A grid-based city integrated management system and method

Abstract

The invention discloses a grid-based city comprehensive management system and method, and relates to the technical field of data management. The grid-based city comprehensive management method collects community grassroots data in a grid-based manner, comprehensively collects community grassroots management information, and provides reliable data support for model construction and anomaly detection; uses historical data to train an information classification model, automatically classifies community abnormal data, and improves classification efficiency and accuracy; quantifies the degree of data deviation according to the abnormal coefficient, and classifies the abnormal data according to set thresholds and classification rules, and gives priority to high-abnormal data; ensures that high-level abnormal data is resolved first through visual identification and sorting, and improves processing efficiency; at the same time, compares the processed grassroots data with the unprocessed grassroots data, evaluates the effect of the processing strategy, and adjusts the management strategy according to the satisfaction of the people; finally forms a closed-loop abnormal data processing mechanism, improves the level of urban management order, and enhances the happiness and satisfaction of the community people.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a grid-based city comprehensive management system and method. Background Art

[0002] As digitalization is further promoted in the management of various urban areas, most cities have promoted the modernization of urban governance through digital transformation. Due to the existence of multiple grid management lines in major cities, a grid-based urban management system has been designed, which lacks macro-top-level design and overall planning and management strategies. For example, resources are scattered and each works independently, which makes it impossible to fully mobilize effective resources for various social governance work, and it is impossible to make the current social governance work scientifically support the refined processing of abnormal data in urban areas and the orderly management of community services.

[0003] To sum up, how to orderly and finely manage abnormal data at the grassroots level of urban communities is an issue that needs to be urgently solved and optimized in the grid-based urban comprehensive management system. Summary of the invention

[0004] The present invention provides a grid-based urban comprehensive management system and method, which solves the technical problem of how to manage abnormal data at the grassroots level of various urban communities in an orderly and precise manner.

[0005] In order to solve the above technical problems, the present invention provides a grid-based urban integrated management system and method, and the specific technical solutions are as follows:

[0006] In a first aspect, a grid-based urban integrated management method comprises the following steps:

[0007] Collect various basic information data of urban areas; obtain community grassroots data sets;

[0008] Obtain the historical community grassroots data set to train and construct an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set;

[0009] According to the classified basic abnormal data items, an abnormality coefficient is obtained; an information control mechanism is set, and the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; different abnormality levels are marked, and the marked abnormality levels are sorted, and based on the sorting results, high-abnormal basic data is processed preferentially; the processed basic abnormal data is compared with the unprocessed basic abnormal data to obtain a comparison result;

[0010] The comparison results are evaluated to obtain a public satisfaction value; based on the public satisfaction value, abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

[0011] As a further optimization scheme of the present invention, a variety of basic information data of urban areas are collected to obtain community grassroots data sets, including:

[0012] Divide the city into multiple grid areas, each grid area is a jurisdiction; based on the jurisdiction, assign grassroots information collection tasks to each jurisdiction;

[0013] Establishing an electronic information file of the jurisdiction based on the collected multiple grassroots information, wherein the electronic information file of the jurisdiction contains five community basic information data types, namely, people, situations, events, places and objects;

[0014] Based on a variety of community basic information data, to form an initial community grassroots data set;

[0015] The initial community grassroots data set is preprocessed to obtain a community grassroots data set.

[0016] As a further optimization scheme of the present invention, the community grassroots data set is obtained to train and construct an information classification model, including:

[0017] The obtained community grassroots data set contains five community basic information data types: people, feelings, events, places and objects; the community grassroots data set generates a data structure, which contains a data matrix and a relationship matrix; each grassroots data item in the data matrix is ​​assigned a position, that is, a unit in the data matrix;

[0018] The relationship matrix is ​​used to represent the correlation between each basic data. The element value of the i-th row and j-th column is 1, indicating correlation, otherwise it is 0, indicating no correlation. The correlation includes that the data form attributes between the basic data are similar or located in the same management unit;

[0019] The data structure contains n data items sorted by time, each of which represents the basic data collected at the corresponding moment; that is, the t-th data point represents the basic data collected at the t-th time point;

[0020] The generated data structure is encoded into sequence data to train and construct the information classification model.

[0021] As a further optimization solution of the present invention, the information classification model includes:

[0022] The information classification model includes inputting the sequence data into the information classification model; the information classification model is composed of an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer; the intermediate data of each hidden layer is transmitted to the output layer through the layers, and finally the classification result representing the community grassroots data is output;

[0023] Based on the classification results of the historical community grassroots data set, the acquired grassroots abnormal data is input into the information classification model to output the category results of the grassroots abnormal data.

[0024] As a further optimization solution of the present invention, obtaining the base abnormality data includes:

[0025] Create a feedback account, which is established using the identity information of residents in the jurisdiction; feedback the jurisdiction abnormal information through the abnormal information window, which includes a text form window and an image form window;

[0026] The text form window collects time information, location information and abnormal situation description text information of the basic abnormal data to obtain initial basic abnormal data; the initial basic abnormal data is combined with the abnormal information feature data for fact verification to obtain first basic abnormal data;

[0027] The image form window collects base abnormal data image information, the image information includes image acquisition time information and scene information, to obtain primary base abnormal data; the primary base abnormal data is encoded by the difference between the 6×6 neighborhood of the image center pixel, through: ; To identify and extract abnormal data image features; where X i (i=1,2,…,8) is the center pixel X 0 The surrounding neighbors, I Xi is pixel X i The pixel value of the single threshold is locally defined as binarization by LBP. When the adjacent pixel is larger than the central pixel, it is encoded as 1, otherwise it is encoded as 0. Specifically, the LBP operator generates 2P different output values, corresponding to 2P different LBP values ​​of each pixel. The abnormal data image features are identified according to the different identification of the LBP values ​​of each pixel to obtain the second base abnormal data.

[0028] The first base abnormality data and the second base abnormality data are fused to obtain a base data set.

[0029] As a further optimization scheme of the present invention, the initial base abnormal data is combined with abnormal information feature data for fact verification to obtain first base abnormal data, including:

[0030] Acquire the initial base abnormal data set, and extract abnormal characters from the initial base abnormal data to obtain abnormal information feature data;

[0031] Retrieving the abnormal scene image at the corresponding time node according to the abnormal information feature data to obtain an abnormal feature index, wherein the abnormal feature index includes the location, type and abnormal degree of the image abnormal information;

[0032] The obtained abnormal characteristic indicators are passed Match with the initial base abnormal data set to obtain the matching degree; where, is the matching degree; u i is an abnormal marker point; is the matching parameter in the i-th abnormal characteristic indicator data item; n represents the number of abnormal characteristic indicator data items;

[0033] According to the matching Set the verification threshold d, when D M ≤d, and d∈(0,10);it means that the abnormal feature index has a high matching degree with the initial base abnormal data; it can be considered that it meets the abnormal data characteristics, otherwise, D M >d, the match is low or not matched;

[0034] The matching degree is used to quantify the proximity between the initial abnormal data and the characteristic data, so as to obtain a valid initial base abnormal data set, namely, the first base abnormal data.

[0035] As a further optimization scheme of the present invention, obtaining an abnormal coefficient includes:

[0036] Based on the abnormal scene image, by setting the origin of the abnormal mark point in the abnormal scene image as the center u 0 According to the matching degree, the abnormal landmark points (u 1 ,u 2 ,u 3 ,…,u i ), establish a circle with the origin as the center u 0 ;by The spatial sphere with a radius of , where h is the number of abnormal marking points; the range of the spatial sphere is the abnormal impact range;

[0037] Based on the abnormal impact range, the abnormal status of each abnormal sign point within the range is divided into five community basic information data types: people, feelings, things, places and objects; the corresponding weights of people, feelings, things, places and objects are set as {W 1 , W 2 , W 3 , W 4 , W 5}; and W 1 >W 2 >W 3 >W 4 >W 5 ;

[0038] pass To obtain the impact risk index, the impact risk index is the abnormal coefficient; where R represents the impact risk index; W i Indicates the data type of the i-th abnormal state; S i Represents the impact risk coefficient.

