Campus safety supervision early warning method and system

Abstract

The invention relates to the technical field of early warning methods, in particular to a campus safety supervision early warning method and system, and the method comprises the steps: obtaining the position data of each camera in a campus; the deviation degree of each camera is calculated, and the importance degree of the camera is obtained based on the product of the deviation degree of the camera and the frequency of dangerous behaviors in the camera monitoring area; and adjusting the danger threshold value of the monitoring area of each camera based on the importance degree of each camera. According to the invention, the danger threshold value in the monitoring area of each camera is adjusted by using the importance degree of each camera, so that the higher the importance degree of each camera is, the lower the danger threshold value of the monitoring area is, and the system has higher responsiveness and flexibility. In an area where the dangerous behavior occurrence frequency is high, the sensitivity of the trigger threshold is improved, and the safety guarantee of a campus or a public place is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of early warning methods, and more specifically, to a campus safety supervision early warning method and system. Background Art

[0002] With the continuous development of the times, people's quality of life is constantly improving, but it is accompanied by parents' high demands and strict management of their children. The psychological pressure of students is constantly increasing, and anxiety is constantly being amplified, leading to frequent campus tragedies. In order to reduce the occurrence of tragic events, schools should not only provide psychological counseling for students, help children enhance their psychological endurance, and cultivate a sound self-awareness and understanding of society, but also take necessary preventive measures, which is also an important part of campus safety management. With the rapid development of modern monitoring technology, Internet of Things technology, artificial intelligence and other fields, video monitoring, sensor technology and big data analysis have been applied to various public safety systems. Traditional monitoring systems mainly rely on manual intervention and regular inspections, cannot respond to campus safety events in a timely manner, and lack effective early warning mechanisms.

[0003] In the related technology, for example, the Chinese patent application document with publication number CN116934550A discloses a campus security incident processing and automatic early warning method based on artificial intelligence. Based on the first mark and the second mark of the abnormal signal, an early warning mechanism corresponding to the type and severity of the security incident is generated, and based on the early warning mechanism, an early warning signal is sent to various devices and related terminals on campus.

[0004] However, when identifying dangerous behaviors through people's trajectories, due to the different degrees of deviation and the probability of dangerous behaviors occurring in different regulatory areas, when using fixed danger thresholds to identify dangerous behaviors, problems such as misidentification or untimely danger identification may occur. Summary of the invention

[0005] The present invention provides a campus safety supervision and early warning method and system, aiming to solve the problem in the related technology that due to the different degrees of deviation and the probability of dangerous behaviors occurring in different supervision areas, misidentification or untimely danger identification may occur when using a fixed danger threshold to identify dangerous behaviors.

[0006] In the first aspect, the present invention provides a campus safety supervision and early warning method, comprising: obtaining the position data of each camera in the campus; calculating the deviation of each camera, and obtaining the importance of the camera based on the product of the deviation of the camera and the frequency of dangerous behaviors in the camera monitoring area; adjusting the danger threshold of each camera monitoring area based on the importance of each camera, and obtaining the danger level of the current behavior of the person to be supervised in the camera monitoring area according to the average value of the similarity between the current behavior of the person to be supervised and each historical dangerous behavior, and issuing an early warning based on the comparison result of the danger level and the adjusted danger threshold, so as to conduct safety supervision on the campus; wherein, the calculation formula of the camera deviation is: ; In the formula, For the The deviation of the camera, is the overall road connectivity. For the The sum of the Euclidean distances between a camera and the rest of the cameras, For the The sum of the Manhattan distances between the camera and the rest of the cameras, is the number of cameras, where the overall road connectivity represents the average value of the ratio of the number of cameras with directly connected roads to each camera to the total number of cameras. By combining multiple data sources (such as camera location, frequency of dangerous behaviors, similarity of human behaviors, etc.), it is not only possible to respond to dangerous behaviors in real time, but also to improve the accuracy of campus safety supervision and early warning by adjusting monitoring rules and optimizing monitoring layout.

