent variables of radius R, N and M are the scales in two perpendicular directions in a circle with radius R, and L is a preset threshold.

[0022] As a further aspect of the present invention: the image application module includes:

[0023] The numerical conversion unit is used to traverse the pixels in the scene image and convert the value of each pixel into an LBP value.

[0024] The contour recognition unit is used to input the converted image into the trained convolutional recognition model and label animal and still life contours.

[0025] The first verification unit is used to acquire access control monitoring data, verify the animal outline based on the access control monitoring data, and obtain a first verification value;

[0026] The second verification unit is used to obtain still life registration information, verify the still life outline based on the still life registration information, and obtain a second verification value.

[0027] The frequency update unit is used to determine anomalies based on the first verification value and the second verification value, and to update the dynamic frequency synchronously based on the anomalies.

[0028] As a further aspect of the present invention: the content of converting the value of each pixel into an LBP value includes:

[0029]

[0030] In the formula, LBP(x c ,y c ) is the position (x c ,y c The LBP value at position (p) represents the value of the p-th pixel in the window excluding the center pixel, and I(c) represents the grayscale value of the center pixel; s(x) is the threshold function, which takes the following values:

[0031]

[0032] The window size is 3×3.

[0033] As a further aspect of the present invention: the first verification unit includes:

[0034] The access control information application subunit is used to acquire the entry and exit personnel information collected by the access control terminal, and to determine the theoretical personnel table in the area in real time based on the entry and exit personnel information;

[0035] The object recognition subunit is used to read animal outlines, identify animal outlines, and determine the actual personnel table; wherein, both the theoretical personnel table and the actual personnel table contain personnel tag items and personnel level items;

[0036] The table conversion sub-cell is used to convert the theoretical personnel table and the actual personnel table into theoretical quantity groups and actual quantity groups based on the personnel level item, respectively.

[0037] The difference application subunit is used to subtract the theoretical quantity group and the actual quantity group to obtain the difference quantity group, and to calculate the first verification value based on the difference quantity group.

[0038] As a further aspect of the present invention: the object recognition subunit includes:

[0039] The traversal sub-unit is used to traverse the pixels in the animal outline, calculate the color difference between adjacent pixels, and mark the corresponding pixel when the color difference is less than a preset color difference threshold.

[0040] The statistical subunit is used to count the marked pixels to obtain the feature contour;

[0041] The safety helmet positioning subunit is used to determine the safety helmet outline based on the outline curve and outline position of the feature contour.

[0042] The level determination subunit is used to read the color value of each pixel in the outline of the safety helmet to determine the personnel level.

[0043] As a further aspect of the present invention: the first verification unit further includes:

[0044] The partitioning subunit is used to read the regional map and receive partitioning information containing permission levels input by the user based on the regional map.

[0045] The permission query sub-unit is used to sequentially read the personnel position and personnel level of each person in the actual personnel table, and read the permission level corresponding to the personnel position;

[0046] The correction subunit is used to compare the permission level and the personnel level, calculate the difference rate, and correct the first verification value based on the difference rate.

[0047] As a further aspect of the present invention: the supply data detection module includes:

[0048] The target selection unit is used to randomly select target contours from animal and still life contours, identify the target contours, and determine the required data.

[0049] The demand query unit is used to query demand locations based on the location of the target outline; wherein, the demand locations are included in the registered power supply locations;

[0050] The single-time risk calculation unit is used to obtain energy supply data at demand points, compare the energy supply data with the demand data, and determine the single-time risk value.

[0051] The comprehensive risk statistics unit is used to count the risk value of a single instance within a preset number of times, calculate and output the risk value.

[0052] As a further aspect of the present invention: the emergency control module includes:

[0053] The first comparison unit is used to compare the outlier with a preset outlier threshold and determine the risk threshold based on the comparison result.

