Campus security system based on Internet of Things and service terminal thereof

Abstract

The invention belongs to the technical field of campus security and protection, and particularly relates to a campus security and protection system based on Internet of Things and a service terminal thereof, comprising a monitoring module, an image recognition module, a screening and classifying module, a potential risk recognition module and a hierarchical monitoring module. According to the invention, by introducing the Internet of Things technology, real-time monitoring and risk assessment of the campus environment are realized, the efficiency and accuracy of campus security management are improved, and by analyzing the environment image and the personnel image, abnormal conditions can be found in time and classified, so that potential security threats can be quickly responded, and in addition, the safety of the campus is improved. In addition, distribution of monitoring resources can be dynamically adjusted according to the risk level, it is ensured that the risk area of the high-risk draught fan gets enough attention, use of the monitoring resources is optimized, resource waste is avoided, based on the mode, the campus security and protection system can be more intelligent and automatic, and reliable technical guarantee is provided for campus security.

Description

Technical Field

[0001] The present invention belongs to the technical field of campus security, and in particular relates to a campus security system based on the Internet of Things and a service terminal thereof. Background Art

[0002] With the rapid development of Internet of Things technology, campus security systems have been significantly improved. Traditional campus security systems mainly rely on manual monitoring and simple alarm devices, which are difficult to meet the needs of modern campus safety management. In order to improve the level of campus safety management, campus security systems based on the Internet of Things have emerged, aiming to achieve real-time monitoring and early warning of the campus environment, so as to timely discover and deal with risk events on campus.

[0003] Campus security systems in the existing technology mostly capture images through monitoring equipment, but these devices often lack intelligent analysis capabilities and cannot effectively identify and handle abnormal situations. They still require human intervention and supervision, and lack an effective risk assessment mechanism. They cannot effectively identify risk events of the same category and the corresponding potential risks, which will undoubtedly lead to a waste of monitoring resources and delays in risk handling. Based on this, this solution provides a campus security system based on the Internet of Things to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a campus security system based on the Internet of Things and a service terminal thereof, which can realize intelligent identification and hierarchical monitoring of potential risks on campus, thereby improving the efficiency and accuracy of campus security management.

[0005] The technical solution adopted by the present invention is as follows: A campus security system based on the Internet of Things includes a monitoring module, an image recognition module, a screening and classification module, a potential risk identification module, and a hierarchical monitoring module: The monitoring module is used to monitor the environment image and the personnel image in the campus, and pre-process the environment image and the personnel image to obtain a first reference image and a second reference image; The image recognition module is used to identify and analyze the first reference image and the second reference image to obtain security status information on the campus; The screening and classification module is used to perform screening and classification processing on the first reference image and the second reference image according to the safety status information to obtain a first abnormal image and a second abnormal image; The potential risk identification module is used to identify risks based on the first abnormal image and the second abnormal image to determine potential risks on campus; The hierarchical monitoring module is used to perform hierarchical monitoring according to the potential risks on campus and determine the monitoring intensity of potential risks at different risk levels.

[0006] In a preferred embodiment, the monitoring module acquires image data in real time through cameras installed at key locations on campus. The environmental images include images of roads, buildings and green areas on campus, and the personnel images include images of students, faculty and visitors.

[0007] In a preferred embodiment, the monitoring module includes a preprocessing unit, which is used to preprocess the environmental image and the personnel image, including image denoising, enhancement and scaling, and record the preprocessed environmental image as a first reference image and record the preprocessed environmental image as a second reference image.

[0008] In a preferred embodiment, the image recognition module includes a feature extraction unit and an analysis unit, wherein the feature extraction unit is used to perform feature extraction on the first reference image and the second reference image, extract key features in the image, calculate the feature distance between the key features and the preset risk area, and record it as the condition parameter to be evaluated; A safety distance is provided in the analysis unit for comparison with the condition parameter to be evaluated; Wherein, when the condition parameter to be evaluated is less than or equal to the safety distance, it indicates that an abnormality exists in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as an abnormal state; When the condition parameter to be evaluated is greater than the safety distance, it indicates that there is no abnormality in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as a safe state.

[0009] In a preferred embodiment, the screening and classification module includes a screening unit and a classification unit, wherein the screening unit is used to screen the first reference image and the second reference image according to the security status information, exclude normal images, and retain abnormal images, wherein the abnormal images include a first abnormal image corresponding to the first reference image, and a second abnormal image corresponding to the second reference image; The classification unit performs classification processing according to the similarity between the first abnormal image and the second abnormal image, and outputs a plurality of classification subsets according to the classification results, and determines the initial monitoring intensity of the risk area corresponding to each of the first abnormal image and the second abnormal image according to the number of the first abnormal image and the second abnormal image in each of the classification subsets; The greater the number of first abnormal images or second abnormal images in the classification subset, the higher the corresponding initial monitoring intensity.