[0007] Furthermore, an early warning is issued based on the comparison result between the danger level and the adjusted danger threshold, including: if the danger level of the current behavior of the person to be supervised is greater than the adjusted danger threshold, an early warning is issued; if the danger level of the current behavior of the person to be supervised is equal to or less than the adjusted danger threshold, no early warning is issued, and supervision continues using the camera.

[0008] Furthermore, the danger threshold of each camera monitoring area is adjusted based on the importance of each camera, and the adjustment formula is: ; In the formula, It is Adjusted danger threshold for each camera monitoring area, is the danger threshold of the camera monitoring area, It is The importance of each camera. The importance of each camera is used to adjust the danger threshold in the monitoring area of ​​each camera, so that the greater the importance of the camera, the lower the danger threshold of its monitoring area, which can make the system more responsive and flexible.

[0009] Furthermore, the similarity between the current behavior of the person to be supervised and each historical dangerous behavior is calculated, including: collecting the current behavior of the person to be supervised, wherein the current behavior includes the current moving speed and current moving trajectory of the person to be supervised, constructing a current moving speed sequence and a current moving trajectory sequence; based on the similarity between the current moving speed sequence and the historical dangerous moving speed sequence , and the similarity between the current movement trajectory sequence and the historical dangerous movement trajectory sequence , calculate the similarity between the current behavior of the supervised person and the historical dangerous behavior , the calculation formula is: , where Represents the standard normalization function.

[0010] Furthermore, the similarity between the current behavior of the person to be supervised and each historical dangerous behavior is also related to the correlation between the historical dangerous behavior and the moving speed and the moving trajectory, and the calculation formula is: ; In the formula, is the degree of similarity between the current behavior of the supervised person and the historical risky behavior, is the mutual information between the mobile trajectory and the historical dangerous behavior, is the mutual information between moving speed and historical dangerous behavior, is the similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence, is the similarity between the current moving speed sequence and the historical dangerous moving speed sequence. This method introduces two mutual information and two similarities, which represent the relationship between the moving trajectory and dangerous behavior, the correlation between the moving speed and historical dangerous behavior, and their similarity with historical dangerous behavior. This method considers multiple dimensions of the behavior pattern (trajectory and speed) and can more comprehensively evaluate the dangerousness of the current behavior.

[0011] Furthermore, the similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence, and the similarity between the current moving speed sequence and the historical dangerous moving speed sequence are both calculated using the DTW algorithm.

[0012] Furthermore, the overall road connectivity is also related to the standard deviation of the ratio of the number of cameras with directly connected roads to each camera to the total number of cameras. The calculation formula for the overall road connectivity is: ; In the formula, is the overall road connectivity. For the The ratio of the number of cameras with direct communication paths to the total number of cameras. is the number of cameras, is the standard deviation of the ratio, The natural constant The introduction of the standard deviation enables the system to capture the uniformity and unevenness of camera connections, thereby more accurately reflecting the actual connectivity of the road network.

[0013] Furthermore, the calculation formula for the importance of the camera is: ; In the formula, For the The importance of a camera, For the The deviation of the camera, For the The frequency of dangerous behaviors in the area monitored by the cameras, is the number of cameras. For areas with high frequency of dangerous behaviors or large deviations, cameras will be given higher weights to ensure that these areas can be monitored more closely.

[0014] Furthermore, the empirical value of the danger threshold of the camera monitoring area is 0.6.

[0015] In a second aspect, a campus safety supervision and early warning system is also provided, comprising a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the campus safety supervision and early warning methods described above.

[0016] Beneficial Effects (I) Using the importance of each camera to adjust the danger threshold in the surveillance area of ​​each camera, the greater the importance of the camera, the lower the danger threshold of its surveillance area, which can make the system more responsive and flexible. In areas where dangerous behaviors occur frequently, increasing the sensitivity of the trigger threshold can more quickly identify and deal with potential dangerous behaviors, take timely intervention measures, and enhance the safety of campuses or public places.

