Campus safety informatization management platform based on artificial intelligence

Abstract

The invention discloses a campus safety informatization management platform based on artificial intelligence, and the platform obtains real-time power consumption data with a dormitory as a unit through a data collection module, carries out the preprocessing of the real-time power consumption data through a data processing module, and generates a power consumption data matrix. Constructing a BP neural network analysis model through a data analysis module, screening abnormal data in the power consumption data matrix, determining an analysis result of the abnormal data in the row rank through constructing the analysis model, and setting an early warning pushing module to divide early warning levels according to the analysis result, according to the method, the campus electricity utilization safety information is pre-warned, the pre-warning grade division result is pushed to a dormitory manager, the dormitory manager manages, checks and analyzes the campus electricity utilization safety information through a visual management platform module, dormitories possibly using large-power electric appliances illegally are screened, the management efficiency is improved, and a guarantee is provided for dormitory electricity utilization safety management.

Description

Technical Field

[0001] The present invention relates to the field of campus safety information technology, and in particular to an artificial intelligence-based campus safety information management platform. Background Art

[0002] In recent years, information technology has developed rapidly, and big data and artificial intelligence have gradually become mainstream technologies. How to analyze and maximize the value of these huge data and apply them to various fields has become an important research topic in big data information technology at this stage. In the context of the information age, new concepts such as "smart medical care", "smart family" and "smart campus" continue to emerge. Among them, campus safety issues have always been the focus of social attention. As society continues to call for attention to the safety construction of the education industry, the application of big data information technology in school safety management has received widespread attention.

[0003] Campus electricity safety is an important part of campus safety construction. To ensure electricity safety, high-power electrical appliances are prohibited in dormitories. In particular, some electrical appliances with potential safety hazards, such as hot-steam heaters and electric blankets, will seriously affect dormitory electricity safety. At present, supervision is mostly achieved through irregular spot checks by dormitory managers. However, due to the large number of dormitories on campus and the difficulty in management, omissions are inevitable, resulting in repeated bans on high-power electrical appliances in dormitories. For this reason, we propose an artificial intelligence-based campus safety information management platform. Summary of the invention

[0004] The main purpose of the present invention is to provide a campus safety information management platform based on artificial intelligence, which is connected to the campus power supply system through a data acquisition module to obtain real-time power consumption data with dormitories as units and fixed time intervals as data acquisition cycles, and pre-process the real-time power consumption data through a data processing module to generate a power consumption data matrix, and construct a BP neural network analysis model through a data analysis module to filter abnormal data in the power consumption data matrix, and determine the analysis results of the abnormal data in the row rank by constructing an analysis model, which can effectively solve the problems in the background technology.

[0005] The technical solution adopted by the present invention is:

[0006] A campus safety information management platform based on artificial intelligence, the management platform includes the following modules:

[0007] The data collection module is connected to the campus power supply system to obtain real-time power consumption data in dormitories with fixed time intervals as data collection cycles;

[0008] A data processing module, which is connected to the data acquisition module, is used to pre-process the real-time power consumption data and generate a power consumption data matrix. Among them, m nt is the real-time power consumption of the nth dormitory in the tth data collection cycle, where n and t are both positive integers;

[0009] A data analysis module, which is connected to the data processing module and is used to obtain preprocessing results of real-time power consumption data, and to construct a BP neural network analysis model based on the acquired data, to screen abnormal data in the row rank of the power consumption data matrix M through the model, and to compare the relationship between the abnormal data in the row rank and the adjacent column rank in the power data matrix M in the same data collection period by constructing an analysis model, thereby determining the analysis result of the abnormal data in the row rank;

[0010] An early warning push module, which is connected to the data analysis module and is used to obtain the analysis results of the data analysis module, classify the early warning levels according to the analysis results, and push the early warning level classification results to dormitory management personnel;

[0011] A management platform module, which manages, views and analyzes campus electricity safety information by providing a visual management platform;

[0012] The data storage module is used to store the collected real-time power consumption data.