[0010] In a preferred solution, when the classification subset is output, a similarity measurement is performed on the first abnormal image in the same risk area and the second abnormal image in the same risk area, and the specific process is as follows: Calling the first abnormal image in the same risk area and performing vectorization processing to obtain a first feature vector; Calling a second abnormal image in the same risk area and performing vectorization processing to obtain a second feature vector; Introduce a preset similarity measurement function; Inputting the first feature vectors of the same risk area into a similarity measurement function to obtain first similarity scores between different first feature vectors; Inputting the second feature vectors of the same risk area into a similarity measurement function to obtain second similarity scores between different second feature vectors; Obtaining a first classification threshold and a second classification threshold, and comparing the first classification threshold with the first similarity score, and comparing the second classification threshold with the second similarity score; When the first classification threshold is higher than the first similarity score, it indicates that there is no correlation between the first abnormal images under the first similarity score, and the first abnormal images under the first similarity score are classified into different classification subsets; When the first classification threshold is lower than or equal to the first similarity score, it indicates that the first abnormal image under the first similarity score is a related image, and the first abnormal image under the first similarity score is classified into the same classification subset; When the second classification threshold is higher than the second similarity score, it indicates that there is no correlation between the second abnormal images under the second similarity score, and the second abnormal images under the second similarity score are classified into different classification subsets; When the second classification threshold is lower than or equal to the second similarity score, it indicates that the second abnormal image under the second similarity score is a related image, and the second abnormal image under the second similarity score is classified into the same binary classification subset.

[0011] In a preferred embodiment, the potential risk identification module includes a risk sampling unit and a risk determination unit, wherein the risk sampling unit collects collection nodes of the first abnormal image and the second abnormal image, and performs backtracking offset processing on the collection nodes to obtain a risk assessment period, and sets a plurality of sampling nodes within the risk assessment period, and collects the historical distance between the key features and the risk area in the first abnormal image and the second abnormal image under each of the sampling nodes, and records the historical distance as a sample condition parameter; The risk determination unit is used to determine the association state between the sample condition parameter and the condition parameter to be evaluated according to the distribution law of the sample condition parameter, wherein the association state includes an ordered state and a disordered state; In the ordered state, it is indicated that the historical distance between the key features and the risk area during the risk assessment period has regularity, and the potential risk distance is calculated based on the sample condition parameters and the monitoring feedback delay; The disordered state indicates that the historical distance between the key feature and the risk area during the risk assessment period is irregular, and the maximum historical distance between the key feature and the risk area during the risk assessment period is recorded as the potential risk distance.

[0012] In a preferred solution, the hierarchical monitoring module includes a hierarchical monitoring unit and a response unit, wherein the hierarchical monitoring unit is used to collect the potential risk distances corresponding to the risk areas in the first abnormal image and the second abnormal image under each classification subset; Inputting the potential risk distance and the number of the first abnormal image and the second abnormal image into a preset evaluation function, and recording the output result of the evaluation function as a risk score; Arrange the risk scores in descending order, and allocate monitoring resources one by one according to the arrangement results of the risk scores; The response unit is used to trigger an early warning mechanism and send an alarm signal when a potential risk corresponding to the risk area in the first abnormal image or the second abnormal image is monitored in the risk area.

[0013] The present invention also provides a campus security service method based on the Internet of Things, which is applied to the campus security system based on the Internet of Things, and includes the following steps: Monitoring the environment image and the personnel image in the campus, and preprocessing the environment image and the personnel image to obtain a first reference image and a second reference image; Identify and analyze the first reference image and the second reference image to obtain security status information on the campus; Screening and classifying the first reference image and the second reference image according to the safety status information to obtain a first abnormal image and a second abnormal image; Perform risk identification based on the first abnormal image and the second abnormal image to determine potential risks on campus; Carry out graded monitoring based on potential risks on campus and determine the monitoring intensity of potential risks at different risk levels.

[0014] And, a service terminal for campus security based on the Internet of Things, the service terminal comprising: at least one processor; and a memory communicatively coupled to the at least one processor; Among them, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned campus security system based on the Internet of Things.

[0015] The technical effects achieved by the present invention are: By introducing the Internet of Things technology, the present invention realizes real-time monitoring and risk assessment of the campus environment, improves the efficiency and accuracy of campus security management, and can promptly discover abnormal situations and classify them by analyzing environmental images and personnel images, thereby quickly responding to potential security threats. In addition, the present invention can also dynamically adjust the allocation of monitoring resources according to the risk level to ensure that the risk areas of high-risk wind turbines receive sufficient attention, while optimizing the use of monitoring resources to avoid waste of resources. Based on this approach, the campus security system can be more intelligent and automated, providing reliable technical guarantees for campus safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the system module of the present invention; Figure 2 It is a schematic flow chart of the method of the present invention; Figure 3 It is a schematic diagram of the service terminal structure of the present invention. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0018] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The term "in a preferred embodiment" that appears in different places in this specification does not refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0020] See also Figure 1 As shown, the present invention provides a campus security system based on the Internet of Things, including a monitoring module, an image recognition module, a screening and classification module, a potential risk identification module and a hierarchical monitoring module: The monitoring module is used to monitor the environment image and the personnel image in the campus, and pre-process the environment image and the personnel image to obtain the first reference image and the second reference image; The image recognition module is used to recognize and analyze the first reference image and the second reference image to obtain security status information on the campus; The screening and classification module is used to screen and classify the first reference image and the second reference image according to the security status information to obtain a first abnormal image and a second abnormal image; The potential risk identification module is used to identify risks based on the first abnormal image and the second abnormal image to determine potential risks on campus; The hierarchical monitoring module is used to conduct hierarchical monitoring based on potential risks within the campus and determine the monitoring intensity of potential risks at different risk levels.

