Campus safety monitoring and early warning solution method and device based on artificial intelligence technology

Abstract

The invention discloses a campus security monitoring and early warning solution method and device based on an artificial intelligence technology, relates to the technical field of security monitoring, and is used for solving the problems of lack of perception of scene change and personnel dynamics and delayed early warning response. The method comprises the following steps: detecting personnel walking information under monitoring equipment within marking time, analyzing image survey conditions in combination with a monitoring range, screening and sorting the monitoring equipment needing image survey, collecting image brightness distribution of the monitoring equipment, setting a light supplement level according to a sorting result, detecting light intensity in real time, and performing abnormity judgment in combination with the brightness distribution and the light supplement level. And calculating a supplementary lighting proofreading value of the abnormal equipment, judging whether early warning is performed according to the personnel passing rate, and implementing supplementary lighting processing by using the supplementary lighting proofreading value during early warning of the abnormal equipment so as to realize intelligent regulation and control of supplementary lighting response.

Description

Technical Field

[0001] The present invention relates to the field of security monitoring technology, and more specifically, to a campus security monitoring and early warning solution and device based on artificial intelligence technology. Background Art

[0002] With the advancement of smart campus construction, security management is increasingly reliant on video surveillance systems. Traditional campus video surveillance systems primarily rely on fixed-site cameras to capture and store images of areas such as teaching buildings, dormitories, and corridors to meet the needs of daily security inspections and emergency incident recovery. Some systems are equipped with fill lighting for image enhancement at night or in low-light conditions, improving image clarity and detail reproduction.

[0003] The existing technology has the following deficiencies: Currently, existing systems have a single supplementary lighting strategy, lack awareness of scene changes and personnel dynamics, and have delayed early warning responses. Therefore, a campus security monitoring and early warning solution and device based on artificial intelligence technology is proposed.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a campus safety monitoring and early warning solution and device based on artificial intelligence technology, which solves the problems raised in the above-mentioned background technology by using different product inspection methods.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: a campus safety monitoring and early warning solution and device based on artificial intelligence technology, comprising: S1: accessing the course schedule in the campus information database to mark the duration, detecting the movement information of people under each monitoring device during the marked time, and calling the monitoring range of each monitoring device; S2: Comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, select the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; S3: Set different fill light levels for the monitoring devices based on the ranking results of the monitoring devices, detect the light intensity of each monitoring device in real time, and judge the abnormality of the monitoring devices based on the image brightness distribution of the monitoring devices and the fill light level, and calculate the fill light correction value of each abnormal device; S4: Set the personnel monitoring interval and collect the personnel passing rate within the monitoring range. Determine whether to issue an early warning based on the personnel passing rate. When an abnormal device issues an early warning, use the fill light calibration value to perform fill light processing on the monitoring device.

[0007] In a preferred embodiment, the campus information database is accessed through an interface driver, and the course time is marked. The course time refers to the time when the school starts and the time when the school ends, and includes the time from the time when the school starts to the time when the school ends. The movement information of people under each monitoring device includes the change rate of the density of people within the monitoring range corresponding to each monitoring device; The number of people in each monitoring area at consecutive moments is obtained through a video image recognition algorithm. Combined with the area of ​​the corresponding monitoring area, the density of people at consecutive moments is obtained. The difference between the density of people at consecutive moments is calculated and the ratio is calculated with the continuous unit time interval to obtain the rate of change of the density of people within the monitoring range. After obtaining the rate of change of the population density within the monitoring range, the monitoring range under each monitoring device is called according to each monitoring device.

[0008] In a preferred embodiment, the change rate of the population density within the monitoring range corresponding to each monitoring device and the monitoring range are standardized and substituted into the geometric mean method to obtain the image survey coefficient.