[0013] Furthermore, the preprocessing step of the real-time power consumption data includes:

[0014] Step 1: remove noise from the collected real-time electricity consumption data and fill in missing data;

[0015] Step 2: Perform a linear transformation on the real-time power consumption data and map the data value to the normalized value between [0,1]. The specific normalization formula is:

[0016]

[0017] Step three, cluster analysis method is used to classify the normalized data of real-time electricity consumption according to the same dormitory and the same data collection period.

[0018] Furthermore, the input sample of the BP neural network analysis model is in, It is the real-time power consumption of the nth dormitory during the Pth data collection cycle.

[0019] Furthermore, the BP neural network analysis model is a three-layer topological structure of an input layer, an output layer and a hidden layer, and the calculation formula of the hidden layer nodes is:

[0020]

[0021] Among them, h is the number of hidden layer nodes, m is the number of input layer nodes, n is the number of output layer nodes, and a is a regulation constant between 1 and 10.

[0022] Furthermore, the method for determining abnormal data is:

[0023] Step 1, setting a reasonable floating interval threshold s of the abnormal data;

[0024] Step 2: Get the real-time power consumption m of the nth dormitory in the tth data collection cycle nt ;

[0025] Step 3: Obtain the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt ';

[0026] Step 4: Calculate the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt ' and real-time power consumption m nt The difference between nt ;

[0027] Step 5, compare Δm nt and the size of s, if Δm nt If ≥s, it is judged as abnormal data, otherwise it is judged as not abnormal data.

[0028] Furthermore, the implementation steps of the analysis model are as follows:

[0029] Step 1, setting a safety interval threshold u of the analysis result;

[0030] Step 2: Obtain the abnormal data m in the row rank of the power consumption data matrix M ij , where 1≤i≤n, 1≤j≤t, and i and j are both integers;

[0031] Step 3: Obtain the abnormal data m in the power consumption data matrix M. ij Adjacent data m in the same column rank (i-v)j , ..., m (i-2)j , m (i-1)j , m (i+1)j , m (i+2)j ,...,m (i+y)j , where v and y are both positive integers, and 1≤v<i, i+y≤n;

[0032] Step 4: Calculate m (i-v)j , ..., m (i-2)j , m (i-1)j , m ij , m (i+1)j , m (i+2)j,...,m (i+y)j The standard deviation σ;

[0033] Step 5, by comparing σ with the safety interval threshold u, when σ ≥ u, the analysis result of the abnormal data is determined to be definite abnormal data, otherwise, the analysis result of the abnormal data is uncertain abnormal data.

[0034] Furthermore, the warning levels are divided as follows:

[0035] Step 1: Obtain the difference P between the determined abnormal data σ and the safety interval threshold u j where j = 1, 2, ..., n;

[0036] Step 2: Use the difference P j The numerical value creates a sample set, denoted as {P 1 , P 2 ,,,,,P n};

[0037] Step 3: Get the mean and standard deviation of the sample set, and use the mean and standard deviation to standardize the data. The standardization formula is: In this formula, z is the standard parameter, σ is the variance of the sample data, and μ is the mean of the sample data;

[0038] Step 4: After standardization is completed, use the standard parameters Adjust the numerical interval to [0,1], and use the function value of f(k) to classify the live broadcast popularity data value of this evaluation indicator. The classification mechanism is:

[0039] when When the warning level is classified as level one;

[0040] when The warning level is classified as Level 2.

[0041] Furthermore, the management platform also includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the functions of the data acquisition module, the data processing module, the data analysis module, the early warning push module, the management platform module, and the data storage module are realized.