[0039] As a further optimization scheme of the present invention, the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; marks different abnormality levels, and sorts the marked abnormality levels, including:

[0040] The comparison result based on the corresponding weights of people, feelings, things, places and objects is W 1 >W 2 >W 3 >W 4 >W 5 ; Set the abnormal level of the basic abnormal data, the abnormal level of the basic abnormal data is , where represents the weight factor; Indicates the number of abnormal judgment items, ranging from [1,5]. When W 1 =5;W 2 , W 3 , W 4 and W 5 Decrease in sequence and are all integers;

[0041] Based on the base abnormal data set, the base abnormal data set includes a plurality of base abnormal data and their corresponding abnormal levels G; the abnormal level G includes five level identifiers of red solid dots, orange solid dots, yellow solid dots, green solid dots and blue solid dots assigned by the system;

[0042] Based on the base abnormal data and its corresponding level identification; using the level identification, the base abnormal data corresponding to the red solid circle is marked as high abnormality, the orange solid circle is marked as medium-high abnormality, the yellow solid circle is marked as medium abnormality, the green solid circle is marked as medium-low abnormality, and the blue solid circle is marked as low abnormality;

[0043] According to the abnormal state of the level identification, the base abnormal data with the red level identification is prioritized, and the corresponding base abnormal data are sorted in turn according to the orange, yellow, green and blue level identifications.

[0044] As a further optimization scheme of the present invention, the processed base abnormal data is compared with the unprocessed base abnormal data to obtain a comparison result, including:

[0045] Process the corresponding base-level abnormal data in sequence according to the level identifier; and re-mark the processed base-level abnormal data as processed;

[0046] Re-extract abnormal keywords from the processed basic abnormal data; and retrieve abnormal scene images with the basic abnormal data marked as processed time nodes;

[0047] Compare the abnormal scene image at the processed time node with the abnormal scene image at the previously unprocessed time node to obtain distinguishing feature points between the two images;

[0048] The processing result is obtained based on the number of distinguishing feature points combined with the anomaly coefficient.

[0049] As a further optimization scheme of the present invention, the comparison result is evaluated to obtain a public satisfaction value; based on the public satisfaction value, the grassroots abnormal data is managed in an orderly and meticulous manner, including:

[0050] Based on the processing results, a processing threshold is set, and the number of distinguishing feature points is combined with the abnormal coefficient through: ; To obtain the public satisfaction value; where S represents the evaluation value; α represents the satisfaction coefficient; C 1 and C 2 They represent the scene distinguishing feature weights respectively; t represents the time interval between processed and unprocessed; N represents the number of distinguishing feature points; M represents the processing threshold;

[0051] A satisfaction threshold is preset. When the evaluation value S is greater than the satisfaction threshold, the public satisfaction value is lower; when the evaluation value S is less than or equal to the satisfaction threshold, the public satisfaction value is higher, so that abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

[0052] In a second aspect, a grid city integrated management system is provided, wherein an electronic device includes a memory, a processor, and a grid city integrated management method program stored in the memory and executable on the processor, wherein the grid city integrated management method program implements the steps of the grid city integrated management method when executed by the processor, and the system includes:

[0053] Data collection module: It is used to collect various basic information data in urban areas to obtain community grassroots data sets;

[0054] Data classification module: It is used to obtain the historical community grassroots data set to train and build an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set;

[0055] Data management module: used to obtain an abnormality coefficient according to the classified basic abnormal data items; set an information control mechanism, the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; mark different abnormality levels, and sort the marked abnormality levels, based on the sorting results, give priority to processing high abnormal basic data; compare the processed basic abnormal data with the unprocessed basic abnormal data to obtain the comparison result;

[0056] Data verification module: It is used to evaluate the comparison results to obtain the public satisfaction value; based on the public satisfaction value, the abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

[0057] The present invention has at least the following beneficial effects: The present invention collects community grassroots data through gridding:

[0058] Through grid-based data collection, we can fully obtain grassroots data in the community, including various information such as demographics and community management, providing comprehensive and reliable data support for model building and anomaly detection;

[0059] Ensure that the data covers the multi-dimensional characteristics of the community and improve the accuracy of the analysis; provide a comprehensive basic data set for subsequent data processing and analysis.

[0060] The information classification model is trained using historical data sets; the model automatically classifies the input community grassroots abnormal data according to data characteristics; realizes automatic classification of data, reduces manual participation, and improves efficiency; based on the regularity of historical data, improves the accuracy and reliability of the classification of community grassroots abnormal data sets; according to the characteristics of historical community grassroots data sets, the grassroots data are refined and classified to improve the data's refined category recognition capabilities, and further optimize data analysis and application; the newly acquired grassroots abnormal data sets are then input into the information classification model, which can quickly output the category results of the grassroots abnormal data sets, realize real-time processing and classification of data, and ensure the timeliness and accuracy of data processing.

[0061] By analyzing the classification results, an abnormal coefficient is calculated for each abnormal data item; the abnormal coefficient reflects the degree to which the abnormal data deviates from the normal state; the abnormal coefficient provides a quantitative basis for the division of abnormal levels; the impact of each abnormal data is refined to avoid extensive processing.

[0062] Based on the anomaly coefficient, set thresholds and classification rules to divide the abnormal data into multiple levels; through the classification rules, the abnormal data processing is made more targeted; priority is given to highly abnormal data and resources are allocated reasonably.

[0063] Label different levels of anomalies and sort the labeled levels. Based on the sorting results, prioritize high-abnormality base data. Label the abnormal data according to the level and use visual identifiers to distinguish different levels of anomalies. Sort according to the labeled level and prioritize high-level abnormal data. Improve processing efficiency and ensure that urgent problems can be solved first. Help to intuitively display the priority of abnormal data and assist decision makers in responding quickly.

[0064] Compare the processed grassroots abnormal data with the unprocessed grassroots abnormal data to obtain the comparison results; compare and analyze the processed data with the unprocessed abnormal data to observe the differences; evaluate the effectiveness of the processing strategy to ensure that the data processing has achieved the expected results; provide a reference basis for subsequent optimization; evaluate the comparison results to obtain the public satisfaction value; provide intuitive service evaluation indicators to facilitate the monitoring of processing quality; adjust and optimize management strategies through feedback from public satisfaction; use the evaluation results of public satisfaction to further optimize the data processing process; manage abnormal data in a hierarchical manner and establish a refined abnormal processing mechanism; form a closed-loop abnormal data processing process to ensure that problems are effectively solved; improve the level of order in urban management and enhance the happiness and satisfaction of community residents. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a schematic diagram of a flow chart of a grid-based city comprehensive management method provided by an embodiment of the present invention;

[0066] Figure 2 It is a schematic diagram of a grid-based city integrated management system process provided by an embodiment of the present invention; DETAILED DESCRIPTION

[0067] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0068] This embodiment provides a grid-based city integrated management system and method, and the specific implementation methods are as follows:

[0069] like Figure 1 As shown, a grid-based urban integrated management method comprises the following steps:

[0070] Collect various basic information data of urban areas; obtain community grassroots data sets;

[0071] Obtain the historical community grassroots data set to train and construct an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set;

[0072] According to the classified basic abnormal data items, an abnormality coefficient is obtained; an information control mechanism is set, and the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; different abnormality levels are marked, and the marked abnormality levels are sorted, and based on the sorting results, high-abnormal basic data is processed preferentially; the processed basic abnormal data is compared with the unprocessed basic abnormal data to obtain a comparison result;

[0073] The comparison results are evaluated to obtain a public satisfaction value; based on the public satisfaction value, abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

[0074] In this embodiment, multi-dimensional information such as population, geography, economy, facilities, and environment of urban areas is collected through grid or digital technology; data is formed into a community grassroots data set through real-time collection, historical data integration, external system docking, etc.; comprehensive, detailed, and reliable community basic information is provided to lay the foundation for subsequent data analysis and anomaly detection; the data covers multi-dimensional features to improve the accuracy of abnormal data classification and processing.