[0017] (ii) By weighting the contribution of historical dangerous speed sequences and historical dangerous trajectory sequences to the identification of dangerous behaviors, the degree of danger of the current behavior of the person to be supervised can be accurately calculated, thereby improving the accuracy of danger identification and warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The following detailed description is read with reference to the accompanying drawings, which illustrate several embodiments of the present invention in an exemplary and non-limiting manner, and the same or corresponding reference numerals are the same or corresponding parts, wherein: Figure 1 The figure schematically shows a flow chart of a campus safety supervision and early warning method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0021] like Figure 1 As shown, S101: collecting data.

[0022] Specifically, high-definition cameras are used to collect historical dangerous behaviors of people on campus, camera location data, and the frequency of dangerous behaviors in the camera monitoring area, where the historical dangerous behaviors of people include the historical dangerous movement speeds and historical dangerous movement trajectories of people, and historical dangerous movement speed sequences and historical dangerous movement trajectory sequences are constructed based on the historical movement speeds and historical movement trajectories, respectively. The historical dangerous movement speed sequence is constructed by collecting the movement speeds from the time a person who has engaged in dangerous behavior enters the camera monitoring area to the time when the person leaves the monitoring area. Similarly, the historical dangerous movement trajectory sequence is also constructed according to the above method, where the collection frequency can be set manually, for example, once per second or once every two seconds.

[0023] For example, person A has performed dangerous behaviors in the monitoring area of ​​camera W1, such as sudden fast running, frequent turns that cause him to fall or collide with other people, or go to some dangerous areas, including construction areas and pool areas. The movement speed and movement trajectory of the above persons from entering the monitoring area to leaving the monitoring area are collected to obtain a dangerous movement speed sequence and a dangerous movement trajectory sequence.

[0024] S102: Calculate the importance of the camera.

[0025] In one embodiment, due to the complexity and diversity of buildings on campus, the installation of cameras is also diverse, resulting in uneven distribution of cameras, and the frequency of dangerous behaviors occurring in monitoring areas covered by different cameras is different, and the frequency and probability of dangerous behaviors occurring in more remote monitoring areas are higher. Therefore, the deviation of the camera can be calculated based on the location data of the camera; the importance of each camera can be calculated through the deviation of the camera and the frequency of dangerous behaviors occurring in the area monitored by the camera. For areas with greater importance, the danger threshold should be smaller when judging the occurrence of dangerous behaviors, thereby improving the accuracy of dangerous behavior identification.

[0026] In one embodiment, the calculation formula of the importance of the camera is: ; In the formula, For the The importance of a camera, The larger the The more dangerous the area monitored by a camera is, the more important the camera is. For the The deviation of the camera, is the number of cameras, For the The frequency of dangerous behaviors in the monitoring area of ​​a camera. The greater the frequency of dangerous behaviors in the monitoring area of ​​the camera, the greater the importance of the camera. The reason is that if the frequency of dangerous behaviors in the monitoring area is low or the deviation is small, the importance of the camera will be low. Accordingly, the danger threshold of these areas can be set higher in the future to reduce overreaction (false alarm). For high-risk areas, the importance of the camera is higher, and the system will automatically increase its sensitivity to reduce the risk of missed alarms.

[0027] In one embodiment, the calculation formula of the deviation of the camera is: ; In the formula, For the The deviation of the camera, is the overall road connectivity. For the The sum of the Euclidean distances between a camera and the rest of the cameras, For the The sum of the Manhattan distances between the camera and the rest of the cameras, is the number of cameras, wherein the overall road connectivity represents the average value of the ratio of the number of cameras that have roads directly connected to each camera to the total number of cameras.

[0028] It should be noted that since there are blocked roads between different cameras and the roads are not straight diagonals, the deviation of the camera cannot be accurately evaluated only by Euclidean distance. Therefore, the overall road connectivity of the campus is calculated by dividing the weights of Euclidean distance and Manhattan distance to calculate the deviation of each camera. The greater the road connectivity, the more accurate the Euclidean distance evaluation, and the smaller the connectivity, the more accurate the Manhattan distance evaluation, thereby improving the accuracy of evaluating the overall road connectivity.