[0009] In a preferred embodiment, the image survey coefficient is compared with a preset survey threshold. If the image survey coefficient is greater than or equal to the survey threshold, the current monitoring device is listed as a candidate device for image survey. If the image survey coefficient is less than the survey threshold, the current monitoring device is determined to be a non-priority survey device and is temporarily not included in the image survey process, while maintaining the basic monitoring function unchanged. The image survey candidate devices are statistically analyzed to obtain edge information integrity factors and color distortion deviation factors of the image survey candidate devices.

[0010] In a preferred embodiment, edge contours are extracted from the device image, and the ratio of the total number of edge points to the total pixel area of ​​the image, the average gradient strength of the edge, and the ratio of the average length to the number of contours are calculated. The values ​​are then substituted into the integrity factor weighting formula to obtain the edge information integrity factor. Perform channel-by-channel analysis on the color information in the surveillance image, extract the color difference distribution of the three color channels of the image under the standard reference brightness value, calculate the root mean square color difference of each channel, and combine it with the overall color deviation of the image and substitute it into the color distortion deviation factor weighting formula to obtain the color distortion deviation factor; The edge information integrity factor and color distortion deviation factor are standardized and substituted into the logistic regression algorithm to obtain the image ranking coefficient; The image ranking coefficient is used as the ranking index to sort all the candidate devices for image survey, and the sorting results are arranged from low to high according to the image quality judgment coefficient.

[0011] In a preferred embodiment, according to the sorting result, the image brightness distribution of each monitoring device is collected in sequence; Obtain the current monitoring image frame, extract and calculate the brightness distribution characteristics of each frame image, and obtain the image brightness distribution of each monitoring device by comparing the image brightness mean and brightness standard deviation; According to the ranking results of the candidate devices of the image survey, different fill light levels are set for the monitoring devices; Use the analytic hierarchy process to set different fill light levels for monitoring equipment. The specific steps are as follows: First, a judgment matrix is ​​constructed to compare the relative fill light levels of each monitoring device. The eigenvector of each monitoring device is calculated in the judgment matrix, and the weighted average of the eigenvectors of each monitoring device is calculated as the fill light level of the corresponding monitoring device. After obtaining the fill light level of each monitoring device, the light intensity of each monitoring device is detected in real time; The light intensity of each monitoring device is obtained by sensing the ambient light intensity of the monitoring area where each monitoring device is located through a light-sensitive sensor.

[0012] In a preferred embodiment, the fill light level of each monitoring device, the light intensity of each monitoring device, and the image brightness distribution of each monitoring device are compared with corresponding preset thresholds; The fill light level of each monitoring device is compared with the fill light level threshold. If the fill light level of the monitoring device is lower than the fill light level threshold, it is marked as a level of abnormal risk. The light intensity of each monitoring device is compared with the light intensity threshold. If the light intensity of the monitoring device is less than the light intensity threshold, it is marked as an abnormal risk level. The image brightness distribution of each monitoring device is compared with the brightness distribution threshold. If the image brightness distribution of the monitoring device is less than the brightness distribution threshold, it is marked as a level of abnormal risk; The abnormal risk level of each monitoring device is counted and compared with the preset abnormal threshold. If the abnormal risk level of the monitoring device is greater than or equal to the abnormal threshold, the monitoring device is marked as an abnormal device. If the abnormal risk level of the monitoring device is less than the abnormal threshold, the monitoring device is marked as a normal device.

[0013] In a preferred embodiment, the fill light level of each abnormal device, the light intensity of each abnormal device, and the image brightness distribution of each abnormal device are standardized and substituted into a weighted aggregation algorithm to obtain a fill light calibration value.

[0014] In a preferred embodiment, abnormal equipment is obtained and a personnel monitoring interval is set; Based on the deep learning target detection algorithm, the individual people in the image frame are identified, and the speed of people from entering to leaving the monitoring range in each interval is calculated and the average is calculated to obtain the passing rate of people in the monitoring range. The rate at which people pass through the monitoring area is compared with the preset rate threshold. If the rate at which people pass through the monitoring area is greater than or equal to the rate threshold, an early warning signal is generated and sent to the abnormal device. When an abnormal device issues an early warning, the fill light calibration value is multiplied by the preset fill light coefficient to obtain the corresponding fill light value, and fill light processing is performed based on the fill light value.