[0054] The second comparison unit is used to compare the risk value with the risk threshold and determine the control command of the emergency terminal based on the comparison result.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention acquires images through a monitoring system, performs contour recognition on the images, and then determines the required power supply parameters. The actual power supply parameters are then identified based on the power supply parameters to determine the risk value. By combining the contour recognition process and the risk judgment process, the present invention can accurately and timely generate reminder information and report it to senior management personnel. There is no need to configure senior management personnel to perform real-time monitoring work, which greatly optimizes the labor costs. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0057] Figure 1 This is a block diagram of the composition structure of a big data security monitoring system. Detailed Implementation

[0058] To make the technical problems, solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0059] Figure 1 This invention provides a big data security monitoring system, comprising: a block diagram illustrating the structural composition of such a system. The system 10 includes:

[0060] Image acquisition module 11 is used to insert an information reader into the camera network, and the information reader acquires scene images at dynamic frequencies;

[0061] Image application module 12 is used to perform numerical conversion on scene images, determine the target distribution in the region based on the converted values, calculate outliers based on the target distribution, and update the dynamic frequency synchronously based on the outliers; the targets in the target distribution include animals and still objects;

[0062] The supply data detection module 13 is used to calculate the demand value based on the target distribution, perform random detection on the supply data based on the demand value, and output the risk value based on the detection result.

[0063] The emergency control module 14 is used to generate control commands for the emergency terminal based on abnormal values ​​and risk values ​​as reference parameters, and send them to the corresponding emergency terminal; wherein, the emergency terminal includes at least an alarm and a safety door.

[0064] In one example of the technical solution of this invention, the focus is on scenarios with monitoring needs, such as a production workshop or a living area. These scenarios are equipped with their own camera systems, referred to as camera networks. Under the existing Internet of Things architecture, the cameras in the camera system include not only cameras installed on the top of the area, but also cameras installed on some devices. Information readers are set up according to the location of these cameras. The information readers acquire the images collected by the cameras, and thus obtain the image of the entire area, referred to as the scene image. The relationship between the information reader and the camera is one-to-many. The information reader acts as a relay in the image acquisition process and belongs to a distributed architecture, which can greatly reduce the data processing pressure of the central control terminal.

[0065] After acquiring the scene image, the scene image is identified to determine the distribution of each object in the scene image. For a stable production and living area, the objects in the scene are almost fixed. Based on the distribution of the objects, it can be determined whether there is a possibility of anomalies. Generally, the animals refer to the staff, and the still objects refer to the fixed equipment.

[0066] There are many application solutions for target distribution. The technical solution of this invention provides a risk assessment process based on energy supply conditions. The demand value can be calculated based on the target distribution, for example, how much electricity is needed. The electricity supply process is obtained by corresponding power meters. Based on the demand value and the actual value obtained by the power meters, it can be determined whether there is an incorrect energy supply situation under the current conditions. In production and living areas, many risks are caused by energy supply risks. Controlling the energy supply process can greatly improve safety. Since the energy supply data is continuous, this application uses a random sampling method for detection, and then outputs a value reflecting safety, called the risk value.

[0067] Outliers reflect abnormal situations in the scenario, while risk values ​​reflect the likelihood of power supply errors under abnormal conditions. Combining these two factors allows for control of the emergency response system's operational status. The emergency response system includes at least an alarm, whose function is to alert management. In addition, the emergency response system may also include a safety door, which is normally closed but will be opened when the risk or outlier value is high.

[0068] In a preferred embodiment of the technical solution of the present invention, the image acquisition module includes:

[0069] The camera parameter acquisition unit is used to query the installed camera terminals with image acquisition functions and obtain the camera's position and camera parameters; the camera parameters include camera direction, camera wide angle, and camera resolution;

[0070] The area calculation and assignment unit is used to calculate the camera area of ​​each camera end based on the camera direction and camera width, and to determine the additional value of the camera area based on the camera sharpness;

[0071] A region overlay unit is used to overlay all camera regions and their additional values ​​to obtain a camera range map; the camera range map is used to characterize the information exposure of each location in the region;

[0072] An execution unit is used to determine the number and location of information readers based on the camera range map, and the information readers acquire scene images at a dynamic frequency.

[0073] In one example of the technical solution of this invention, the installed camera terminals are queried. Some of these cameras are installed on the top of the area, some are mounted on equipment, and some are even equipped on the staff. These are collectively referred to as camera terminals with image acquisition functions. Each camera terminal is analyzed to obtain its camera parameters. The camera parameters can determine the monitoring position of each camera terminal in the area, which is called the camera area. Different camera resolutions correspond to different values ​​for the camera area, which are called additional values. The higher the camera resolution, the more fully the information is displayed and the higher the exposure. By statistically analyzing all camera areas and their additional values, a map corresponding to the entire area can be obtained, which is called the camera range map. The value of each point in the camera range map represents the exposure of each position in the area.