[0075] Use historical community grassroots data as a training set to build a classification model based on machine learning or deep learning; the model forms automatic classification rules through data feature extraction and learning; the model can adapt to changes in data patterns and improve the reliability and efficiency of classification; reduce manual classification work and improve the level of classification automation; input the grassroots abnormal data collected in real time into the trained classification model; the classification model outputs the category results of each abnormal data based on the data characteristics; quickly complete the classification of grassroots abnormal data to ensure the real-time nature of data processing; clarify the nature of abnormal data to provide a basis for subsequent classification and processing.

[0076] According to the degree to which the abnormal data deviates from the normal state, the abnormal coefficient is obtained; the abnormal coefficient is generated based on the classification feature value or statistical method to reflect the degree of data abnormality; the quantitative indicators of abnormal data are refined to provide a basis for grade division; the differentiated management capabilities of abnormal data are improved; thresholds are set according to the abnormal coefficient to divide the abnormal levels (high abnormality, medium-high abnormality, medium abnormality, medium-low abnormality and low abnormality); the information control mechanism uses the threshold rules to grade and visually identify the abnormal data (red, orange, yellow, green and blue); abnormal data is processed in a graded manner to make resource allocation more targeted; high abnormal data is resolved first to improve the efficiency and effectiveness of data processing.

[0077] Label and sort abnormal data (for example, from high to low); optimize resource allocation based on the sorting results and prioritize high-level data; improve data processing efficiency and ensure that key issues are resolved quickly; provide intuitive priority display to assist decision makers in taking quick action.

[0078] Compare the processed abnormal data with the unprocessed data to analyze the processing effect; the comparison content includes changes in quantity, reduction in the degree of abnormality, etc.; evaluate the effectiveness of the abnormal data processing strategy to verify whether the expected goals are achieved; clarify the direction of improvement through difference analysis to provide a scientific basis for subsequent optimization.

[0079] Combined with the comparison results, the satisfaction of community residents with the abnormal data processing results is evaluated; the satisfaction value reflects the matching degree between the processing quality and community needs through quantitative indicators; provides intuitive service evaluation indicators, and optimizes the management and control of abnormal data processing procedures; enhances the scientific nature of decision-making through public feedback, and further adjusts management strategies; optimizes the classification, grading and processing strategies of abnormal data based on public satisfaction feedback; establishes a data closed-loop management mechanism, and dynamically adjusts the abnormal data processing process; ensures that abnormal data processing meets the actual needs of the community and improves governance efficiency; promotes the orderly development of grassroots management, and improves the quality of life and happiness of community residents.

[0080] Through the collaborative work between the above steps, a complete and orderly processing flow is formed from data collection to processing, evaluation, and feedback; thereby improving data processing efficiency, optimizing resource allocation, accurately solving community problems, improving public satisfaction, and ultimately achieving efficient and orderly management of grassroots data.

[0081] In a preferred embodiment of the present invention, the collecting of various basic information data of urban areas to obtain community grassroots data sets includes the following steps:

[0082] Step 1: Divide the city into multiple grid areas, each grid area is a jurisdiction; based on the jurisdiction, assign a grassroots information collection task to each jurisdiction;

[0083] Step 12, establishing an electronic information file of the jurisdiction based on the collected multiple grassroots information, wherein the electronic information file of the jurisdiction contains five types of community basic information data, namely, people, situations, events, places and objects;

[0084] Step 13, based on various community basic information data, to form an initial community grassroots data set;

[0085] Step 14: preprocess the initial community grassroots data set to obtain a community grassroots data set.

[0086] In the embodiment of the present invention, the main goal in step 1 is to divide the city into multiple grid communities based on the city's geographical and demographic characteristics. Specifically: clarifying the collection targets includes garbage disposal, road maintenance, community services, building information within the jurisdiction, number of rooms, personnel conditions, street problems, etc. in the urban community; determining the collection area includes obtaining the regional boundaries (urban divisions, natural geographical boundaries, etc.); using geographic information system (GIS) software, calibrating the coordinate range of the area.

[0087] Grid division includes selecting the size of the grid unit according to the purpose and required data accuracy. The unit can usually be meters, kilometers, etc.; generating the grid includes using GIS software (such as ArcGIS, QGIS, etc.) or programming languages ​​(such as Python's geopandas library) to generate the grid; determining the shape of the grid: square, rectangle, hexagon, etc.; using tools to divide the target area into several grid units to ensure that the grid seamlessly covers the entire collection area.

[0088] The collection methods include staff carrying equipment to each grid to collect data; automated equipment collection: deploying a sensor network to automatically collect data at a fixed time; remote sensing technology: remote data collection through satellites, drones and other equipment, and using automated equipment and remote sensing technology to assist manual collection to ensure the accuracy of manually collected data.

[0089] Each grid unit is checked and verified. This division is based on the size of the area, population density, and requirements for regional management. This division method ensures that each grid area can reflect the different management needs in the city and adapt to the population and resource distribution in the area. The specific division will take into account factors such as resource demand in the area, the feasibility of regional management, and population mobility; grid division allows management departments to manage the city more accurately, especially when responding to emergencies, to quickly determine the area of ​​responsibility. At the same time, this division method reduces the complexity of management and facilitates the reasonable allocation of data collection work. In addition, the definition of grid areas can avoid uneven distribution of resources and ensure that each grid area is effectively managed.

[0090] In step 12, based on the urban area grid division results in step 1, each grid will be assigned a corresponding data collection task. Specific tasks will be assigned according to the population size and characteristics of the grid area to ensure the comprehensiveness of data collection. Five categories of basic information data for each jurisdiction are collected through legal channels including community surveys, online forms, sensors, and government databases:

[0091] People: including basic personal information of residents in the jurisdiction (such as name, age, gender, occupation, family members, etc.);

[0092] Emotion: The social dynamics and emotional information of the community, including residents’ social participation, feedback, emotional state, etc.;

[0093] Events: key events or activities occurring within the jurisdiction, such as public security conditions, social service activities, public affairs, etc.;

[0094] Location: Geographic information of the jurisdiction, such as regional divisions, roads, distribution of public facilities, natural environment, etc.;

[0095] Things: physical resources within the jurisdiction, including infrastructure, public facilities, equipment, streets and houses, etc.

[0096] According to the five types of data collected, the information of each jurisdiction will be integrated into an electronic file. Each file will include data from different data sources, formats and types to ensure that the data of each jurisdiction can be uniformly stored, updated and managed; and this electronic file records the detailed information of each person in the grid area. The "community grassroots data" included includes the identity information, residential address, contact information, etc. of the personnel, and also records the information of special groups (such as the elderly, the disabled, etc.); the personnel information of each grid community is digitized and saved in the form of electronic files, which can greatly improve the efficiency of data management. In this way, the status of each grid community can be tracked in real time, making it easy to detect abnormal situations or obtain accurate data in a timely manner.