[0015] The campus security monitoring and early warning solution device based on artificial intelligence technology includes an access marking device, an image survey device, a fill light calibration device, and a fill light execution device; The access marking device is used to access the course timetable in the campus information database to mark the time, detect the movement information of people under each monitoring device during the marked time, and call the monitoring range under each monitoring device; The image survey device is used to comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, screen out the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; The fill light calibration device sets different fill light levels for the monitoring devices according to the ranking results of the monitoring devices, detects the light intensity of each monitoring device in real time, and judges the abnormality of the monitoring devices based on the image brightness distribution and fill light level of the monitoring devices and calculates the fill light calibration value of each abnormal device; The fill light execution device sets the personnel monitoring interval according to the abnormal equipment and collects the personnel passing rate within the monitoring range. It determines whether to issue an early warning based on the personnel passing rate. When the abnormal equipment issues an early warning, the fill light calibration value is used to perform fill light processing on the monitoring equipment.

[0016] Technical effects and advantages of the present invention: 1. The present invention accesses the course schedules in the campus information database and marks them. During the marked time, it detects the movement of people under the monitoring equipment and analyzes the image survey conditions based on the monitoring range. It then screens and sorts the monitoring equipment that require image survey, collects the brightness distribution of their images, sets the fill light level based on the sorting results, detects light intensity in real time, and uses the brightness distribution and fill light level to determine abnormalities. It calculates the fill light calibration value for abnormal equipment and determines whether to issue an early warning based on the personnel passing rate. When an abnormal equipment issues an early warning, it uses the fill light calibration value to implement fill light processing, thereby achieving precise image survey and scheduling, improving the quality of monitoring images, and realizing intelligent regulation of fill light response. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the method for solving the campus safety monitoring and early warning solution based on artificial intelligence technology of the present invention.

[0018] Figure 2 This is a schematic diagram of the campus safety monitoring and early warning solution device based on artificial intelligence technology of the present invention.