[0021] In the present invention, with the continuous development of information technology, its application fields are becoming more and more extensive. Campus security is one of them. It can grasp the security status of the campus in real time through online monitoring. The security system provided by this embodiment includes a monitoring module, an image recognition module, a screening and classification module, a potential risk identification module and a hierarchical monitoring module. The main responsibility of the monitoring module is to monitor the environmental images and personnel images in the campus in real time. Through the cameras installed in various corners of the campus, the real-time image screens in the campus are captured and corresponding preprocessing is performed. The preprocessing process includes operations such as image denoising and enhancement, and finally the first reference image and the second reference image are obtained as the basis for subsequent analysis. The image recognition module is responsible for performing corresponding identification and analysis on the first reference image and the second reference image provided by the monitoring module, extracting key information from the image, such as greening distribution, personnel identity, behavioral characteristics, etc., and determining the security status information in the campus based on the key information, including but not limited to abnormal behavior of personnel, the appearance of suspicious items, etc. The role of the screening and classification module is to further process the security status information provided by the image recognition module, screen and classify the first reference image and the second reference image, so as to identify the first abnormal image and the second abnormal image, such as The appearance of unidentified persons, abnormal behavior, or the lush greenery affecting the width of the road, etc. The potential risk identification module identifies risks based on the first abnormal image and the second abnormal image provided by the screening and classification module. By analyzing the abnormal behavior and characteristics in the image, the potential risks in the campus and the risk level of the potential risks can be determined. The hierarchical monitoring module performs hierarchical monitoring according to the risk level provided by the potential risk identification module. For example, if unidentified persons or suspicious items frequently appear in a certain area, the potential risk identification module will mark it as a high-risk area. This process not only relies on image analysis, but also combines historical data and real-time information to ensure the comprehensiveness and accuracy of risk assessment. The hierarchical monitoring module performs hierarchical monitoring based on the risk area information provided by the potential risk identification module. According to the different risk levels, the monitoring intensity of potential risks of different risk levels is determined. For the campus areas corresponding to the first abnormal image and the second abnormal image of high wind, the system will increase the monitoring frequency and the investment of monitoring resources to ensure that these areas can be more closely monitored. The implementation of the hierarchical monitoring module not only improves the monitoring efficiency, but also ensures the flexibility and adaptability of the campus security prevention system, providing strong technical support for campus security prevention.

[0022] In a preferred embodiment, the monitoring module acquires image data in real time through cameras installed at key locations on campus. Environmental images include images of roads, buildings and green areas on campus, and personnel images include images of students, faculty and visitors.

[0023] In this embodiment, the monitoring module captures and obtains real-time image data in real time through high-definition cameras deployed at various key locations on campus. The real-time image data covers multiple aspects of the campus, including but not limited to images of roads, various buildings and green areas on campus. Environmental images can help understand the environmental conditions of the campus and ensure the safety and cleanliness of the campus environment. At the same time, the monitoring module is also responsible for capturing personnel images, which include images of students, faculty and staff, and various visitors on campus, to assist in managing the flow of personnel on campus and ensure the safety and order of the campus.

[0024] In a preferred embodiment, the monitoring module includes a preprocessing unit, which is used to preprocess the environmental image and the personnel image, including image denoising, enhancement and scaling, and record the preprocessed environmental image as a first reference image and record the preprocessed environmental image as a second reference image.

[0025] In this embodiment, a preprocessing unit is also provided in the monitoring module, and the preprocessing unit is used to perform preliminary processing on the environmental image and the personnel image. Specifically, the main functions of the preprocessing unit include denoising, enhancing and scaling the image. Denoising is to remove noise interference in the image and ensure the image quality, thereby improving the accuracy of subsequent processing. Enhancement is to improve the contrast and clarity of the image, making the details in the image more obvious and convenient for subsequent analysis. The scaling operation is to adjust the size of the image to adapt it to the needs of subsequent processing. After completing these preprocessing steps, the preprocessing unit will record the processed environmental image as the first reference image and the processed personnel image as the second reference image to facilitate the subsequent corresponding risk analysis.