[0074] By analyzing the camera range diagram, the number and location of the information readers can be determined. The information readers are devices connected to the camera end and used to read the camera data.

[0075] In a preferred embodiment of the technical solution of the present invention, the acquisition execution unit includes:

[0076] The initial position determination subunit is used to determine the initial installation position in the camera range map based on the boundary dimensions of the camera range map and the pre-set fixed-length grid.

[0077] A circle creation subunit is used to create a circle based on the initial installation position, wherein the radius of the circle is related to the value of each pixel point within the circle;

[0078] The circle removal subunit is used to remove circles from the camera range map and repeatedly perform the circle creation process in the remaining area until the union of the circles contains the camera range map;

[0079] The circular application sub-unit is used to determine the number of all circles as the installation quantity and the center position of the circle as the installation position.

[0080] The rule for calculating the radius of a circle is as follows: In the formula, R is the radius, N and M are the dependent variables of radius R, N and M are the scales in two perpendicular directions in a circle with radius R, and L is a preset threshold.

[0081] In one embodiment of the technical solution of the present invention, the installation process of the information reader is defined. The value of each pixel in the camera range diagram indicates how many cameras monitor each pixel. The larger the value, the higher the information exposure. Correspondingly, the more cameras are around it. At this time, the working range of the information reader will be smaller.

[0082] Regarding the working range mentioned above, it is represented by a circle. Based on the boundary dimensions of the camera range map and the pre-set fixed-length grid, some points can be selected. This process involves inserting a grid into the map. The grid size is input in advance by the staff. Centered on each point, the surrounding values ​​are counted and the sum of the values ​​is calculated. A circle is then created with each point as the center. The radius of the circle is inversely proportional to the sum of the values. The practical significance of this process is that the larger the value, the more cameras are around, and the smaller the working range of the information reader needs to be. Through the above process, multiple circles of different sizes can be obtained.

[0083] By removing these circles from the camera's field of view and using the remaining area as a reference, the above process is repeated to obtain multiple circles of different sizes until all areas are covered by the working range of the information reader, which is the circular area.

[0084] Specifically, the calculation of the radius of a circle is a recursive process. The radius increases continuously, which affects N and M. N and M, in turn, affect the radius, eventually reaching a maximum value.

[0085] As a preferred embodiment of the technical solution of the present invention, the image application module includes:

[0086] The numerical conversion unit is used to traverse the pixels in the scene image and convert the value of each pixel into an LBP value.

[0087] The contour recognition unit is used to input the converted image into the trained convolutional recognition model and label animal and still life contours.

[0088] The first verification unit is used to acquire access control monitoring data, verify the animal outline based on the access control monitoring data, and obtain a first verification value;

[0089] The second verification unit is used to obtain still life registration information, verify the still life outline based on the still life registration information, and obtain a second verification value.

[0090] The frequency update unit is used to determine anomalies based on the first verification value and the second verification value, and to update the dynamic frequency synchronously based on the anomalies.

[0091] In one example of the technical solution of this invention, the working process of the image application module is defined. First, each channel in the scene image is extracted or the scene image is converted into a grayscale image, so that each analysis process is a single-value image analysis. The pixels in the single-value image are traversed and converted into LBP values. The converted LBP values ​​retain the texture features of the image and are clearer. Then, the converted image is input into a trained convolutional recognition model, which can quickly identify animal outlines and still life outlines. Among them, animal outlines mainly correspond to staff members, and still life outlines mainly correspond to equipment. Access control monitoring data can determine which personnel are in the current area, and still life registration data can determine which still lifes should be in the area. Access control monitoring data is the entry and exit information obtained by the access control terminal, and still life registration data is the placement location of the equipment, etc. The access control monitoring data verifies the animal outline to obtain a first verification value, which is used to reflect whether there is any abnormality of the staff. The still life registration data verifies the still life outline to obtain a second verification value, which is used to reflect whether there is any abnormality of the equipment.

[0092] Using the first and second verification values ​​as independent variables, and the relationships (such as summation) pre-entered by the staff, a comprehensive value can be obtained, called the outlier, which is used to reflect the overall anomaly situation of the area. The dynamic frequency is updated by the outlier. Generally, the higher the outlier, the greater the dynamic frequency, and the more area images are obtained.