[0097] The main task of step 13 is to form a systematic "grassroots data set" based on the community grassroots data collected in step 12. This data set contains three core contents: personal identity information (such as name, ID number, etc.), neighborhood living status information (such as household head information, house type, etc.) and special group information (such as low-income groups, unemployed people, etc.). By integrating and classifying these data, a basic information data set that can represent the entire grid community is constructed; the purpose of constructing the grassroots data set is to provide a data basis for subsequent urban management and decision-making.

[0098] The five basic information data are integrated; the data integration process includes data cleaning (removing redundancy and unifying the format), data denoising (eliminating erroneous information), and data standardization (unifying data format and units, etc.); so that multiple community grassroots data types will be combined into an initial community grassroots data set. This data set will be stored in a standardized data format (such as database, Excel spreadsheet, JSON, etc.); different types of information are integrated to form a preliminary grassroots data set, providing strong data support for subsequent intelligent analysis, abnormal monitoring and decision-making;

[0099] The integrated community grassroots data in step 14 is complicated and has high duplication, and duplicate records may be caused by multiple entries, system errors, etc. These redundant information will cause distortion of data analysis results. By comparing the unique identification items of the data (such as ID, name, date, etc.), duplicate records are screened out and removed; the deduplicated data set is more streamlined, which improves data quality and reduces the computational burden.

[0100] When entering data, spelling errors can produce inconsistent data and affect data accuracy. Using natural language processing (NLP) algorithms or preset word libraries, common spelling errors can be automatically identified and corrected; correcting spelling errors can improve data reliability and ensure the accuracy of subsequent analysis.

[0101] Then outlier processing and missing value filling are performed. Outliers are extreme data points that are beyond the normal range, which may be caused by equipment failure, human input errors, etc. These data points are usually identified using statistical analysis (such as Z-score, box plot method) or machine learning algorithms (such as isolation forest, DBSCAN, etc.); after identifying outliers, you can decide whether to keep, modify or delete them according to the actual situation, so as to ensure the representativeness and accuracy of the data.

[0102] In a data set, missing values ​​can lead to inaccurate model training and analysis results. Filling methods include mean filling, median filling, regression filling, interpolation, etc. The specific choice depends on the characteristics of the data; after missing value filling, the data set is more complete, and more comprehensive and accurate results can be obtained in subsequent analysis.

[0103] The data is then converted to a unified measurement scale through standardization, usually by returning the mean to zero, normalizing the standard deviation to one, or normalizing the data to a specific range (such as 0 to 1). This step is often used to eliminate dimensional differences between data features; the standardized data set eliminates the impact of different units and ranges, which helps to ensure consistency in model training and analysis, and ultimately obtains a high-quality "community grassroots data set."

[0104] In a preferred embodiment of the present invention, the above-mentioned acquisition of the historical community grassroots data set training and construction of the information classification model includes the following steps:

[0105] Step 2, the obtained historical community grassroots data set, which contains five types of community basic information data, namely, people, feelings, events, places and objects; the historical community grassroots data set is used to generate a data structure, which contains a data matrix and a relationship matrix; each grassroots data item in the data matrix is ​​assigned a position, namely, a unit in the data matrix;

[0106] Step 21, the relationship matrix is ​​used to represent the correlation between each basic data, and the element value of the i-th row and j-th column is 1 to indicate correlation, otherwise it is 0 to indicate no correlation; the correlation includes that the data form attributes between the basic data are similar or located in the same management unit;

[0107] Step 22, the data structure includes n data items sorted by time, each item represents the basic data collected at the corresponding time; that is, the t-th data point represents the basic data collected at the t-th time point;

[0108] Step 23, encoding the generated data structure into sequence data to train and construct the information classification model.

[0109] Step 24, the information classification model includes inputting the sequence data into the information classification model; the information classification model consists of an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer; the intermediate data of each hidden layer is transmitted to the output layer through the layers, and finally the classification result representing the community grassroots data is output;

[0110] Step 25, based on the classification results of the historical community grassroots data set, the obtained grassroots abnormal data is input into the information classification model to output the category results of the grassroots abnormal data.

[0111] In the embodiment of the present invention, the information classification model in step 24 includes:

[0112] The sequence data is input into the information classification model; the information classification model includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer, the intermediate representation data of each hidden layer is transmitted to the output layer, and the output layer outputs the category recognition result representing the basic data, which is as follows:

[0113]

[0114] Where W u , W r , W c represents the weight parameter, b u , b r , b c represents the bias parameter, represents the dot product, u （t） 、r （t） and c （t） Represent the intermediate states of the first, second, and third hidden layers respectively, X （t） represents the tth data item of the base dataset, H （t） and H （t-1） They represent the intermediate representation data of the t-th and t-1-th base data sets, respectively. n ≥ t ≥ 1, where n represents the total number of input base data items. When t = 1, H （t-1）=X （t） , tanh is the hyperbolic tangent function, σ represents the sigmoid function; to obtain a trained information classification model;

[0115] Based on the output layer, the classification result of the historical basic data set is output by identifying the category of the historical basic data set.

[0116] In this embodiment, each data item X(t) in the base data set is input into the classification model. These data items may be arranged in chronological order, as mentioned in the subsequent description. (t−1) , which means the data item pushed forward one time step from the current data item X(t); the time relationship of the sequence data is preserved, and the model can be learned and inferred through time information, which is suitable for time series data analysis tasks such as prediction and classification.

[0117] The model structure is input layer, hidden layer and output layer; when the input layer receives data X(t); the first hidden layer, the second hidden layer and the third hidden layer receive the output of the previous layer in turn, and each layer is weighted by W y and bias b y Weighting and biasing are performed, and then the activation function (hyperbolic tangent function tanh) is applied to produce the intermediate state u (t) 、r (t) and c (t) The intermediate states are passed through multiple layers to gradually extract and express high-level features of the input data.

[0118] The output layer receives the output of the last hidden layer, calculates the weighted sum through weights and biases, and then applies the sigmoid function to output the classification result; multiple hidden layers can learn different levels of abstract features of the data, improving the expressiveness and classification accuracy of the model.

[0119] The intermediate state u of the last hidden layer (t) 、r (t) and c (t) It is transmitted to the output layer, the weighted sum is calculated through weights and biases, and the sigmoid function is applied to generate the classification result; through the transmission of intermediate states, the model can map the learned complex features to the final classification decision, thereby improving the accuracy and interpretability of the classification results; the multi-layer hidden layer structure allows the model to extract and combine the complex features of the input data layer by layer, enhancing the expressive power of the model.

[0120] The model can process time series data by retaining historical information H (t−1) , effectively utilizing time correlation; through the sigmoid function processing of the output layer, the model can generate accurate classification results and is suitable for multi-category classification tasks.

[0121] Based on step 24, in step 25, the obtained community-based abnormal data set is input into the information classification model to output the classification result of the community-based abnormal data set;

[0122] The output layer:

[0123] In the formula, It represents the feature representation of the community-based abnormal data set obtained. represents the intermediate representation data of the vth unit of the data matrix, M represents the set of all units of the data matrix, and W y is the weight parameter, b y is the bias parameter, representing the sigmoid function.

[0124] In this embodiment, in step 25, the community-based abnormal data set is input into the trained information classification model. This model has been trained with the previous historical data set, so it can classify the community-based abnormal data set; through the classification ability of the model, the category results of the community-based abnormal data set can be output quickly and accurately. This provides support for real-time data classification and ensures that the system can respond to changes in the community-based abnormal data set in a timely manner.

[0125] The intermediate representation data K of each cell of the data matrix v , represents the state of the vth unit.

[0126] M represents the set of the entire data matrix, which captures the characteristic information of the community-based anomaly dataset.