[0019] Figure 3 This is a method mind map of the campus safety monitoring and early warning solution based on artificial intelligence technology of the present invention; DETAILED DESCRIPTION The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] Example 1 See also Figure 1 , a campus safety monitoring and early warning solution based on artificial intelligence technology, the specific operation process is as follows: S1: Access the course schedule in the campus information database to mark the duration, detect the movement of people under each monitoring device within the marked time, and call the monitoring range of each monitoring device; Among them, the campus information database is a structured database in the campus management system, which contains student information, course schedules, attendance records, class information and schedules; Access the campus information database through the interface driver and mark the course time. Specifically, accumulate the class hours in the course time to obtain the course time and mark it as the marked time. It should be noted that interface drivers include RESTful APIs or JDBC. The course schedule is a table that lists the order of courses scheduled by the school for the current semester and the duration of each class. The course schedule duration refers to the start and end times of school on campus, and includes the time between the start and end times of school. Obtain the marking time, detect the movement information of people under each monitoring device during the marking time, and obtain the change rate of people density within the monitoring range corresponding to each monitoring device; The rate of change of the population density within the monitoring range is the numerical change in the population density within a set unit time. Its acquisition logic is to obtain the number of people in each monitoring area at consecutive moments through a video image recognition algorithm, and combine it with the area of ​​the corresponding monitoring area to obtain the population density value at consecutive moments. The difference between the population density values ​​at consecutive moments is calculated, and the ratio is calculated with the continuous unit time interval to obtain the rate of change of the population density within the monitoring range. Among them, multiple unit times are set within the marking time. The specific division and definition of unit time are obtained by the experimenters based on the personnel flow frequency evaluation results of the monitoring scene and the analysis of the recognition frame rate adaptability of the system performance and image processing algorithm, which will not be elaborated here. Specifically, the video image recognition algorithm is used to automatically identify the movement of people within the monitoring range. Optionally, a deep convolutional neural network is used to detect and classify targets in real-time image frame sequences. By identifying key points and tracking the positions of people, it is determined whether people are moving. The total number of people at each timestamp is output as the basis for calculating the density of people. After obtaining the change rate of the population density within the monitoring range, the monitoring range of each monitoring device is called according to each monitoring device; Specifically, the monitoring range is determined by the field of view parameters of the lens carried by different monitoring devices and the device installation posture information, which will not be described in detail here; S2: Comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, select the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; The change rate of population density and the monitoring range within the monitoring range of each monitoring device are standardized and substituted into the geometric mean method to obtain the image survey coefficient; It should be noted that the standardization methods include but are not limited to standard linear transformation based on interval scaling, Z-Score standardization method based on statistics, or normalization method based on nonlinear mapping function. The application methods of standardization are not described in detail here. Furthermore, in the geometric mean method, the standardized change rate of the population density within the monitoring range corresponding to each monitoring device and the monitoring range are calculated. The specific calculation formula is as follows: ; Where, is the image survey coefficient, is the standardized result of the change rate of personnel density within the monitoring range of each monitoring device. To standardize the results of monitoring scope; The image survey coefficient is compared with the preset survey threshold. If the image survey coefficient is greater than or equal to the survey threshold, the current monitoring device is listed as a candidate device for image survey. If the image survey coefficient is less than the survey threshold, the current monitoring device is determined to be a non-priority survey device and is temporarily not included in the image survey process, maintaining the basic monitoring function unchanged. It should be noted that the survey threshold was determined by the experimenters based on the distribution characteristics of historical monitoring data in the experimental scenario and the accuracy requirements for abnormal image behavior recognition, and will not be elaborated here; Counting the candidate devices for image survey, and obtaining edge information integrity factors and color distortion deviation factors of the candidate devices for image survey; The logic for obtaining the edge information integrity factor is to extract edge contours from the device image, calculate the ratio of the total number of edge points to the total pixel area of ​​the image, the average gradient strength of the edge, and the ratio of the average length to the number of contours, and substitute these into the integrity factor weighting formula to obtain the edge information integrity factor. Specifically, the weighted formula of the integrity factor is: ; Where, is the edge information completeness factor, is the total