[0093] Specifically, the process of converting the value of each pixel into an LBP value includes:

[0094]

[0095] In the formula, LBP(x c ,y c ) is the position (x c ,y c The LBP value at position (p) represents the value of the p-th pixel in the window excluding the center pixel, and I(c) represents the grayscale value of the center pixel; s(x) is the threshold function, which takes the following values:

[0096]

[0097] The window size is 3×3.

[0098] It should be noted that, regarding the above values, if the region image is converted to a grayscale image for processing, the above conversion process and application are very simple. However, if each channel in the region image is extracted for individual analysis, then it is necessary to sum the converted values ​​of each channel and calculate the average.

[0099] Furthermore, regarding the convolutional recognition model, it is a neural network model based on convolutional features-contours. The convolutional features are the features of the image in the LBP value space. When constructing samples, image data of each contour can be obtained with the help of big data. The image data is then converted into LBP values ​​to obtain image features represented by LBP values. When the number of image feature-contour samples is large enough, a fast recognition model with extremely high recognition accuracy can be trained, which is called the convolutional recognition model.

[0100] Specifically, the first verification unit includes:

[0101] The access control information application subunit is used to acquire the entry and exit personnel information collected by the access control terminal, and to determine the theoretical personnel table in the area in real time based on the entry and exit personnel information;

[0102] The object recognition subunit is used to read animal outlines, identify animal outlines, and determine the actual personnel table; wherein, both the theoretical personnel table and the actual personnel table contain personnel tag items and personnel level items;

[0103] The table conversion sub-cell is used to convert the theoretical personnel table and the actual personnel table into theoretical quantity groups and actual quantity groups based on the personnel level item, respectively.

[0104] The difference application subunit is used to subtract the theoretical quantity group and the actual quantity group to obtain the difference quantity group, and to calculate the first verification value based on the difference quantity group.

[0105] The above describes the process of generating the first verification value. Access control information is obtained from the access control system to determine the possible personnel within the area, referred to as theoretical personnel, and represented in tabular form. Animal silhouettes are read and identified to obtain the actual personnel in the scene, also represented in tabular form. The two tables are converted into arrays to obtain a theoretical quantity group and an actual quantity group, representing the number of personnel at various levels. The horizontal axis of the quantity group is the level, and the vertical axis is the quantity. The difference between the theoretical and actual quantity groups yields a difference quantity group, which reflects abnormal situations in the number of personnel at various levels. If the area is in a safe state, the difference quantity group is an array of all zeros.

[0106] The process of calculating the first verification value from the difference count is an accumulation process. Each value (or its absolute value) is multiplied by a weight determined by the current value order, and then summed to obtain the first verification value.

[0107] Furthermore, the object recognition subunit includes:

[0108] The traversal sub-unit is used to traverse the pixels in the animal outline, calculate the color difference between adjacent pixels, and mark the corresponding pixel when the color difference is less than a preset color difference threshold.

[0109] The statistical subunit is used to count the marked pixels to obtain the feature contour;

[0110] The safety helmet positioning subunit is used to determine the safety helmet outline based on the outline curve and outline position of the feature contour.

[0111] The level determination subunit is used to read the color value of each pixel in the outline of the safety helmet to determine the personnel level.

[0112] The function of the object recognition sub-unit is to identify personnel and their level. If the recognition algorithm is accurate enough, it can directly perform facial or body recognition to determine the identity of people in the area. However, in actual use, it is not necessary to be accurate to the specific identity of each person. Therefore, it is only necessary to locate a certain part based on the outline to determine the user level. For example, safety helmets are a necessary protective gear in the production area. In addition, if the work clothes have color or other types of markings, the personnel level can also be determined by the outline of the work clothes.

[0113] As a preferred embodiment of the technical solution of the present invention, the first verification unit further includes:

[0114] The partitioning subunit is used to read the regional map and receive partitioning information containing permission levels input by the user based on the regional map.

[0115] The permission query sub-unit is used to sequentially read the personnel position and personnel level of each person in the actual personnel table, and read the permission level corresponding to the personnel position;

[0116] The correction subunit is used to compare the permission level and the personnel level, calculate the difference rate, and correct the first verification value based on the difference rate.