[0042] The specific implementation steps of the technical solution adopted by the present invention are:

[0043] Step 1), the data acquisition module is connected to the campus power supply system to obtain real-time power consumption data with dormitory as the unit and fixed time interval as the data acquisition cycle;

[0044] Step 2), the data processing module obtains the real-time power consumption data collected by the data collection module, and pre-processes the real-time power consumption data, including removing noise from the collected real-time power consumption data and filling in missing data; linearly transforms the real-time power consumption data and maps the data value to the normalized value between [0,1]; uses cluster analysis to classify the normalized data of real-time power consumption according to the same dormitory and the same data collection cycle, and generates a power consumption data matrix, wherein each element in the row rank of the matrix is ​​the real-time power consumption data of the same dormitory in t data collection cycles, and each element in the column rank is the real-time power consumption data of different dormitories in the same collection cycle;

[0045] Step 3), the data analysis module obtains the preprocessing results of the real-time power consumption data, and constructs a BP neural network analysis model based on the acquired data, screens the abnormal data in the row rank of the power consumption data matrix M through the model, and compares the abnormal data in the row rank with the adjacent column rank in the power data matrix M in the same data collection period by constructing an analysis model, thereby determining the analysis result of the abnormal data in the row rank;

[0046] Step 4), the warning push module obtains the analysis results of the data analysis module, divides the warning level according to the analysis results, and pushes the warning level division results to the dormitory management staff;

[0047] Step 5), dormitory management personnel obtain visual classification results of warning levels through the management platform module, identify suspicious dormitories, and conduct targeted inspections and processing.

[0048] The present invention has the following beneficial effects:

[0049] 1) The technical solution of the present invention is connected to the campus power supply system through a data acquisition module to obtain real-time power consumption data with dormitories as units and fixed time intervals as data acquisition cycles, and the real-time power consumption data is preprocessed by a data processing module to generate a power consumption data matrix, and a BP neural network analysis model is constructed by a data analysis module to screen abnormal data in the power consumption data matrix, and the analysis results of the abnormal data in the row rank are determined by constructing an analysis model, and an early warning push module is set to divide the early warning level according to the analysis results, and the early warning level division results are pushed to the dormitory management personnel. The dormitory management personnel manage, view and analyze the campus power safety information through a visual management platform module, screen dormitories that may have illegal use of high-power electrical appliances, improve management efficiency, and provide protection for dormitory power safety management;

[0050] 2) The technical solution of the present invention obtains abnormal data through a data analysis module, which can help dormitory managers to make auxiliary judgments on whether there is any behavior of forgetting to turn off electrical equipment in the dormitory, so as to facilitate the managers to manage the electricity usage in the dormitory. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is an overall structural block diagram of an artificial intelligence-based campus safety information management platform of the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0053] The specific implementation steps of one embodiment of the solution of the present invention are:

[0054] Step 1: The data collection module is connected to the campus power supply system to obtain real-time power consumption data with dormitory as the unit and fixed time interval as the data collection cycle;

[0055] Step 2: The data processing module obtains the real-time power consumption data collected by the data collection module and pre-processes the real-time power consumption data, including

[0056] 1. Remove noise from the collected real-time electricity consumption data and fill in missing data;

[0057] 2. Perform a linear transformation on the real-time power consumption data and map the data value to the normalized value between [0,1]. The specific normalization formula is:

[0058]

[0059] 3. Use cluster analysis to classify the normalized data of real-time power consumption according to the same dormitory and the same data collection period, and generate a power consumption data matrix Among them, each element in the row rank of the matrix is ​​the real-time power consumption data of the same dormitory in t data collection cycles, and each element in the column rank is the real-time power consumption data of different dormitories in the same collection cycle. It should be noted that the total duration of the t data collection cycles in the power data matrix M does not exceed one day, that is, the data in the row rank represents the real-time power consumption data of the same dormitory in t data collection cycles within one day;