[0026] In a preferred embodiment, the image recognition module includes a feature extraction unit and an analysis unit, the feature extraction unit is used to extract features from the first reference image and the second reference image, extract key features in the image, and calculate the feature distance between the key features and the preset risk area, and record it as the condition parameter to be evaluated; A safety distance is set in the analysis unit for comparison with the condition parameters to be evaluated; Wherein, when the condition parameter to be evaluated is less than or equal to the safety distance, it indicates that an abnormality exists in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as an abnormal state; When the condition parameter to be evaluated is greater than the safety distance, it indicates that there is no abnormality in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as a safe state.

[0027] In this implementation, the image recognition module is mainly composed of two parts: a feature extraction unit and an analysis unit. The feature extraction unit mainly performs corresponding feature extraction on the first reference image and the second reference image. The process involves identifying and extracting key features in the image. The key features can be personnel features, object features or scene features, etc. After the key features are extracted, the feature distances between these key features and the preset risk areas will be further calculated. Here, it should be clarified that different key features correspond to different risk areas. For example, adult groups such as faculty and staff on campus are allowed to approach areas, while student groups may not be allowed to approach (such as the edge of the campus wall during normal teaching periods, etc.). For example, the branches of landscape trees are not allowed to extend into the internal roads of the campus where students normally travel. Then the internal roads of the campus are relative to the landscape trees. The branches of the tree are risk areas. The risk areas of different key features need to be formulated according to the management needs of school administrators and the safety needs on campus. The feature distance is an indicator to measure whether there is a potential risk in the image. In this embodiment, the calculated feature distance will be recorded as a condition parameter to be evaluated. The analysis unit performs further analysis based on the feature extraction unit. A safety distance threshold is pre-set in the analysis unit. The safety distance threshold is used to compare with the condition parameter to be evaluated. The safety distance threshold is pre-set according to the actual application scenario and safety needs, and is used to determine whether the key features in the image are in a safe state. Specifically, when the condition parameter to be evaluated is less than or equal to the set safety distance threshold, it means that there may be an abnormality in the first reference image or the second reference image corresponding to the key feature. In this case, the safety status information corresponding to the key feature will be recorded as an abnormal state, indicating that there is a safety risk in the image. On the contrary, when the condition parameter to be evaluated is greater than the safety distance threshold, it indicates that there is no abnormality in the first reference image or the second reference image corresponding to the key feature. At this time, the safety status information corresponding to the key feature will be recorded as a safe state, indicating that the key feature in the image is within a safe range.

[0028] In a preferred embodiment, the screening and classification module includes a screening unit and a classification unit, the screening unit is used to screen the first reference image and the second reference image according to the security status information, exclude normal images, and retain abnormal images, wherein the abnormal images include a first abnormal image corresponding to the first reference image, and a second abnormal image corresponding to the second reference image; The classification unit performs classification processing according to the similarity between the first abnormal image and the second abnormal image, outputs a plurality of classification subsets according to the classification results, and determines the initial monitoring intensity of the risk area corresponding to each of the first abnormal image and the second abnormal image according to the number of the first abnormal image and the second abnormal image in each classification subset; The greater the number of first abnormal images or second abnormal images in the classification subset, the higher the corresponding initial monitoring intensity.

[0029] In this embodiment, the screening and classification module is mainly composed of a screening unit and a classification unit. The main responsibility of the screening unit is to use the security status information to perform corresponding screening processing on the first reference image and the second reference image. Specifically, the screening unit will exclude normal images and retain images with abnormal conditions. Abnormal images with abnormal conditions can be further divided into two categories: one is the first abnormal image corresponding to the first reference image, and the other is the second abnormal image corresponding to the second reference image. After that, the classification unit classifies the first abnormal image and the second abnormal image. The core of the classification unit is to perform detailed classification based on the similarity between the first abnormal image and the second abnormal image. Through this type of classification processing, the system can divide abnormal images into multiple classification subsets. The number of images and feature similarity in each classification subset will be recorded and analyzed by the system. After the classification is completed, the system will further determine the initial monitoring intensity of the risk area corresponding to each abnormal image based on the number of first abnormal images and second abnormal images in each classification subset. Specifically, if the number of first abnormal images or second abnormal images in a classification subset is large, the system will automatically determine that the initial monitoring intensity of the area is high, and vice versa. This mechanism ensures that monitoring resources can be more efficiently allocated to areas with higher potential risks, thereby improving the safety and response efficiency of the entire system.