[0127] Using the weight parameter W y and the bias parameter b y These intermediate representation data are weighted and combined. The Sigmoid function, σ, is used to compress the output value so that it falls within the interval [0,1]. Through this processing method, the model provides category information of the community grassroots anomaly dataset.

[0128] W y and b y The weight parameters and bias parameters are used to perform linear transformation and offset on the feature data, respectively, so that the model can adjust the output results according to the weights of different features. σ represents the Sigmoid function, which is used to map the result of linear combination to [0,1] to facilitate the output of stable feature representation.

[0129] In a preferred embodiment of the present invention, the above-mentioned acquisition of base abnormal data includes the following steps:

[0130] Step 3, creating a feedback account, the feedback account is established using the identity information of the residents in the jurisdiction; feedback the jurisdiction abnormal information through the abnormal information window, the abnormal information window includes a text form window and an image form window;

[0131] Step 31, the text form window collects time information, location information and abnormal situation description text information of the grassroots abnormal data to obtain initial grassroots abnormal data; the initial grassroots abnormal data is combined with the abnormal information feature data for fact verification to obtain the first grassroots abnormal data;

[0132] Step 32, the image form window collects base abnormal data image information, the image information includes image acquisition time information and scene information, to obtain primary base abnormal data; the primary base abnormal data is encoded by the difference between the 6×6 neighborhood of the image center pixel, by:

[0133] ; To identify and extract abnormal data image features;

[0134] In the formula, X i (i=1,2,…,8) is the center pixel X 0 The surrounding neighbors, I Xi is pixel X i The pixel value of the single threshold is locally defined as binarization by LBP. When the adjacent pixel is larger than the central pixel, it is encoded as 1, otherwise it is encoded as 0. Specifically, the LBP operator generates 2P different output values, corresponding to 2P different LBP values ​​of each pixel. The abnormal data image features are identified according to the different identification of the LBP values ​​of each pixel to obtain the second base abnormal data.

[0135] Step 33: Fusing the first base abnormality data with the second base abnormality data to obtain a base data set.

[0136] In this embodiment, in step 3, a feedback account is created through the identity information of the public; in this way, abnormal information at the community level can be fed back and linked to specific individuals or groups. The feedback account will be used for user identity authentication, data tracing and other purposes to ensure the traceability and authenticity of the source of abnormal information; by associating with the public's identity information, the authenticity and effectiveness of information feedback can be improved, while ensuring that the system can understand and process feedback information from the public in a timely manner. It can also effectively avoid false feedback caused by anonymous information.

[0137] Step 31 is based on the feedback window created in step 3, and collects the basic abnormality information through the text window therein, including the time, location and description of the abnormal situation; these data form the initial basic abnormality data.

[0138] Time information: Indicates the specific time when the abnormal event occurs.

[0139] Location information: Indicates the specific location where the abnormal event occurred.

[0140] Description of abnormal situation: Describe the nature and characteristics of the abnormal event.

[0141] Then, fact verification is performed by combining the characteristic data of the abnormal information to verify the accuracy and reliability of the abnormal information, and finally the "first grassroots abnormal data" is formed; by collecting detailed text information, it is helpful to fully understand the abnormal events, and at the same time, fact verification is performed in combination with the characteristic data to reduce false alarms and omissions and ensure the accuracy of the data.

[0142] Step 32 is to collect image information of the base abnormality data through the image form window in step 3. The image information includes the time information and scene information of the image acquisition, so as to further obtain preliminary base abnormality data;

[0143] Image acquisition time information: Indicates the time when the image was captured.

[0144] Scene information: Indicates the environment or background information in the image to help analyze the specific context of abnormal events.

[0145] The local binary pattern is used to extract features from the data in the image. Specifically, the LBP algorithm generates a binary code by comparing the neighboring pixels around each pixel in the image. The coding rules are as follows: compare the central pixel with the surrounding n neighboring pixels. If the neighboring pixel value is greater than the central pixel value, it is encoded as 1, otherwise it is 0; each pixel can get a local binary pattern (LBP value), which can reflect the local texture characteristics of the image; through the output value of the LBP operator, the abnormal features in the image can be identified, and finally the second-level abnormal data is generated; image information can provide more intuitive and verifiable evidence for abnormal events. Using the LBP feature extraction method, abnormal features in the image can be effectively identified and encoded, especially in complex scenes, which helps to extract effective abnormal information, thereby improving the recognition accuracy of abnormal data.

[0146] Step 33 forms a complete base data set by fusing the text data (first base abnormal data) of step 31 and the image data (second base abnormal data) of step 32. This fusion can make comprehensive use of text descriptions and image evidence to provide more accurate and comprehensive abnormal event identification; by combining two different forms of data, text data provides background information such as time and location, and image data provides more intuitive and visual abnormal evidence. This fusion helps to comprehensively evaluate abnormal events in multiple dimensions and angles; data fusion can eliminate the limitations that may exist in a single data source. Although text data provides detailed background information, it may lack intuitiveness; while image data provides intuitive evidence, but may lack a specific description of the event. By fusing the two, the analysis of abnormal data can be made more complete and accurate, thereby better supporting subsequent processing and decision-making.

[0147] In summary, by combining different types of data (text, images), a comprehensive analysis of abnormal events can be carried out to improve the quality of data; image feature extraction can help the system extract abnormal information from complex environments more accurately and enhance the system's intelligent recognition capabilities; by integrating multiple information feedback channels (such as text and images), grassroots abnormal events can be processed and responded to more quickly; the fused data set can provide more adequate support for decision makers, helping them make more accurate judgments and decisions, thereby achieving effective heterogeneous data collection, feature extraction and fusion, and building a more intelligent and efficient grassroots abnormal information processing system, thereby improving social governance and emergency response capabilities.

[0148] In a preferred embodiment of the present invention, the above-mentioned combining the initial base abnormal data with the abnormal information characteristic data for fact verification to obtain the first base abnormal data includes the following steps:

[0149] Step 4, obtaining the initial base abnormal data set, and extracting abnormal characters from the initial base abnormal data to obtain abnormal information feature data;

[0150] Step 41, retrieve the abnormal scene image of the corresponding time node according to the abnormal information feature data to obtain an abnormal feature index, wherein the abnormal feature index includes the location, type and abnormal degree of the image abnormal information;

[0151] Step 42: The acquired abnormal characteristic index is passed through Match with the initial base abnormal data set to obtain the matching degree; where, is the matching degree; u i is an abnormal marker point; is the matching parameter in the i-th abnormal characteristic indicator data item; n represents the number of abnormal characteristic indicator data items;

[0152] Step 43: according to the matching degree Set the verification threshold d, when D M ≤d, and d∈(0,10);it means that the abnormal feature index has a high matching degree with the initial base abnormal data; it can be considered that it meets the abnormal data characteristics, otherwise, D M >d, the match is low or not matched;

[0153] Step 44: quantify the degree of proximity between the initial abnormal data and the characteristic data by using the matching degree to obtain a valid initial base abnormal data set, namely the first base abnormal data.

[0154] In this embodiment, an abnormal data set is extracted from the original data through step 4, and feature extraction is performed on the initial abnormal data, usually the "abnormal" part (such as abnormal keywords, abnormal sentences) in the original data is extracted; these abnormal characters are deviations of certain data points, or some states that do not conform to the normal pattern, such as conflicts among residents, damage to public facilities, traffic accidents, congestion of people and vehicles, etc.; by screening out abnormal data that may have problems from the initial data set, the interference of irrelevant data in subsequent analysis can be effectively reduced, and the efficiency of subsequent processing can be improved; by extracting key features from the abnormal data, it is helpful to analyze and model abnormal situations more accurately in the future.