[0029] In another embodiment, the overall road connectivity is not only related to the ratio of the number of cameras with directly connected roads to the total number of cameras, but also to the standard deviation of each ratio, and the size of the standard deviation represents the difference in connectivity between cameras. The larger the standard deviation, the greater the difference in connectivity between cameras, and the worse the overall road connectivity. The calculation formula for the overall road connectivity is: ; In the formula, is the overall road connectivity. For the The ratio of the number of cameras with direct communication paths to the total number of cameras. is the number of cameras, is the standard deviation of the ratio, The natural constant An exponential function with base .

[0030] In one embodiment, due to the different locations of different cameras, the probability of dangerous behavior in remote locations is usually higher. Traditional danger warnings are based on fixed danger thresholds for safety warnings, which will lead to untimely warnings of dangerous events in some areas or misjudgment of safety accidents. Therefore, it is necessary to adaptively adjust the danger threshold for dangerous behavior identification in each area according to the importance of the area where each camera is located. In addition, the more important the camera is, the lower its danger threshold should be. The reason is that the more important the camera is, the higher the probability of dangerous behavior by people. Therefore, the danger threshold should be lowered to improve the sensitivity and accuracy of the warning.

[0031] S103: Adjust the danger threshold according to the importance of the camera.

[0032] In one embodiment, Take the first camera as an example, adjust The adjustment formula for the danger threshold corresponding to the camera monitoring area is: ; In the formula, It is Adjusted danger threshold for each camera monitoring area, is the danger threshold of the camera monitoring area, It is The importance of each camera. The greater the importance, the lower the adjusted danger threshold of its monitoring area; the smaller the importance, the greater the adjusted danger threshold of its monitoring area. The empirical value of the danger threshold is 0.6. In other embodiments, the danger threshold can also be 0.5 or 0.55, etc., which can be adjusted according to the specific implementation situation. By continuously tracking the frequency of dangerous behaviors in the monitoring area, the importance of the camera can be dynamically adjusted according to the real-time situation in different time periods. As the dangerous situation in certain areas changes, the warning sensitivity is automatically adjusted to provide more accurate security precautions.

[0033] In one embodiment, by lowering the danger threshold of cameras with higher importance, the system can provide more sensitive monitoring in high-risk areas. This is critical for monitoring dangerous behavior in the area. For example, cameras with higher importance may be located in key locations or transportation hubs. The lower danger threshold after adjustment can ensure that the system can quickly detect dangerous behavior, and provide early warning and response.

[0034] Through the above steps, the importance of each camera can be obtained through the deviation of the camera and the frequency of dangerous behaviors in the area monitored by the camera, and the importance of each camera can be used to adjust the danger threshold in the area monitored by each camera. The greater the importance of the camera, the lower the adjusted danger threshold of its monitored area, which can make the system more responsive and flexible. In areas where dangerous behaviors occur frequently, increasing the sensitivity of the trigger threshold can more quickly identify and deal with potential dangerous behaviors, take timely intervention measures, and enhance the safety of campuses or public places.

[0035] In one embodiment, different areas have different dangerous behavior patterns. Dangerous behaviors of people on campus usually include sudden fast running, frequent changes of direction, abnormal deviations from trajectories, and other phenomena. Traditional safety warnings only match the current trajectory sequence with the historical dangerous trajectory sequence. Dangerous behaviors caused by abnormal deviations from trajectories can be more accurately identified through the trajectory sequence, but sudden fast running has little effect on the trajectory sequence. At this time, using the trajectory sequence for dangerous identification will cause inaccurate identification results. Therefore, weights are divided according to the contribution of historical dangerous speed sequences and historical dangerous trajectory sequences to the identification of dangerous behaviors, so that the degree of danger of the current behavior of the person to be supervised can be accurately calculated.

[0036] Specifically, the degree of danger of the current behavior of the person to be supervised is obtained based on the similarity between the current behavior of the person to be supervised and each historical dangerous behavior in the camera monitoring area. And based on the comparison result between the danger level and the adjusted danger threshold, an early warning is issued to conduct safety supervision on the campus.

[0037] S104: Calculate the dangerousness level of the current behavior of the person to be supervised.