[0060] Step 3: The data analysis module obtains the preprocessing results of the real-time power consumption data, and constructs a BP neural network analysis model based on the obtained data. The BP neural network analysis model is a three-layer topology structure of input layer, output layer and hidden layer. The input sample of the BP neural network analysis model is in, The real-time power consumption of the nth dormitory in the Pth data collection cycle in the tth data collection cycle is obtained. The BP neural network analysis model can be constructed through Matlab. In Matlab, newff is used to create a function, and the training parameters of the network are set. The network is trained. After the training is successful, the analysis results are output. The output value of the analysis result is the predicted value of the real-time power consumption data of each dormitory in t data collection cycles on the same day. After obtaining the predicted value result of the real-time power consumption data, the abnormal data in the predicted value is filtered out. The specific method is:

[0061] Set a reasonable floating interval threshold s for the abnormal data;

[0062] Get the real-time power consumption m of the nth dormitory in the tth data collection cycle nt ;

[0063] Get the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt ';

[0064] Calculate the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt ' and real-time power consumption m nt The difference between nt ;

[0065] Compare Δm nt and the size of s, if Δm nt ≥s, it is judged as abnormal data, otherwise it is judged as not abnormal data;

[0066] After the abnormal data is initially screened out, since the real-time power consumption will be interfered by objective factors, in order to make the judgment result of the abnormal data more in line with the actual situation, the dormitory that generates the abnormal data and the adjacent dormitory are further analyzed by building an analysis model. The specific method is as follows:

[0067] Set the safety interval threshold u of the analysis result;

[0068] Get the abnormal data m in the row rank of the power consumption data matrix M ij , where 1≤i≤n, 1≤j≤t, and i and j are both integers;

[0069] Get the abnormal data m in the power consumption data matrix M ij Adjacent data m in the same column rank(i-v)j , ..., m (i-2)j , m (i-1)j , m (i+1)j , m (i+2)j ,...,m (i+y)j , where v and y are both positive integers, and 1≤v<i, i+y≤n;

[0070] Calculate m (i-v)j , ..., m (i-2)j , m (i-1)j , m ij , m (i+1)j , m (i+2)j ,...,m (i+y)j The standard deviation σ;

[0071] By comparing σ with the safety interval threshold u, when σ ≥ u, the analysis result of the abnormal data is determined to be definite abnormal data, otherwise, the analysis result of the abnormal data is uncertain abnormal data;

[0072] By comparing and analyzing the dormitory that generates abnormal data with the adjacent dormitories, it can be determined that when σ ≥ u, it means that there is a large difference between the real-time power consumption of the dormitory that generates abnormal data in the corresponding sampling period and the real-time power consumption of the adjacent dormitories in the same period, which can indicate that the dormitory may have violated the regulations in using high-power electrical appliances in this sampling period, resulting in a sudden increase in its real-time power consumption;

[0073] In order to further reduce the interference of objective factors, we further process the analysis results, obtain the analysis results of the data analysis module through the early warning push module, and divide the early warning level according to the analysis results. The specific method is as follows:

[0074] Get the difference P between the determined abnormal data σ and the safety interval threshold u j where j = 1, 2, ..., n;

[0075] Using the difference P j The numerical value creates a sample set, denoted as {P 1 , P 2 ,,,,,P n};

[0076] Get the mean and standard deviation of the sample set, and use the mean and standard deviation to standardize the data. The standardization formula is: In this formula, z is the standard parameter, σ is the variance of the sample data, and μ is the mean of the sample data;

[0077] After standardization is completed, the standard parameters are used Adjust the numerical interval to [0,1], and use the function value of f(k) to classify the live broadcast popularity data value of this evaluation indicator. The classification mechanism is:

[0078] when When the warning level is classified as level one;

[0079] when When the warning level is classified as Level 2;

[0080] Step 4: By dividing the analysis results into warning levels, the results of different warning levels can be selectively pushed to dormitory managers through the warning push module. Dormitory managers obtain the visual division results of the warning levels through the management platform module, determine suspicious dormitories, and conduct targeted inspections and processing.