[0117] In one embodiment of the technical solution of this invention, a correction process for the first verification value is provided. Its practical significance lies in determining whether there are personnel in locations where they should not be when the personnel's location is known, and thus correcting the abnormal value. For example, if an intern appears in a danger zone, it can be considered that there is an abnormality in the personnel. Specifically, the rules for correction are generally set by the management. For example, if multiple personnel who do not meet the permission level requirements appear in an area with a certain permission level, the level difference is calculated, and the correction rate is determined by the level difference. The correction rate is a percentage, which is multiplied by the first verification value to obtain the corrected first verification value.

[0118] As a preferred embodiment of the technical solution of the present invention, the supply data detection module includes:

[0119] The target selection unit is used to randomly select target contours from animal and still life contours, identify the target contours, and determine the required data.

[0120] The demand query unit is used to query demand locations based on the location of the target outline; wherein, the demand locations are included in the registered power supply locations;

[0121] The single-time risk calculation unit is used to obtain energy supply data at demand points, compare the energy supply data with the demand data, and determine the single-time risk value.

[0122] The comprehensive risk statistics unit is used to count the risk value of a single instance within a preset number of times, calculate and output the risk value.

[0123] In one example of the technical solution of the present invention, the power supply analysis process is defined. Some contours are randomly selected from animal contours and still life contours, which are called target contours. Both animal contours and still life contours may be selected as target contours. The required points can be queried based on the position of the target contours (generated by the image application module 12). Its practical significance is to query which power supply interface corresponds to each object. The simplest type of power supply data is electrical energy.

[0124] By identifying the target contour, we can determine its demand data, such as how much electricity is needed. By querying the energy supply data of the demand point and comparing the energy supply data with the demand data, we can calculate the risk value. Since it corresponds to a contour, it is called a single risk value. For each target contour, there is a corresponding single risk value. By statistically analyzing multiple single risk values ​​and performing comprehensive calculations, we can obtain the total risk value.

[0125] As a preferred embodiment of the technical solution of the present invention, the emergency control module includes:

[0126] The first comparison unit is used to compare the outlier with a preset outlier threshold and determine the risk threshold based on the comparison result.

[0127] The second comparison unit is used to compare the risk value with the risk threshold and determine the control command of the emergency terminal based on the comparison result.

[0128] The above content defines the priority of outliers and risk values. Outliers refer to personnel abnormalities. If the outlier is zero, it means that the staff in the production area are normal. At this time, the risk is under control and the risk threshold can be set slightly higher. If the outlier is high, it means that the production area itself is in an uncontrollable state. In this case, the risk threshold can be lowered. As long as some current problems occur, an alarm will be triggered. When the risk value is high, the safety door will be opened.

[0129] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A big data security monitoring system, characterized in that, The system includes: The image acquisition module is used to insert information readers into the camera network, which acquire scene images at dynamic frequencies. The camera network is a scene-on-site camera system, including cameras mounted on the top of the area, cameras mounted on equipment, and cameras worn by staff. The image application module is used to perform numerical conversion on scene images, determine the target distribution in the region based on the converted values, calculate outliers based on the target distribution, and update the dynamic frequency synchronously based on the outliers; the targets in the target distribution include animals and still objects. The supply data detection module is used to calculate the demand value based on the target distribution, perform random detection on the supply data based on the demand value, and output the risk value based on the detection result. An emergency control module is used to generate control commands for the emergency terminal based on abnormal values ​​and risk values ​​as reference parameters, and send them to the corresponding emergency terminal; wherein, the emergency terminal includes at least an alarm and a safety door; The image acquisition module includes: The camera parameter acquisition unit is used to query the installed camera terminals with image acquisition functions and obtain the camera's position and camera parameters; the camera parameters include camera direction, camera wide angle, and camera resolution; The area calculation and assignment unit is used to calculate the camera area of ​​each camera end based on the camera direction and camera width, and to determine the additional value of the camera area based on the camera sharpness; The region overlay unit is used to overlay all camera regions and their additional values ​​to obtain a camera range map; the camera range map is used to characterize the information exposure of each location in the region; different values ​​for camera regions correspond to different camera resolutions, which are called additional values. The higher the camera resolution, the more fully the information is displayed and the higher the information exposure; the higher the information exposure, the more camera ends are around the corresponding location. An execution unit is used to determine the number and location of information readers based on the camera range map, and the information readers acquire scene images at a dynamic frequency. The acquisition execution unit includes: The initial position determination subunit is used to determine the initial installation position in the camera range map based on the boundary dimensions of the camera range map and the pre-set fixed-length grid. A circle creation subunit is used to create a circle based on the initial installation position, wherein the radius of the circle is related to the value of each pixel point within the circle; The circle removal subunit is used to remove circles from the camera range map and repeatedly perform the circle creation process in the remaining area until the union of the circles contains the camera range map; The circular application sub-unit is used to determine the number of all circles as the installation quantity and the center position of the circle as the installation position. The calculation rule for the radius of the circle is as follows: as the radius increases, the sum of the information exposure values ​​of each pixel within the circle is calculated, and the maximum radius corresponding to the sum of values ​​being less than a preset threshold is selected as the final radius.