[0155] Step 41 retrieves scene images related to the abnormal information feature data extracted in step 4. These images may come from a monitoring system, a sensor, or other image acquisition devices.

[0156] Extract abnormal feature indicators from these images, which can include:

[0157] Location: The spatial location where the anomaly occurs.

[0158] Type: The category or type of the anomaly, involving five community-based data types: people, feelings, events, places, and objects;

[0159] Abnormality: The severity or intensity of the abnormal phenomenon.

[0160] By acquiring scene images, anomalies in the data can be presented in an intuitive and visual way, helping analysts to better understand the actual impact of abnormal data; images can accurately locate the location where the anomaly occurs, which is helpful for subsequent analysis, repair or emergency response.

[0161] Step 42 is performed by matching the abnormal feature index extracted in step 4 with the data in the initial base abnormal data set in step 41. The "matching degree" can be obtained through matching; the matching degree is an indicator that measures the similarity between the abnormal feature and the initial data set. Specifically, the similarity between the two may be determined by comparing the abnormal landmark point (ui) and the matching parameters in each abnormal feature data item.

[0162] The matching parameters of the abnormal feature indicator data items are about the specific characteristics of the anomaly, such as the type, time, status, etc. of the anomaly; by calculating the matching degree, the similarity between the initial abnormal data and the feature data can be evaluated, and then it can be determined whether it is a real anomaly; the automated matching process reduces manual intervention and improves processing efficiency and accuracy.

[0163] Step 43 determines whether the matching degree of the abnormal data is high enough by setting a verification threshold (d); when the matching degree (D M ) is less than or equal to the threshold d, it means that the abnormal data has a high similarity with the characteristic data and can be considered to meet the abnormal data characteristics; when D M If it is greater than d, the match is considered to be low or non-matching; the threshold range is set to (0, 10), which means that the match is considered valid only when it is within this range; by setting a reasonable threshold, the system's sensitivity to abnormal data can be adjusted to handle abnormal situations of different types and severity; automated threshold judgment avoids excessive manual intervention, making the screening of abnormal data more accurate and efficient.

[0164] Step 44 quantifies the similarity between the initial abnormal data and the characteristic data according to the previous matching degree calculation result. M ≤d), these data are considered to be valid initial basic abnormal data sets and can be used as basic data for subsequent analysis; valid data sets usually represent the most representative or important abnormal features in the system and can provide a basis for in-depth analysis of system abnormalities; quantified data can more accurately reflect the actual abnormal situation and help with subsequent analysis, modeling or decision-making; by effectively screening out data sets that meet the abnormal characteristics, false alarms can be reduced, the misdiagnosis rate can be reduced, and the stability and credibility of the system can be improved.

[0165] By effectively extracting and identifying abnormal data from raw data, and further accurately locating and verifying anomalies through image and feature matching, a high-quality and reliable abnormal data set is finally generated. These operations are very important for applications such as anomaly detection, fault diagnosis, early warning systems, and data quality monitoring, improving the accuracy and efficiency of data analysis; providing intuitive visual support; automating the processing flow to reduce manual intervention; accurately identifying and verifying abnormal data, and enhancing the robustness of the system.

[0166] In a preferred embodiment of the present invention, the above-mentioned obtaining of the abnormal coefficient includes the following steps:

[0167] Step 5: Based on the abnormal scene image, by setting the origin of the abnormal mark point in the abnormal scene image as the center u 0 According to the matching degree, the abnormal landmark points (u 1 ,u 2 ,u 3 ,…,u i ), establish a circle with the origin as the center u 0 ;by The spatial sphere with a radius of , where h is the number of abnormal marking points; the range of the spatial sphere is the abnormal impact range;

[0168] Step 51: Based on the abnormal impact range, the abnormal status of each abnormal mark point within the range is divided into five community basic information data types: people, feelings, things, places and objects; the corresponding weights of people, feelings, things, places and objects are set as {W 1 , W 2 , W 3 , W 4 , W 5}; and W 1 >W 2 >W 3 >W 4 >W 5 ;

[0169] Step 52, pass To obtain the impact risk index, the impact risk index is the abnormal coefficient; where R represents the impact risk index; W i Indicates the data type of the i-th abnormal state; S i Represents the impact risk coefficient.

[0170] In this embodiment, in the abnormal scene image obtained in step 4, the origin of the abnormal marker point (u 0 ) is taken as the center of the circle, and other abnormal landmarks are found by using image analysis methods. According to the matching degree between the abnormal landmarks, the three-dimensional distribution of these points is constructed to form a spatial model containing all the abnormal points.

[0171] By u 0 With the center as the circle and the maximum distance as the radius r, a spherical range containing all the landmark points is established, and the boundary of the spherical space represents the impact range of the anomaly. Complex anomaly scenarios are simplified into geometric space problems, which facilitates accurate analysis of anomaly impacts in three-dimensional space. A global view of the spatial distribution of anomaly points is provided, which is helpful for subsequent risk assessment.

[0172] Step 51 extracts the specific abnormal status (such as location, characteristic value, impact degree, etc.) of each abnormal sign point within the abnormal impact range of step 5; corresponds the abnormal status of each sign point to the data type of community basic information:

[0173] People (W 1 ): Impacts related to people, such as population density and people’s activity status.

[0174] Love (W 2 ): Impacts related to the event context, such as urgency and public opinion.

[0175] Things (W 3 ): Anomalies related to specific events, such as accidents or activities.

[0176] Ground (W 4 ): Location-related information, such as community street changes and regional planning distribution.

[0177] Object (W 5 ): Exceptions related to items, such as community supplies and public resources.

[0178] According to the impact of anomalies on community information, set the weight relationship:

[0179] W 1 >W 2 >W 3 >W 4 >W 5 ; That is, people>emotions>events>places>objects;

[0180] This weight distribution ensures that more important abnormal states (such as those involving people and emotions) have higher priority in subsequent analysis; it provides a structured method to accurately attribute abnormal impacts to specific information types; the weight design highlights the priority of key abnormal points and improves the pertinence and accuracy of the analysis.

[0181] Step 52 comprehensively considers the abnormal impact of multiple factors by calculating the impact risk; the result directly quantifies the risk, which is convenient for decision-making and resource allocation;

[0182] From the global space (step 5) to the local classification (step 51), and then to the specific numerical quantification (step 52), the full-link analysis of abnormal scenarios from spatial modeling to risk assessment is realized; combined with the classification and weight allocation of five types of community basic information, the key factors affecting the abnormality are accurately located, providing a scientific basis for subsequent risk management; high efficiency and scalability: the calculation process is concise and clear, which is easy to promote and adapt in different scenarios.

[0183] In a preferred embodiment of the present invention, the information control mechanism described above obtains the abnormality level of the base abnormality data according to the abnormality coefficient; marks different abnormality levels, and sorts the marked abnormality levels, including the following steps:

[0184] Step 6: Compare the weights of people, feelings, events, places and objects to get W. 1 >W 2 >W 3 >W 4 >W 5 ; Set the abnormal level of the basic abnormal data, the abnormal level of the basic abnormal data is , where represents the weight factor; Indicates the number of abnormal judgment items, ranging from [1,5]. When W 1 =5;W 2 , W 3 , W 4 and W 5 Decrease in sequence and are all integers;

[0185] Step 61, based on the base abnormal data set, the base abnormal data set includes a plurality of base abnormal data and their corresponding abnormal levels G; the abnormal level G includes five level identifiers assigned by the system, namely, a red solid dot, an orange solid dot, a yellow solid dot, a green solid dot and a blue solid dot;

[0186] Step 62, based on the base abnormal data and its corresponding level identification; using the level identification, the base abnormal data corresponding to the red solid circle is marked as high abnormality, the orange solid circle is marked as medium-high abnormality, the yellow solid circle is marked as medium abnormality, the green solid circle is marked as medium-low abnormality, and the blue solid circle is marked as low abnormality;

[0187] Step 63 , according to the abnormal state of the level identification, the base abnormal data with the red level identification is prioritized, and the corresponding base abnormal data are sorted in turn according to the orange, yellow, green and blue level identifications.