[0081] Through the above scheme, the data analysis module constructs a BP neural network analysis model, and uses the model to screen abnormal data. First, the dormitories that may use high-power electrical appliances are checked, and then the abnormal data are further analyzed by building an analysis model. By comparing the dormitories that generate abnormal data with adjacent dormitories and determining the target dormitory, the interference of objective factors can be reduced, making the analysis results closer to the actual situation.

[0082] The specific implementation steps of another embodiment of the solution of the present invention are:

[0083] Step 1: The data collection module is connected to the campus power supply system to obtain real-time power consumption data with dormitory as the unit and fixed time interval as the data collection cycle;

[0084] Step 2: The data processing module obtains the real-time power consumption data collected by the data collection module and pre-processes the real-time power consumption data, including

[0085] 1. Remove noise from the collected real-time electricity consumption data and fill in missing data;

[0086] 2. Perform linear transformation on the real-time power consumption data and map the data values ​​to the normalized value between [0,1];

[0087] 3. Use cluster analysis to classify the normalized data of real-time power consumption according to the same dormitory and the same data collection period, and generate a power consumption data matrix, where each element in the row rank of the matrix is ​​the real-time power consumption data of the same dormitory in t data collection periods, and each element in the column rank is the real-time power consumption data of different dormitories in the same collection period. It should be noted that the total duration of the t data collection periods in the power data matrix M does not exceed one day, that is, the data in the row rank represents the real-time power consumption data of the same dormitory in t data collection periods within one day;

[0088] Step 3, the data analysis module obtains the preprocessing results of the real-time electricity consumption data, and constructs a BP neural network analysis model according to the obtained data, wherein the BP neural network analysis model is a three-layer topological structure of an input layer, an output layer and a hidden layer, and the input sample of the BP neural network analysis model is, wherein, is the real-time electricity consumption of the nth dormitory for the Pth time in the tth data collection cycle, and the BP neural network analysis model can be constructed through Mat l ab, and the newff function is used in Mat l ab to create a function, and the training parameters of the network are set to train the network. After the training is successful, the analysis result is output, wherein the output value of the analysis result is the real-time electricity consumption data prediction value of each dormitory in t data collection cycles on the same day, and after obtaining the real-time electricity consumption data prediction value result, the abnormal data in the prediction value is filtered;

[0089] In step 4, dormitory managers obtain abnormal data through the management platform module, and by obtaining the class schedule information of students in each dormitory, they can make an auxiliary judgment on whether there is any behavior of forgetting to turn off the electrical equipment in the dormitory. If the dormitory has classes scheduled during the tth data collection cycle but still generates electricity consumption, it can be preliminarily judged that the dormitory may have forgotten to turn off the electrical equipment in the dormitory, so that the management personnel can manage the electricity consumption in the dormitory.

[0090] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A campus safety information management platform based on artificial intelligence, characterized by: The management platform includes the following modules: The data collection module is connected to the campus power supply system to obtain real-time power consumption data in dormitories with fixed time intervals as data collection cycles; A data processing module, which is connected to the data acquisition module, is used to pre-process the real-time power consumption data and generate a power consumption data matrix. Among them, m nt is the real-time power consumption of the nth dormitory in the tth data collection cycle, where n and t are both positive integers; A data analysis module, which is connected to the data processing module and is used to obtain preprocessing results of real-time power consumption data, and to construct a BP neural network analysis model based on the acquired data, to screen abnormal data in the row rank of the power consumption data matrix M through the model, and to compare the relationship between the abnormal data in the row rank and the adjacent column rank in the power data matrix M in the same data collection period by constructing an analysis model, thereby determining the analysis result of the abnormal data in the row rank; An early warning push module, which is connected to the data analysis module and is used to obtain the analysis results of the data analysis module, classify the early warning levels according to the analysis results, and push the early warning level classification results to dormitory management personnel; A management platform module, which manages, views and analyzes campus electricity safety information by providing a visual management platform; The data storage module is used to store the collected real-time power consumption data.