[0038] In one embodiment, since there are multiple historical dangerous behaviors that have occurred in the monitoring area of ​​the same camera, it is necessary to calculate the average value of the similarity between the current behavior of the person to be supervised in the monitoring area of ​​the camera and each historical dangerous behavior as the dangerousness level of the current behavior of the person to be supervised.

[0039] In one embodiment, the similarity between the current behavior of the supervised person in the camera monitoring area and each historical dangerous behavior is calculated. , the calculation formula is: , where represents the standard normalization function, is the similarity between the current moving speed sequence and the historical dangerous moving speed sequence, is the similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence.

[0040] In another embodiment, when calculating the similarity between the current behavior of the person to be supervised and each historical dangerous behavior, the similarity between the current behavior of the person to be supervised and each historical dangerous behavior is also related to the correlation between the historical dangerous behavior and the moving speed and the moving trajectory, thereby providing another calculation method, and the calculation formula is: ; In the formula, is the degree of similarity between the current behavior of the supervised person and any historical dangerous behavior, is the mutual information between the mobile trajectory and the historical dangerous behavior, is the mutual information between the moving speed and the historical dangerous behavior. The size of the mutual information indicates the correlation with the trajectory or speed when identifying dangerous behaviors. The larger the mutual information, the greater the correlation, and the greater the weight of the similarity should be. is the similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence, It is the similarity between the current moving speed sequence and the historical dangerous moving speed sequence. Among them, the similarity can be calculated using the DTW algorithm (dynamic time warping). This method can provide early warning of potential dangerous behaviors by comparing the current behavior with historical dangerous events. Especially in large-scale regulatory systems, when the behavior patterns of personnel tend to be dangerous, the system can quickly identify and provide accurate assessments to reduce false alarms and missed alarms. In addition, two mutual information are introduced to represent the relationship between movement trajectory and dangerous behavior, and the correlation between movement speed and historical dangerous behavior. Multiple dimensions of the behavior pattern (trajectory and speed) are considered, which can more comprehensively evaluate the degree of danger of the current behavior.

[0041] By dividing the weights according to the contribution of historical dangerous speed sequences and historical dangerous trajectory sequences to the identification of dangerous behaviors, the dangerousness of the current behavior of the person to be supervised can be accurately calculated, thereby improving the accuracy of dangerous identification and early warning.

[0042] For example, for camera W2, in the past monitoring process of the monitoring area of ​​camera W2, four historical dangerous behaviors have occurred, and then the similarity between the current behavior of the supervised person and the four historical dangerous behaviors is calculated to obtain the similarity , similarity , similarity and similarity , and then calculate the average value as , and The value of is used as the dangerousness level of the current behavior of the person to be supervised.

[0043] S105: issuing a warning based on the comparison result between the danger level and the adjusted danger threshold.

[0044] In one embodiment, since there are multiple historical dangerous behaviors that occurred in the monitoring area of ​​the same camera, the average value of the similarity between the current behavior of the person to be supervised and all historical dangerous behaviors in the monitoring area of ​​the camera is calculated as the danger level of the current behavior of the person to be supervised. If the danger level of the current behavior of the person to be supervised is greater than the adjusted danger threshold, an early warning is issued; if the danger level of the current behavior of the person to be supervised is equal to or less than the adjusted danger threshold, no early warning is issued, and the camera is used to continue supervision, thereby realizing campus safety supervision early warning.

[0045] The present invention also provides a campus safety supervision and early warning system. The system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a campus safety supervision and early warning method according to the first aspect of the present invention is implemented.

[0046] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface, whose configuration and functions are known in the art and thus will not be described in detail here.

[0047] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions stored or otherwise maintained in such a computer-readable medium.