[0188] In this embodiment, the weight distribution of people, feelings, events, places, and objects in step 6 satisfies: 1 >W 2 >W3>W4>W5, reflecting the priority attention to people and emotions; after obtaining the abnormal level G value through the abnormal level formula of the grassroots abnormal data, the grassroots abnormal data are graded according to their impact degree, and are divided into five levels from high to low, and marked with colors (red to blue, see step 61 for details); quantitative indicators are provided to facilitate the classification of different abnormalities; and it is ensured that high-weight factors have a more significant impact on the abnormal level.

[0189] Based on step 6, in step 61, the composition of the base abnormal data set includes: multiple abnormal data points and their corresponding abnormal levels G; five level identifiers are automatically assigned according to the G value:

[0190] Red solid dot: highest level anomaly.

[0191] Orange solid dot: Second-highest level anomaly.

[0192] Yellow solid dot: medium-level anomaly.

[0193] Green solid dots: lower level anomalies.

[0194] Solid blue dots: low-level anomalies.

[0195] Provides visual grading labels to facilitate quick identification and location of abnormal levels.

[0196] Based on step 6, in step 62, red: marked as high abnormality, requiring priority attention and processing; orange: marked as medium-high abnormality, lower than the red level; yellow: marked as medium abnormality, requiring regular monitoring; green: marked as medium-low abnormality, which can be processed later; blue: marked as low abnormality, which falls within the scope of routine management.

[0197] Provide clear classification strategies based on the severity of abnormal conditions to optimize the allocation of management resources; ensure that high-priority abnormal data is quickly identified and handled;

[0198] According to step 62, the basic abnormal data of red level are put in the front through step 63; they are sorted in order according to the orange, yellow, green and blue levels; the sorting basis is: red (high abnormality)>orange (medium-high abnormality)>yellow (medium abnormality)>green (medium-low abnormality)>blue (low abnormality); after sorting, the abnormal data enters the processing queue in order of priority; ensure that the high abnormal data can be identified and processed in the first time; thereby avoiding waste or delay in resource allocation; and improving the response speed and efficiency of abnormal management.

[0199] The anomaly level calculation formula combines weights and the number of anomaly judgment items to make the classification logic clear and accurate, and provide a scientific basis for high-priority anomaly data. Colors are used to identify anomaly levels to improve user identification and ease of use. High-abnormal status data is processed first to optimize resource allocation and decision-making efficiency. A standardized process is established for the handling of grassroots abnormal data, which is suitable for different scenarios. Quantitative indicators and a grading system provide strong support for subsequent risk management, emergency response and optimization of abnormal scenarios.

[0200] In a preferred embodiment of the present invention, the above-mentioned comparing the processed base abnormal data with the unprocessed base abnormal data to obtain a comparison result includes the following steps:

[0201] Step 7, processing the corresponding base-level abnormal data in sequence according to the level identifier; re-marking the processed base-level abnormal data as processed;

[0202] Step 71, re-extracting abnormal keywords from the processed base abnormal data; and retrieving abnormal scene images marked as processed time nodes in the base abnormal data;

[0203] Step 72, comparing the abnormal scene image at the processed time node with the abnormal scene image at the previously unprocessed time node to obtain distinguishing feature points between the two images;

[0204] Step 73, obtaining a processing result according to the number of distinguishing feature points and the abnormality coefficient.

[0205] In this embodiment, step 7 processes the grassroots abnormal data in sequence and re-labels them, and processes the data in sequence according to the level identifier: the abnormal data is processed in the order of red, orange, yellow, green, and blue; the processed abnormal data is marked with "processed" to facilitate the distinction between processed and unprocessed data.

[0206] By sorting the exception levels by priority, we can ensure that high-priority exceptions are handled in a timely manner; after re-marking as "processed", data closed-loop management can be achieved; the systematicness and efficiency of abnormal data processing are improved; and clear status identification is provided for subsequent analysis and tracking.

[0207] Step 71 re-extracts abnormal keywords from the "processed" basic abnormal data in step 7; the abnormal keywords represent the main features when the abnormality occurs (such as location, crowd, environment, time, etc.); according to the record of the "processed" time node in the abnormal data, the corresponding abnormal scene image is called; the abnormal scene image may include monitoring footage or other related records; the extraction of abnormal keywords can be completed through natural language processing (NLP) technology or predefined rules; the call of scene images is realized by associating timestamps with abnormal data; through the combination of keywords and scene images, data support is provided for subsequent analysis of abnormal processing effects; the image information supplements the visual details of the abnormal data, which is helpful for review and tracing.

[0208] Step 72 compares the abnormal scene image of the "processed" time node obtained in step 71 with the abnormal scene image of the "unprocessed" time node point by point; identifies the distinguishing feature points between the two images, such as object position, color change, and personnel behavior; uses image processing technology (such as edge detection and feature point matching algorithm) to achieve image comparison; the identification of distinguishing feature points is based on image pixel changes or object detection results; the degree of change of the abnormal state is obtained through image comparison technology; the comparison results provide an intuitive basis for subsequent evaluation of processing effects; image comparison enhances the visualization capability of data processing; helps to analyze the specific changes before and after processing, and provides a reference for improving processing solutions.

[0209] Step 73 uses the number of distinguishing feature points obtained in step 72 to represent the degree of change of the abnormal state in the image; evaluates the processing effect based on the abnormality level and feature point weight; the higher the processing result score, the more significant the processing effect; if the score is low, further review or improvement of the processing method may be required; the processing effect is quantified by combining image features and abnormal data; the weight factor in the formula can be flexibly adjusted according to the specific scenario; provides an objective quantitative evaluation standard for the abnormal processing effect; helps to summarize processing experience and optimize subsequent abnormal management processes.

[0210] In a preferred embodiment of the present invention, the comparison result is evaluated to obtain a public satisfaction value; based on the public satisfaction value, the grassroots abnormal data is processed in an orderly and accurate manner, including the following steps:

[0211] Step 8: Based on the processing result, set the processing threshold, and distinguish the number of feature points and the abnormal coefficient by: ; To obtain the public satisfaction value; where S represents the evaluation value; α represents the satisfaction coefficient; C 1 and C 2 They represent the scene distinguishing feature weights respectively; t represents the time interval between processed and unprocessed; N represents the number of distinguishing feature points; M represents the processing threshold;

[0212] Step 81, preset a satisfaction threshold. When the evaluation value S is greater than the satisfaction threshold, the public satisfaction value is lower; when the evaluation value S is less than or equal to the satisfaction threshold, the public satisfaction value is higher, so that the grassroots abnormal data can be processed in an orderly and accurate manner.

[0213] In this embodiment, historical data or expert experience is set as the basis in step 8 to ensure that the calculation of the evaluation value has reference significance; the effect of exception handling is comprehensively evaluated by integrating the factors of feature point changes and time intervals; the standardized measurement method of the processing effect is clarified; and the comprehensiveness and accuracy of the evaluation are improved by integrating multi-dimensional factors.

[0214] Step 81 sets a satisfaction threshold based on step 8 to define the public satisfaction level; when the evaluation value is greater than the satisfaction threshold, it means that the processing result has not met expectations and the public satisfaction value is low; when the evaluation value is less than or equal to the satisfaction threshold, it means that the processing result has met expectations and the public satisfaction value is high.