number of edge points, is the total pixel area of ​​the image, is the average gradient strength of the edge, is the average length, is the number of contours, 、 as well as is the weight coefficient, satisfying ; Specifically, a higher value of the edge information completeness factor indicates more complete edge information; The ratio of the total number of edge points to the total pixel area of ​​the image, the average gradient strength of the edge, and the ratio of the average length to the number of contours were all calculated by our experimenters based on the image feature structure modeling experimental plan and edge quality assessment standards under various scenarios, and are not detailed here. The logic for obtaining the color distortion deviation factor is to perform a channel-by-channel analysis of the color information in the surveillance image, extract the color difference distribution of the three color channels of the image under the standard reference brightness value, calculate the root mean square color difference of each channel, and combine it with the overall color deviation of the image. Substitute it into the color distortion deviation factor weighting formula to obtain the color distortion deviation factor; The root mean square value of the image channel color difference and the overall hue deviation were calculated by our researchers based on the image color restoration error benchmark model and the color distortion analysis standard for monitoring images under multiple lighting environments, and will not be elaborated here. Specifically, the weighted formula for the color distortion deviation factor can be based on weighted average, fuzzy comprehensive evaluation method, or principal component analysis method. The specific selection depends on the complexity characteristics of the monitoring equipment deployment environment and the accuracy and sensitivity requirements of the image acquisition task, and will not be elaborated here. The edge information integrity factor and color distortion deviation factor are standardized and substituted into the logistic regression algorithm to obtain the image ranking coefficient; Specifically, the standardization process has been described in the above content and will not be repeated here; Among them, the logistic regression algorithm formula is expressed as: ; Where, is the result of logistic regression calculation, that is, the image ranking coefficient, e is the natural base, and y is the linear combination term of the logistic regression model. The specific y is set as: ; Where, is the bias term, After standardization edge information completeness factor, After standardization Color distortion deviation factor, and are the edge information integrity factor and the color distortion deviation factor respectively; The image ranking coefficient is used to characterize the probability confidence value of whether the image meets the precision acquisition standard. The closer the value is to 1, the more reliable the image quality is, and the closer it is to 0, the image quality cannot be used for high-precision image survey. Furthermore, the image ranking coefficient is used as a ranking index to rank all candidate devices for image survey, and the ranking results are arranged from low to high according to the image quality judgment coefficient; According to the sorting results, the image brightness distribution of each monitoring device is collected in sequence; The image brightness distribution of each monitoring device is obtained by obtaining the current monitoring image frame, extracting and calculating the brightness distribution characteristics of each frame, and obtaining the image brightness distribution of each monitoring device by comparing the image brightness mean and brightness standard deviation. Specifically, the calculation formula for the image brightness distribution of each monitoring device is as follows: ; Where, is the image brightness distribution of each monitoring device, is the mean brightness of the image, reflecting the overall brightness level of the image. is the standard deviation of image brightness, which is used to reflect the discrete degree of brightness in the image; Among them, the brightness mean and brightness standard deviation are processed by graying the image, extracting the brightness channel, and calculating the average value of all pixels in the gray channel as , calculate all pixels in the grayscale channel relative to The standard deviation of ,Will and Substituting the above formula, we can get the image brightness distribution of each monitoring device: ; The larger the image brightness distribution value of each monitoring device, the more balanced the image brightness distribution and the higher the image quality. S3: Set different fill light levels for the monitoring devices based on the ranking results of the monitoring devices, detect the light intensity of each monitoring device in real time, and judge the abnormality of the monitoring devices based on the image brightness distribution of the monitoring devices and the fill light level, and calculate the fill light correction value of each abnormal device; According to the ranking results of the candidate devices of the image survey, different fill light levels are set for the monitoring devices; The fill light level refers to the illumination enhancement response level set for different image survey candidate devices. It is used to differentiate the illumination compensation intensity required for each device in real-time monitoring scenarios. The higher the fill light level, the less fill light is needed and the less abnormalities there are. The lower the fill light level, the more fill light is needed and the more abnormalities there are. Use the analytic hierarchy process to set different fill light levels for monitoring equipment. The specific steps are as follows: Construct a judgment matrix: First, construct a judgment matrix to compare the relative fill light levels between each monitoring device; Use a 3-point scale of 1 to 3, where 1 represents the same fill light level, 2 represents a slight fill light level difference, and 3 represents a moderate fill light level difference.