[0048] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0049] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0050] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent application. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A campus safety supervision and early warning method, characterized in that: include: Get the location data of each camera on campus; The deviation of each camera is calculated, and the importance of the camera is obtained based on the product of the deviation of the camera and the frequency of dangerous behavior in the camera monitoring area; the danger threshold of each camera monitoring area is adjusted based on the importance of each camera to obtain the adjusted danger threshold of each camera monitoring area, wherein the adjusted danger threshold is negatively correlated with the importance; According to the average value of the similarity between the current behavior of the supervised person and each historical dangerous behavior in the camera monitoring area, the dangerousness of the current behavior of the supervised person is obtained, and an early warning is issued based on the comparison result between the dangerousness and the adjusted dangerous threshold, so as to conduct safety supervision on the campus; The calculation formula of the camera deviation is: ; In the formula, For the The deviation of the camera, is the overall road connectivity. For the The sum of the Euclidean distances between a camera and the rest of the cameras, For the The sum of the Manhattan distances between the camera and the rest of the cameras, is the number of cameras, where the overall road connectivity represents the average value of the ratio of the number of cameras that have directly connected roads with each camera to the total number of cameras.

2. The campus safety supervision and early warning method according to claim 1 is characterized in that: Early warning based on the comparison of the risk level with the adjusted risk threshold, including: If the risk level of the current behavior of the person to be supervised is greater than the adjusted risk threshold, an early warning is issued; If the danger level of the current behavior of the person to be supervised is equal to or less than the adjusted danger threshold, no warning will be issued and supervision will continue using the camera.

3. The campus safety supervision and early warning method according to claim 1 is characterized in that: The danger threshold of each camera monitoring area is adjusted based on the importance of each camera. The adjustment formula is: ； In the formula, It is Adjusted danger threshold for each camera monitoring area, is the danger threshold of the camera monitoring area, It is The importance of a camera.

4. The campus safety supervision and early warning method according to claim 1 is characterized in that: Calculate the similarity between the current behavior of the supervised person and each historical risky behavior, including: Collecting the current behavior of the person to be supervised, wherein the current behavior includes the current moving speed and the current moving trajectory of the person to be supervised, and constructing a current moving speed sequence and a current moving trajectory sequence; Based on the similarity between the current moving speed sequence and the historical dangerous moving speed sequence , and the similarity between the current movement trajectory sequence and the historical dangerous movement trajectory sequence , calculate the similarity between the current behavior of the supervised person and the historical dangerous behavior , the calculation formula is: , where Represents the standard normalization function.

5. The campus safety supervision and early warning method according to claim 4 is characterized in that: The similarity between the current behavior of the supervised person and each historical dangerous behavior is also related to the correlation between the historical dangerous behavior and the moving speed and the moving trajectory, and the calculation formula is: ； In the formula, is the degree of similarity between the current behavior of the supervised person and the historical risky behavior, is the mutual information between the mobile trajectory and the historical dangerous behavior, is the mutual information between moving speed and historical dangerous behavior, is the similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence, It is the similarity between the current moving speed sequence and the historical dangerous moving speed sequence.

6. The campus safety supervision and early warning method according to claim 5 is characterized in that: The similarity between the current moving trajectory sequence and the historical dangerous moving trajectory sequence, and the similarity between the current moving speed sequence and the historical dangerous moving speed sequence are both calculated using the DTW algorithm.

7. The campus safety supervision and early warning method according to claim 1 is characterized in that: The overall road connectivity is also related to the standard deviation of the ratio of the number of cameras with directly connected roads to each camera to the total number of cameras. The calculation formula for the overall road connectivity is: ； In the formula, is the overall road connectivity. For the The ratio of the number of cameras with direct communication paths to the total number of cameras. is the number of cameras, is the standard deviation of the ratio, The natural constant An exponential function with base .

8. The campus safety supervision and early warning method according to claim 1 is characterized in that: The calculation formula of the importance of the camera is: ； In the formula, For the The importance of a camera, For the The deviation of the camera, For the The frequency of dangerous behaviors in the area monitored by the cameras, The number of cameras.

9. The campus safety supervision and early warning method according to claim 3 is characterized in that: The empirical value of the danger threshold of the camera monitoring area is 0.

6.

10. A campus safety supervision and early warning system, comprising a processor and a memory, characterized in that: The memory stores a computer program, and the processor executes the computer program to implement the campus safety supervision and early warning method as described in any one of claims 1-9.

Images