[0030] In a preferred embodiment, when the classification subset is output, the similarity measurement is performed on the first abnormal image in the same risk area and the second abnormal image in the same risk area, and the specific process is as follows: Calling the first abnormal image in the same risk area and performing vectorization processing to obtain a first feature vector; Calling a second abnormal image in the same risk area and performing vectorization processing to obtain a second feature vector; Introduce a preset similarity measurement function; Inputting the first eigenvectors of the same risk area into the similarity measurement function to obtain the first similarity scores between different first eigenvectors; Inputting the second eigenvectors of the same risk area into the similarity measurement function to obtain the second similarity scores between different second eigenvectors; Obtaining a first classification threshold and a second classification threshold, and comparing the first classification threshold with the first similarity score, and comparing the second classification threshold with the second similarity score; When the first classification threshold is higher than the first similarity score, it indicates that there is no correlation between the first abnormal images under the first similarity score, and the first abnormal images under the first similarity score are classified into different classification subsets; When the first classification threshold is lower than or equal to the first similarity score, it indicates that the first abnormal image under the first similarity score is a related image, and the first abnormal image under the first similarity score is classified into the same classification subset; When the second classification threshold is higher than the second similarity score, it indicates that there is no correlation between the second abnormal images under the second similarity score, and the second abnormal images under the second similarity score are classified into different classification subsets; When the second classification threshold is lower than or equal to the second similarity score, it indicates that the second abnormal image under the second similarity score is a related image, and the second abnormal image under the second similarity score is classified into the same binary classification subset.

[0031] In the above, in the process of outputting the classification subset, it is necessary to perform corresponding similarity measurement on the first abnormal image and the second abnormal image in the same risk area. First, it is necessary to call the first abnormal image in the same risk area and perform vectorization processing on it. Through this process, the vector of each first abnormal image feature can be obtained and recorded as the first feature vector. Similarly, the second abnormal image in the same risk area is also called and vectorized to obtain the vector of each second abnormal image feature, that is, the second feature vector. The similarity between the first abnormal images. In order to measure the similarity between the two feature vectors, a preset similarity measurement function is introduced for calculation, wherein the expression of the similarity measurement function is: , where represents a similarity score (including a first similarity score and a second similarity score, which are specifically determined according to the variables input into the similarity measurement function), represents the number of feature points in the first eigenvector or the second eigenvector, and Represents different first eigenvectors, or different second eigenvectors. Based on the above formula, the first similarity scores between different first abnormal images in the same risk area, and the similarity scores between different second abnormal images in the same risk area can be input. Then, the correlation between these abnormal images will be further determined. Specifically, two preset classification thresholds need to be obtained: the first classification threshold and the second classification threshold. Then, the first classification threshold is compared with the first similarity score to determine the correlation between the first abnormal images. If the first classification threshold is higher than the first similarity score, it means that there is no obvious correlation between the first abnormal images under the first similarity score. Therefore, these first abnormal images are classified into different classification subsets. On the contrary, if the first classification threshold is lower than or equal to the first similarity score, The similarity score indicates that there is a correlation between the first abnormal images under the first similarity score, so these first abnormal images will be classified into the same classification subset. Similarly, the second classification threshold is compared with the second similarity score to determine the correlation between the second abnormal images. If the second classification threshold is higher than the second similarity score, it means that there is no obvious correlation between the second abnormal images under the second similarity score, so these second abnormal images are classified into different classification subsets. On the contrary, if the second classification threshold is lower than or equal to the second similarity score, it indicates that there is a correlation between the second abnormal images under the second similarity score, and these second abnormal images are classified into the same classification subset, so the first abnormal images and the second abnormal images in the same risk area can be classified one by one.

[0032] In a preferred embodiment, the potential risk identification module includes a risk sampling unit and a risk determination unit. The risk sampling unit collects the collection nodes of the first abnormal image and the second abnormal image, and performs backtracking offset processing on the collection nodes to obtain a risk assessment period, and sets multiple sampling nodes within the risk assessment period, and collects the historical distance between the key features and the risk area in the first abnormal image and the second abnormal image under each sampling node, and records it as a sample condition parameter; The risk determination unit is used to determine the association state between the sample condition parameter and the condition parameter to be evaluated according to the distribution law of the sample condition parameter, wherein the association state includes an ordered state and a disordered state; In the ordered state, it indicates that the historical distance between the key features and the risk area during the risk assessment period is regular, and the potential risk distance is calculated based on the sample condition parameters and the monitoring feedback delay; In the disordered state, it indicates that the historical distance between the key features and the risk area during the risk assessment period is irregular, and the maximum historical distance between the key features and the risk area during the risk assessment period is recorded as the potential risk distance.