[0021]

[0022] Calculate the eigenvector: Calculate the eigenvector of each monitoring device in the judgment matrix. The eigenvector is the weighted average of each column of the judgment matrix. First, normalize each column of the judgment matrix so that the sum of each column is equal to 1. Then, for each column, calculate the weighted average as the element value of the eigenvector. For the judgment matrix above, the normalized matrix is ​​as follows:

[0023] Then, the weighted average of the feature vectors of each monitoring device is calculated as the fill light level of the corresponding monitoring device: A equipment fill light level ; B device fill light level 2972; C device fill light level .

[0024] Devices A, B, and C correspond to the three monitoring devices in this example. After sorting the image sorting coefficients according to their numerical values, they correspond to devices A, B, and C in ascending order. After obtaining the fill light level of each monitoring device, the light intensity of each monitoring device is detected in real time; The light intensity of each monitoring device is obtained by using a light sensor to sense the ambient light intensity of the monitoring area where each monitoring device is located, and obtain the light intensity of each monitoring device. Compare the fill light level of each monitoring device, the light intensity of each monitoring device, and the image brightness distribution of each monitoring device with the corresponding preset thresholds; Specifically, the fill light level of each monitoring device is compared with the fill light level threshold. If the fill light level of the monitoring device is less than the fill light level threshold, it is marked as a level of abnormal risk; The light intensity of each monitoring device is compared with the light intensity threshold. If the light intensity of the monitoring device is less than the light intensity threshold, it is marked as an abnormal risk level. The image brightness distribution of each monitoring device is compared with the brightness distribution threshold. If the image brightness distribution of the monitoring device is less than the brightness distribution threshold, it is marked as a level of abnormal risk; Count the abnormal risk levels of each monitoring device and compare it with the preset abnormal threshold. If the abnormal risk level of a monitoring device is greater than or equal to the abnormal threshold, the monitoring device will be marked as an abnormal device. If the abnormal risk level of a monitoring device is less than the abnormal threshold, the monitoring device will be marked as a normal device and no fill light is required. It should be noted that the fill light level threshold, light intensity threshold, brightness distribution threshold, and abnormality threshold were obtained by our experimenters based on the illumination compensation demand assessment model and the image acquisition reliability experimental parameter tuning plan, and will not be detailed here. The design of the number of layers for abnormal risk was determined by our researchers based on the modeling method of image degradation hierarchical response strategy and the theoretical basis of multi-parameter joint abnormal clustering. The rules for stacking the number of layers for abnormal risk are not limited and will not be elaborated here. The fill light level, light intensity, and image brightness distribution of each abnormal device are standardized and substituted into the weighted aggregation algorithm to obtain the fill light calibration value; Furthermore, in the weighted aggregation model, weight coefficients are set for the normalized fill light level of each abnormal device, the light intensity of each abnormal device, and the image brightness distribution of each abnormal device, respectively, and are expressed as 、 as well as , then the fill light correction value can be expressed by the following formula: ; Where, is the normalized fill light level of the jth abnormal device, is the normalized light intensity of the jth abnormal device, is the normalized image brightness distribution of the jth abnormal device, is the fill light calibration value corresponding to the j-th abnormal device, and j is the j-th abnormal device; Among them, each weight coefficient is set by our experimenters based on the image response priority scheduling rules and the multi-source parameter fill light feedback sensitivity test scheme, which will not be described in detail here; S4: Set the personnel monitoring interval and collect the personnel passing rate within the monitoring range. Determine whether to issue an early warning based on the personnel passing rate. When an abnormal device issues an early warning, use the fill light calibration value to fill light the monitoring device. Obtain abnormal devices and set the personnel monitoring range through cluster analysis of historical monitoring data and crowd flow fluctuation modeling results; Specifically, the monitoring interval corresponds to the abnormal equipment, and n people are set as a personnel monitoring interval; The logic for obtaining the rate of people passing through the monitoring range is to identify individual people in the image frame based on the deep learning target detection algorithm, calculate the speed of people from entering the monitoring range to leaving the monitoring range in each interval, and calculate the average to obtain the rate of people passing through the monitoring range. The rate at which people pass through the monitoring area is compared with the preset rate threshold. If the rate at which people pass through the monitoring area is greater than or equal to the rate threshold, an early warning signal is generated and sent to the abnormal device. When an abnormal device issues an early warning, the fill light calibration value is multiplied by the preset fill light coefficient to obtain the corresponding fill light value, and fill light processing is performed based on the fill light value; It should be noted that the rate threshold was obtained by our experimenters based on the changing patterns of pedestrian density in different time periods within the monitoring area and the historical traffic efficiency model, and will not be elaborated here. Example 2 See also Figure 2 , a campus security monitoring and early warning solution device based on artificial intelligence technology, including an access marking device, an image survey device, a fill light calibration device, and a fill light execution device; The access marking device is used to access the course timetable in the campus information database to mark the time, detect the movement information of people under each monitoring device during the marked time, and call the monitoring range under each monitoring device; The image survey device is used to comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, screen out the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; The fill light calibration device sets different fill light levels for the monitoring devices according to the ranking results of the monitoring devices, detects the light intensity of each monitoring device in real time, and judges the abnormality of the monitoring devices based on the image brightness distribution and fill light level of the monitoring devices and calculates the fill light calibration value of each abnormal device; The fill light execution device sets the personnel monitoring interval according to the abnormal equipment and collects the personnel passing rate within the monitoring range. It determines whether to issue an early warning based on the personnel passing rate. When the abnormal equipment issues an early warning, the fill light calibration value is used to perform fill light processing on the monitoring equipment.