2. According to the artificial intelligence-based campus safety information management platform of claim 1, it is characterized by: The preprocessing step of the real-time power consumption data includes: Step 1: remove noise from the collected real-time electricity consumption data and fill in missing data; Step 2: Perform a linear transformation on the real-time power consumption data and map the data value to the normalized value between [0, 1]; Step three, cluster analysis method is used to classify the normalized data of real-time electricity consumption according to the same dormitory and the same data collection period.

3. According to the campus safety information management platform based on artificial intelligence according to claim 1, it is characterized by: The input sample of the BP neural network analysis model is in, It is the real-time electricity consumption of the nth dormitory during the Pth data collection cycle. The total duration of t data collection cycles does not exceed one day.

4. According to the artificial intelligence-based campus safety information management platform of claim 1, it is characterized by: The BP neural network analysis model is a three-layer topological structure of input layer, output layer and hidden layer. The calculation formula of the hidden layer nodes is: Among them, h is the number of hidden layer nodes, m is the number of input layer nodes, n is the number of output layer nodes, and a is a regulation constant between 1 and 10.

5. According to the artificial intelligence-based campus safety information management platform of claim 1, it is characterized by: The method for determining the abnormal data is as follows: Step 1, setting a reasonable floating interval threshold s of the abnormal data; Step 2: Get the real-time power consumption m of the nth dormitory in the tth data collection cycle nt ; Step 3: Obtain the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt '; Step 4: Calculate the real-time power consumption forecast value m of the nth dormitory in the tth data collection cycle nt ' and real-time power consumption m nt The difference between nt ; Step 5, compare Δm nt and the size of s, if Δm nt If ≥s, it is judged as abnormal data, otherwise it is judged as not abnormal data.

6. According to the artificial intelligence-based campus safety information management platform of claim 1, it is characterized by: The implementation steps of the analysis model are as follows: Step 1, setting a safety interval threshold u of the analysis result; Step 2: Obtain the abnormal data m in the row rank of the power consumption data matrix M ij , where 1≤i≤n, 1≤j≤t, and i and j are both integers; Step 3: Obtain the abnormal data m in the power consumption data matrix M. ij Adjacent data m in the same column rank (i-v)j , ..., m (i-2)j , m (i-1)j , m (i+1)j , m (i+2)j ,...,m (i+y)j , where v and y are both positive integers, and 1≤v<i, i+y≤n; Step 4: Calculate m (i-v)j , ..., m (i-2)j , m (i-1)j , m ij , m (i+1)j , m (i+2)j ,...,m (i+y)j The standard deviation σ; Step 5, by comparing σ with the safety interval threshold u, when σ ≥ u, the analysis result of the abnormal data is determined to be definite abnormal data, otherwise, the analysis result of the abnormal data is uncertain abnormal data.

7. According to the artificial intelligence-based campus safety information management platform of claim 1, it is characterized by: The warning levels are divided as follows: Step 1: Obtain the difference P between the determined abnormal data σ and the safety interval threshold u j where j = 1, 2, ..., n; Step 2: Use the difference P j The sample set is created by the numerical value, denoted as {P1, P2,,,, P n }; Step 3: Get the mean and standard deviation of the sample set, and use the mean and standard deviation to standardize the data. The standardization formula is: In this formula, z is the standard parameter, σ is the variance of the sample data, and μ is the mean of the sample data; Step 4: After standardization is completed, use the standard parameters Adjust the numerical interval to [0,1], and use the function value of f(k) to classify the live broadcast popularity data value of this evaluation indicator. The classification mechanism is: when When the warning level is classified as level one; when The warning level is classified as Level 2.

8. The campus safety information management platform based on artificial intelligence according to claim 1 is characterized by: The management platform also includes a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the function of any one of the modules when executing the program.

Images