[0033] In this implementation, the potential risk identification module is composed of a risk sampling unit and a risk determination unit. The main responsibility of the risk sampling unit is to collect the collection nodes of the first abnormal image and the second abnormal image. The collection nodes of the first abnormal image and the second abnormal image are the key nodes where the abnormal situation occurs. In order to perform risk assessment more accurately, the collection nodes will be backtracked and offset, that is, offset to the time before the abnormal situation occurs, so as to determine the risk assessment period. During the risk assessment period, the risk sampling unit will set multiple sampling nodes. Under each sampling node, the risk sampling unit will collect the historical distance between the key features and the risk area in the first abnormal image and the second abnormal image, and record it as a sample condition parameter to provide basic data support for subsequent risk assessment. The risk determination unit determines the association state between the sample condition parameter and the condition parameter to be evaluated based on the distribution law of the sample condition parameter. The association state can be divided into an ordered state and a disordered state. In the ordered state, it indicates that the historical distance between the key features and the risk area has a certain regularity during the risk assessment period. This regularity can provide an important reference for risk assessment. In this case, the risk determination unit will calculate the potential risk distance based on the sample condition parameters and the monitoring feedback delay, thereby providing decision makers with more accurate risk assessment results. In the disordered state, it indicates that the historical distance between the key features and the risk area has no obvious regularity during the risk assessment period. In this case, the risk determination unit will record the maximum historical distance between the key features and the risk area during the risk assessment period as the potential risk distance, providing a safety margin for decision makers to ensure that appropriate preventive measures can be taken even in the most unfavorable circumstances. After the potential risk distance is output, the original risk area will be offset according to the potential risk distance to obtain the potential risk area. When the same type of events occur in the subsequent potential risk area, timely feedback can be obtained. In the above, when determining the association state between the sample condition parameter and the condition parameter to be evaluated, it is first necessary to introduce a preset condition function, wherein the expression of the condition function is: , where Indicates the verification condition parameters. Indicates the length of the risk assessment period, represents the number of sampling nodes, and Represents the sample condition parameters under adjacent sampling nodes. After the verification condition parameters are output, the verification process will be performed based on the sample condition parameters under adjacent sampling nodes and the time interval between adjacent sampling nodes to determine the association state between the sample condition parameters and the condition parameters to be evaluated. The specific formula is: , where Indicates the calibration deviation. Indicates the sample condition parameter that occurs earlier in time among adjacent sampling nodes. Indicates the sample condition parameter that occurs later in time among adjacent sampling nodes. Represents the time interval between adjacent sampling nodes, and then calculates the sample condition parameters under adjacent sampling nodes one by one to obtain multiple verification check quantities, and then sums and averages the verification deviations to obtain the standard verification deviation, and compares the standard verification deviation with the preset allowable verification deviation. When the standard verification deviation is less than the allowable verification deviation, the association state between the corresponding sample condition parameter and the condition parameter to be evaluated is recorded as an ordered state, otherwise, it is recorded as a disordered state. In the ordered state, combined with the function Calculate the potential risk distance, where Indicates the potential risk distance, Indicates the conditional parameter to be evaluated. Indicates the duration of monitoring feedback.

[0034] In a preferred embodiment, the hierarchical monitoring module includes a hierarchical monitoring unit and a response unit, and the hierarchical monitoring unit is used to collect the potential risk distance corresponding to the risk area in the first abnormal image and the second abnormal image under each classification subset; Inputting the potential risk distance, and the number of the first abnormal image and the second abnormal image into a preset evaluation function, and recording the output result of the evaluation function as a risk score; Arrange the risk scores in descending order and allocate monitoring resources one by one according to the risk score arrangement results; The response unit is used to trigger the early warning mechanism and send an alarm signal when a potential risk corresponding to the risk area in the first abnormal image or the second abnormal image is monitored in the risk area.

[0035] In this embodiment, the hierarchical monitoring module mainly includes a hierarchical monitoring unit and a response unit. The hierarchical monitoring unit is responsible for real-time acquisition and monitoring of the corresponding potential risk distances of the first abnormal image and the second abnormal image under each classification subset. In order to more accurately assess the risk, the potential risk distance and the number of the first abnormal image and the second abnormal image are input into a pre-set evaluation function, thereby calculating a comprehensive risk score to quantify the potential risk under the current monitoring environment, wherein the expression of the evaluation function is: , where represents the risk score, represents the weight factor in risk distance, represents the potential risk distance corresponding to the first abnormal image or the potential risk distance corresponding to the second abnormal image, A weight factor representing the number of first abnormal images or the number of second abnormal images, Indicates the number of first abnormal images or the number of second abnormal images. After obtaining the risk score, the hierarchical monitoring unit will arrange the risk scores in order from high to low. Through this sorting, the monitoring system can allocate monitoring resources more effectively. Risk scores with higher priority mean higher potential risks, so more monitoring resources are needed to deal with them, ensuring that monitoring resources are reasonably allocated to where they are most needed, thereby improving the efficiency and effectiveness of the entire monitoring system. The main task of the response unit is to conduct real-time monitoring in the risk area. When potential risks are detected in the first abnormal image or the second abnormal image, the response unit will immediately trigger the early warning mechanism and send an alarm signal. This alarm signal can be in various forms, such as notifying relevant personnel or systems through sound, light, text message, email, etc. The purpose of the alarm signal is to ensure that the potential risk situation is notified to the relevant personnel at the first time so that they can quickly take corresponding countermeasures to minimize potential risks.