2. The big data security monitoring system according to claim 1, characterized in that, The image application module includes: The numerical conversion unit is used to traverse the pixels in the scene image and convert the value of each pixel into an LBP value. The contour recognition unit is used to input the converted image into a trained convolutional recognition model to label animal and still life contours. The first verification unit is used to acquire access control monitoring data, verify the animal outline based on the access control monitoring data, and obtain a first verification value; The second verification unit is used to obtain still life registration information, verify the still life outline based on the still life registration information, and obtain a second verification value. The frequency update unit is used to determine anomalies based on the first verification value and the second verification value, and to update the dynamic frequency synchronously based on the anomalies.

3. The big data security monitoring system according to claim 2, characterized in that, The content that converts the value of each pixel to an LBP value includes: ； In the formula, For position LBP value at that location, This represents the grayscale value of the p-th pixel in the window, excluding the center pixel. This indicates the total number of pixels in the window excluding the center pixel. This represents the grayscale value of the center pixel; Let be the threshold function, and its values ​​are: ； The window size is 3×3.

4. The big data security monitoring system according to claim 2, characterized in that, The first verification unit includes: The access control information application subunit is used to acquire the entry and exit personnel information collected by the access control terminal, and to determine the theoretical personnel table in the area in real time based on the entry and exit personnel information; The object recognition subunit is used to read animal outlines, identify animal outlines, and determine the actual personnel table; wherein, both the theoretical personnel table and the actual personnel table contain personnel tag items and personnel level items; The table conversion sub-cell is used to convert the theoretical personnel table and the actual personnel table into theoretical quantity groups and actual quantity groups based on the personnel level item, respectively. The difference application subunit is used to subtract the theoretical quantity group and the actual quantity group to obtain the difference quantity group, and to calculate the first verification value based on the difference quantity group.

5. The big data security monitoring system according to claim 4, characterized in that, The object recognition subunit includes: The traversal sub-unit is used to traverse the pixels in the animal outline, calculate the color difference between adjacent pixels, and mark the corresponding pixel when the color difference is less than a preset color difference threshold. The statistical subunit is used to count the marked pixels to obtain the feature contour; The safety helmet positioning subunit is used to determine the safety helmet outline based on the outline curve and outline position of the feature contour. The level determination subunit is used to read the color value of each pixel in the outline of the safety helmet to determine the personnel level.

6. The big data security monitoring system according to claim 4, characterized in that, The first verification unit further includes: The partitioning subunit is used to read the regional map and receive partitioning information containing permission levels input by the user based on the regional map. The permission query sub-unit is used to sequentially read the personnel position and personnel level of each person in the actual personnel table, and read the permission level corresponding to the personnel position; The correction subunit is used to compare the permission level and the personnel level, calculate the difference rate, and correct the first verification value based on the difference rate.

7. The big data security monitoring system according to claim 2, characterized in that, The supply data detection module includes: The target selection unit is used to randomly select target contours from animal and still life contours, identify the target contours, and determine the required data. The demand query unit is used to query demand locations based on the location of the target outline; wherein, the demand locations are included in the registered power supply locations; The single-time risk calculation unit is used to obtain energy supply data at demand points, compare the energy supply data with the demand data, and determine the single-time risk value. The comprehensive risk statistics unit is used to count the risk value of a single instance within a preset number of times, calculate and output the risk value.

8. The big data security monitoring system according to any one of claims 1 to 7, characterized in that, The emergency control module includes: The first comparison unit is used to compare the outlier with a preset outlier threshold and determine the risk threshold based on the comparison result. The second comparison unit is used to compare the risk value with the risk threshold and determine the control command of the emergency terminal based on the comparison result.

Images