[0025] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0026] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0027] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0028] In this application, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0029] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0030] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0031] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0032] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0033] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0034] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0035] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0036] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A campus safety monitoring and early warning solution based on artificial intelligence technology, characterized by: include: S1: Access the course schedule in the campus information database to mark the duration, detect the movement of people under each monitoring device within the marked time, and call the monitoring range of each monitoring device; S2: Comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, select the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; S3: Set different fill light levels for the monitoring devices based on the ranking results of the monitoring devices, detect the light intensity of each monitoring device in real time, and judge the abnormality of the monitoring devices based on the image brightness distribution of the monitoring devices and the fill light level, and calculate the fill light correction value of each abnormal device; S4: Set the personnel monitoring interval and collect the personnel passing rate within the monitoring range. Determine whether to issue an early warning based on the personnel passing rate. When an abnormal device issues an early warning, use the fill light calibration value to perform fill light processing on the monitoring device.

2. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 1 is characterized by: Access the campus information database through the interface driver and mark the course time. The course time refers to the time when the school starts and the time when the school ends, and includes the time from the start time to the end time. The movement information of people under each monitoring device includes the change rate of the density of people within the monitoring range corresponding to each monitoring device; The number of people in each monitoring area at consecutive moments is obtained through a video image recognition algorithm. Combined with the area of ​​the corresponding monitoring area, the density of people at consecutive moments is obtained. The difference between the density of people at consecutive moments is calculated and the ratio is calculated with the continuous unit time interval to obtain the rate of change of the density of people within the monitoring range. After obtaining the rate of change of the population density within the monitoring range, the monitoring range under each monitoring device is called according to each monitoring device.

3. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 2 is characterized by: The population density change rate and monitoring range within the monitoring range corresponding to each monitoring device are standardized and substituted into the geometric mean method to obtain the image survey coefficient.

4. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 3 is characterized by: The image survey coefficient is compared with the preset survey threshold. If the image survey coefficient is greater than or equal to the survey threshold, the current monitoring device is listed as a candidate device for image survey. If the image survey coefficient is less than the survey threshold, the current monitoring device is determined to be a non-priority survey device and is temporarily not included in the image survey process, maintaining the basic monitoring function unchanged. The image survey candidate devices are statistically analyzed to obtain edge information integrity factors and color distortion deviation factors of the image survey candidate devices.

5. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 4 is characterized by: Extract edge contours from device images, calculate the ratio of the total number of edge points to the total pixel area of ​​the image, the average gradient strength of the edge, and the ratio of the average length to the number of contours, and substitute these into the integrity factor weighted formula to obtain the edge information integrity factor. Perform channel-by-channel analysis on the color information in the surveillance image, extract the color difference distribution of the three color channels of the image under the standard reference brightness value, calculate the root mean square color difference of each channel, and combine it with the overall color deviation of the image and substitute it into the color distortion deviation factor weighting formula to obtain the color distortion deviation factor; The edge information integrity factor and color distortion deviation factor are standardized and substituted into the logistic regression algorithm to obtain the image ranking coefficient; The image ranking coefficient is used as the ranking index to sort all the candidate devices for image survey, and the sorting results are arranged from low to high according to the image quality judgment coefficient.

6. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 5 is characterized by: According to the sorting results, the image brightness distribution of each monitoring device is collected in sequence; Obtain the current monitoring image frame, extract and calculate the brightness distribution characteristics of each frame image, and obtain the image brightness distribution of each monitoring device by comparing the image brightness mean and brightness standard deviation; According to the ranking results of the candidate devices of the image survey, different fill light levels are set for the monitoring devices; Use the analytic hierarchy process to set different fill light levels for monitoring equipment. The specific steps are as follows: First, a judgment matrix is ​​constructed to compare the relative fill light levels of each monitoring device. The eigenvector of each monitoring device is calculated in the judgment matrix, and the weighted average of the eigenvectors of each monitoring device is calculated as the fill light level of the corresponding monitoring device. After obtaining the fill light level of each monitoring device, the light intensity of each monitoring device is detected in real time; The light intensity of each monitoring device is obtained by sensing the ambient light intensity of the monitoring area where each monitoring device is located through a light-sensitive sensor.

7. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 6 is characterized by: Compare the fill light level of each monitoring device, the light intensity of each monitoring device, and the image brightness distribution of each monitoring device with the corresponding preset threshold value; The fill light level of each monitoring device is compared with the fill light level threshold. If the fill light level of the monitoring device is lower than the fill light level threshold, it is marked as a level of abnormal risk. The light intensity of each monitoring device is compared with the light intensity threshold. If the light intensity of the monitoring device is less than the light intensity threshold, it is marked as an abnormal risk level. The image brightness distribution of each monitoring device is compared with the brightness distribution threshold. If the image brightness distribution of the monitoring device is less than the brightness distribution threshold, it is marked as a level of abnormal risk; The abnormal risk level of each monitoring device is counted and compared with the preset abnormal threshold. If the abnormal risk level of the monitoring device is greater than or equal to the abnormal threshold, the monitoring device is marked as an abnormal device. If the abnormal risk level of the monitoring device is less than the abnormal threshold, the monitoring device is marked as a normal device.

8. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 7 is characterized by: The fill light level, light intensity and image brightness distribution of each abnormal device are standardized and substituted into the weighted aggregation algorithm to obtain the fill light correction value.

9. The campus safety monitoring and early warning solution based on artificial intelligence technology according to claim 8 is characterized by: Obtain abnormal equipment and set the monitoring interval for the number of personnel; Based on the deep learning target detection algorithm, the individual people in the image frame are identified, and the speed of people from entering to leaving the monitoring range in each interval is calculated and the average is calculated to obtain the passing rate of people in the monitoring range. The rate at which people pass through the monitoring area is compared with the preset rate threshold. If the rate at which people pass through the monitoring area is greater than or equal to the rate threshold, an early warning signal is generated and sent to the abnormal device. When an abnormal device issues an early warning, the fill light calibration value is multiplied by the preset fill light coefficient to obtain the corresponding fill light value, and fill light processing is performed based on the fill light value.

10. A campus security monitoring and early warning solution device based on artificial intelligence technology, used to implement the campus security monitoring and early warning solution based on artificial intelligence technology as described in any one of claims 1 to 9, characterized in that: It includes an access marking device, an image surveying device, a fill light calibration device, and a fill light execution device; The access marking device is used to access the course timetable in the campus information database to mark it, detect the movement information of people under each monitoring device during the marked time, and call the monitoring range under each monitoring device; The image survey device is used to comprehensively analyze the movement information of people under each monitoring device and the monitoring range to determine whether to conduct image survey on the monitoring device, screen out the monitoring devices for image survey and sort them, and collect the image brightness distribution of each monitoring device after sorting; The fill light calibration device sets different fill light levels for the monitoring devices according to the ranking results of the monitoring devices, detects the light intensity of each monitoring device in real time, and judges the abnormality of the monitoring devices based on the image brightness distribution and fill light level of the monitoring devices and calculates the fill light calibration value of each abnormal device; The fill light execution device sets the personnel monitoring interval according to the abnormal equipment and collects the personnel passing rate within the monitoring range. It determines whether to issue an early warning based on the personnel passing rate. When the abnormal equipment issues an early warning, the fill light calibration value is used to perform fill light processing on the monitoring equipment.

Images