[0036] See also Figure 2 A campus security service method based on the Internet of Things is applied to the above-mentioned campus security system based on the Internet of Things, comprising the following steps: Monitoring the environment image and the personnel image in the campus, and preprocessing the environment image and the personnel image to obtain a first reference image and a second reference image; Identify and analyze the first reference image and the second reference image to obtain security status information on the campus; Screening and classifying the first reference image and the second reference image according to the security status information to obtain a first abnormal image and a second abnormal image; Perform risk identification based on the first abnormal image and the second abnormal image to determine potential risks on campus; Carry out graded monitoring based on potential risks on campus and determine the monitoring intensity of potential risks at different risk levels.

[0037] When the above method is executed, first, the environmental image and personnel image in the campus are monitored in real time through the sensors and cameras installed in various corners of the campus. The environmental image and personnel image data are transmitted to the central processing system for preliminary preprocessing operations to ensure that the image quality meets the requirements of subsequent analysis. The preprocessing steps include image denoising, enhancement and standardization, and finally a clear first reference image and a second reference image are obtained. Next, the first reference image and the second reference image are deeply identified and analyzed, and key information such as personnel identity and behavioral characteristics are extracted from the image to evaluate the security status in the campus. Then, the first reference image and the second reference image are screened and classified according to the security status information. Through the set thresholds and rules, abnormal images, that is, images that do not conform to the normal mode or have potential threats, can be identified. The abnormal images are marked as the first abnormal image and the second abnormal image. Based on the first abnormal image and the second abnormal image, risk identification will be performed to identify the potential risk area and potential risk distance in the campus, which will help to timely discover and warn possible security threats when similar phenomena occur. Finally, according to the identified potential risks in the campus, hierarchical monitoring will be carried out. Different risk levels will correspond to different monitoring intensities and response measures to ensure campus safety in a targeted manner.

[0038] See also Figure 3 , a service terminal for campus security based on the Internet of Things, the service terminal includes: at least one processor; and a memory communicatively coupled to the at least one processor; Among them, the memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the above-mentioned campus security system based on the Internet of Things.

[0039] The processor of the service terminal mentioned above can be any general-purpose processor, such as a central processing unit (CPU) or a graphics processing unit (GPU), or a specially designed processor, such as a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC). The memory can be a random access memory (RAM), a read-only memory (ROM), a flash memory or other types of non-volatile memory. The computer program contains a series of instructions that can guide the processor to perform specific tasks, such as data acquisition, image processing, risk assessment and triggering of early warning mechanisms. The service terminal may also include an operator, an input device and an output device. The operator can perform complex mathematical operations, input devices such as keyboards and mice are used to operate and control the service terminal, and output devices such as displays and printers are used to display processing results and output information. The computer program realizes real-time monitoring and management of campus security conditions through the collaborative work of these hardware components.

[0040] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0041] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A campus security system based on the Internet of Things, characterized by: It includes monitoring module, image recognition module, screening and classification module, potential risk identification module and hierarchical monitoring module: The monitoring module is used to monitor the environment image and the personnel image in the campus, and pre-process the environment image and the personnel image to obtain a first reference image and a second reference image; The image recognition module is used to identify and analyze the first reference image and the second reference image to obtain security status information on the campus; The screening and classification module is used to perform screening and classification processing on the first reference image and the second reference image according to the safety status information to obtain a first abnormal image and a second abnormal image; The potential risk identification module is used to identify risks based on the first abnormal image and the second abnormal image to determine potential risks on campus; The hierarchical monitoring module is used to perform hierarchical monitoring according to the potential risks on campus and determine the monitoring intensity of potential risks at different risk levels.

2. The campus security system based on the Internet of Things according to claim 1 is characterized in that: The monitoring module acquires image data in real time through cameras installed at key locations on campus. The environmental images include images of roads, buildings and green areas on campus, and the personnel images include images of students, faculty and staff and visitors.

3. The campus security system based on the Internet of Things according to claim 1 is characterized in that: The monitoring module includes a preprocessing unit, which is used to preprocess the environment image and the personnel image, including image denoising, enhancement and scaling, and record the preprocessed environment image as a first reference image and record the preprocessed environment image as a second reference image.

4. The campus security system based on the Internet of Things according to claim 1 is characterized in that: The image recognition module includes a feature extraction unit and an analysis unit, wherein the feature extraction unit is used to perform feature extraction on the first reference image and the second reference image, extract key features in the image, calculate a feature distance between the key features and a preset risk area, and record the distance as a condition parameter to be evaluated; A safety distance is provided in the analysis unit for comparison with the condition parameter to be evaluated; Wherein, when the condition parameter to be evaluated is less than or equal to the safety distance, it indicates that an abnormality exists in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as an abnormal state; When the condition parameter to be evaluated is greater than the safety distance, it indicates that there is no abnormality in the first reference image or the second reference image corresponding to the key feature, and the safety status information corresponding to the key feature is recorded as a safe state.

5. The campus security system based on the Internet of Things according to claim 1 is characterized in that: The screening and classification module includes a screening unit and a classification unit, wherein the screening unit is used to screen the first reference image and the second reference image according to the security status information, exclude normal images, and retain abnormal images, wherein the abnormal images include a first abnormal image corresponding to the first reference image, and a second abnormal image corresponding to the second reference image; The classification unit performs classification processing according to the similarity between the first abnormal image and the second abnormal image, and outputs a plurality of classification subsets according to the classification results, and determines the initial monitoring intensity of the risk area corresponding to each of the first abnormal image and the second abnormal image according to the number of the first abnormal image and the second abnormal image in each of the classification subsets; The greater the number of first abnormal images or second abnormal images in the classification subset, the higher the corresponding initial monitoring intensity.

6. The campus security system based on the Internet of Things according to claim 1 is characterized in that: When the classification subset is output, the similarity measurement is performed on the first abnormal image in the same risk area and the second abnormal image in the same risk area. The specific process is as follows: Calling the first abnormal image in the same risk area and performing vectorization processing to obtain a first feature vector; Calling a second abnormal image in the same risk area and performing vectorization processing to obtain a second feature vector; Introduce a preset similarity measurement function; Inputting the first feature vectors under the same risk area into a similarity measurement function to obtain first similarity scores between different first feature vectors; Inputting the second feature vectors of the same risk area into a similarity measurement function to obtain second similarity scores between different second feature vectors; Obtaining a first classification threshold and a second classification threshold, and comparing the first classification threshold with the first similarity score, and comparing the second classification threshold with the second similarity score; When the first classification threshold is higher than the first similarity score, it indicates that there is no correlation between the first abnormal images under the first similarity score, and the first abnormal images under the first similarity score are classified into different classification subsets; When the first classification threshold is lower than or equal to the first similarity score, it indicates that the first abnormal image under the first similarity score is a related image, and the first abnormal image under the first similarity score is classified into the same classification subset; When the second classification threshold is higher than the second similarity score, it indicates that there is no correlation between the second abnormal images under the second similarity score, and the second abnormal images under the second similarity score are classified into different classification subsets; When the second classification threshold is lower than or equal to the second similarity score, it indicates that the second abnormal image under the second similarity score is a related image, and the second abnormal image under the second similarity score is classified into the same binary classification subset.

7. The campus security system based on the Internet of Things according to claim 1 is characterized in that: The potential risk identification module includes a risk sampling unit and a risk determination unit, wherein the risk sampling unit collects collection nodes of the first abnormal image and the second abnormal image, and performs backtracking offset processing on the collection nodes to obtain a risk assessment period, and sets a plurality of sampling nodes within the risk assessment period, and collects the historical distance between the key features and the risk area in the first abnormal image and the second abnormal image at each of the sampling nodes, and records the historical distance as a sample condition parameter; The risk determination unit is used to determine the association state between the sample condition parameter and the condition parameter to be evaluated according to the distribution law of the sample condition parameter, wherein the association state includes an ordered state and a disordered state; In the ordered state, it is indicated that the historical distance between the key features and the risk area during the risk assessment period has regularity, and the potential risk distance is calculated based on the sample condition parameters and the monitoring feedback delay; The disordered state indicates that the historical distance between the key feature and the risk area during the risk assessment period is irregular, and the maximum historical distance between the key feature and the risk area during the risk assessment period is recorded as the potential risk distance.

8. The campus security system based on the Internet of Things according to claim 7 is characterized in that: The hierarchical monitoring module includes a hierarchical monitoring unit and a response unit, wherein the hierarchical monitoring unit is used to collect the potential risk distances corresponding to the risk areas in the first abnormal image and the second abnormal image under each classification subset; Inputting the potential risk distance and the number of the first abnormal image and the second abnormal image into a preset evaluation function, and recording the output result of the evaluation function as a risk score; Arrange the risk scores in descending order, and allocate monitoring resources one by one according to the arrangement results of the risk scores; The response unit is used to trigger an early warning mechanism and send an alarm signal when a potential risk corresponding to the risk area in the first abnormal image or the second abnormal image is monitored in the risk area.

9. A campus security service method based on the Internet of Things, characterized in that: The campus security system based on the Internet of Things as described in any one of claims 1 to 8 comprises the following steps: Monitoring the environment image and the personnel image in the campus, and preprocessing the environment image and the personnel image to obtain a first reference image and a second reference image; Identify and analyze the first reference image and the second reference image to obtain security status information on the campus; Screening and classifying the first reference image and the second reference image according to the safety status information to obtain a first abnormal image and a second abnormal image; Perform risk identification based on the first abnormal image and the second abnormal image to determine potential risks on campus; Carry out graded monitoring based on potential risks within the campus and determine the monitoring intensity of potential risks at different risk levels.

10. A campus security service terminal based on the Internet of Things, characterized by: The service terminal comprises: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the campus security system based on the Internet of Things as described in any one of claims 1 to 8.

Images