[0215] By comparing the evaluation value with the satisfaction threshold, the processing strategy is adjusted in real time:

[0216] If the evaluation value is greater than the satisfaction threshold, it is fed back to the data management step for further optimization;

[0217] If the evaluation value is ≤ the satisfaction threshold, it will be recorded as a success case.

[0218] The satisfaction threshold is set based on the statistical analysis of historical processing data or survey results; the dynamic adjustment of the evaluation value is achieved through real-time data feedback; the quantitative management of public satisfaction is achieved; and through dynamic adjustment, the accuracy and specificity of exception processing are improved.

[0219] like Figure 2 As shown, a grid city integrated management system is provided, the system is provided with an electronic device including a memory, a processor and a grid city integrated management method program stored in the memory and executable on the processor, the grid city integrated management method program implements the grid city integrated management method steps when executed by the processor, the system includes:

[0220] Data collection module 201: It is used to collect various basic information data of urban areas to obtain community grassroots data sets;

[0221] Data classification module 202: It is used to obtain the historical community grassroots data set to train and build an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set;

[0222] Data management module 203: used to obtain an abnormality coefficient according to the classified basic abnormal data items; set an information control mechanism, the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; mark different abnormality levels, and sort the marked abnormality levels, based on the sorting results, give priority to processing high abnormal basic data; compare the processed basic abnormal data with the unprocessed basic abnormal data to obtain a comparison result;

[0223] Data verification module 204: It is used to evaluate the comparison results to obtain the public satisfaction value; based on the public satisfaction value, the grassroots abnormal data can be managed in an orderly and meticulous manner.

[0224] When the functions of the above modules are implemented in the form of software functional units and used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, 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, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks or optical disks.

[0225] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general intelligent device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or system. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

Claims

1. A grid-based urban integrated management method, characterized in that: The following steps are involved: Collect various basic information data of urban areas; To obtain community-based datasets; Among them, a variety of basic information data of urban areas are collected to obtain community grassroots data sets, including: Divide the city into multiple grid areas, each grid area is a jurisdiction; based on the jurisdiction, assign grassroots information collection tasks to each jurisdiction; Establishing an electronic information file of the jurisdiction based on the collected multiple grassroots information, wherein the electronic information file of the jurisdiction contains five community basic information data types, namely, people, situations, events, places and objects; Among them, people represent the basic personal information of residents in the jurisdiction; Situation refers to the social dynamics and emotional information of the community, including residents’ social participation, feedback, and emotional state; Events refer to key incidents or activities occurring within the jurisdiction, including public security situations, social service activities, and public affairs; The land represents the geographical information of the jurisdiction, including regional divisions, roads, and natural environment; Things represent physical resources within the jurisdiction, including infrastructure and equipment; Based on a variety of community basic information data, to form an initial community grassroots data set; Preprocessing the initial community grassroots data set to obtain a community grassroots data set; Obtain the historical community grassroots data set to train and construct an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set; According to the classified basic abnormal data items, an abnormality coefficient is obtained; an information control mechanism is set, and the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; different abnormality levels are marked, and the marked abnormality levels are sorted, and based on the sorting results, high-abnormal basic data is processed preferentially; the processed basic abnormal data is compared with the unprocessed basic abnormal data to obtain a comparison result; Among them, obtaining grassroots abnormal data includes: Create a feedback account, which is established using the identity information of residents in the jurisdiction; feedback the jurisdiction abnormal information through the abnormal information window, which includes a text form window and an image form window; The text form window collects time information, location information and abnormal situation description text information of the basic abnormal data to obtain initial basic abnormal data; the initial basic abnormal data is combined with the abnormal information feature data for fact verification to obtain first basic abnormal data; The image form window collects base abnormal data image information, the image information includes image acquisition time information and scene information, to obtain primary base abnormal data; the primary base abnormal data is encoded by the difference between the 6×6 neighborhood of the image center pixel, through: ; To identify and extract abnormal data image features; Where, X i (i=1,2,…,8) are the neighboring points around the center pixel X0, I Xi is pixel X i The pixel value of the single threshold is locally defined as binarization by LBP. When the adjacent pixel is larger than the central pixel, it is encoded as 1, otherwise it is encoded as 0. Specifically, the LBP operator generates 2P different output values, corresponding to 2P different LBP values ​​of each pixel. The abnormal data image features are identified according to the different identification of the LBP values ​​of each pixel to obtain the second base abnormal data. Merging the first base abnormal data with the second base abnormal data to obtain a base data set; The comparison results are evaluated to obtain a public satisfaction value; based on the public satisfaction value, abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

2. A grid-based city comprehensive management method according to claim 1, characterized in that: Obtain the community grassroots data set described in history to train and build an information classification model, including: The obtained community grassroots data set contains five community basic information data types: people, feelings, events, places and objects; the community grassroots data set generates a data structure, which contains a data matrix and a relationship matrix; each grassroots data item in the data matrix is ​​assigned a position, that is, a unit in the data matrix; The relationship matrix is ​​used to represent the correlation between each basic data. The element value of the i-th row and j-th column is 1, indicating correlation, otherwise it is 0, indicating no correlation. The correlation includes that the data form attributes between the basic data are similar or located in the same management unit. The data structure contains n data items sorted by time, each item represents the basic data collected at the corresponding time; that is, the t-th data point represents the basic data collected at the t-th time point. The generated data structure is encoded into sequence data to train and construct the information classification model; the information classification model includes inputting the sequence data into the information classification model; the information classification model consists of an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer; the intermediate data of each hidden layer is transmitted to the output layer through the layers, and finally the classification result representing the community grassroots data is output; Based on the classification results of the historical community grassroots data set, the acquired grassroots abnormal data is input into the information classification model to output the category results of the grassroots abnormal data.

3. A grid-based city comprehensive management method according to claim 2, characterized in that: The initial base abnormal data is combined with the abnormal information feature data for fact verification to obtain first base abnormal data, including: Acquire the initial base abnormal data set, and extract abnormal characters from the initial base abnormal data to obtain abnormal information feature data; Retrieving the abnormal scene image at the corresponding time node according to the abnormal information feature data to obtain an abnormal feature index, wherein the abnormal feature index includes the location, type and abnormal degree of the image abnormal information; The obtained abnormal characteristic indicators are passed Match with the initial base abnormal data set to obtain the matching degree; where, is the matching degree; u i It is an abnormal mark point; is the matching parameter in the i-th abnormal characteristic indicator data item; n represents the number of abnormal characteristic indicator data items; According to the matching Set the verification threshold d, when D M ≤d, and d∈(0,10);it means that the abnormal feature index has a high matching degree with the initial base abnormal data; it can be considered that it meets the abnormal data characteristics, otherwise, D M >d, the match is low or non-matching; The matching degree is used to quantify the proximity between the initial abnormal data and the characteristic data, so as to obtain a valid initial base abnormal data set, namely, the first base abnormal data.

4. A grid-based city comprehensive management method according to claim 3, characterized in that: Get anomaly coefficients, including: Based on the abnormal scene image, by setting the origin of the abnormal mark point in the abnormal scene image as the circle center u0, the abnormal mark points (u1, u2, u3, ..., u4, ...) in the abnormal scene image environment are obtained in a three-dimensional form according to the matching degree. i ), establish a circle with the origin as the center u0; The spatial sphere with a radius of , where h is the number of abnormal marking points; the range of the spatial sphere is the abnormal impact range; Based on the abnormal impact range, the abnormal status of each abnormal sign point within the range is divided into five community basic information data types: people, feelings, things, places and objects; the corresponding weights of people, feelings, things, places and objects are set to {W1, W2, W3, W4, W5}; and W1>W2>W3>W4>W5; pass To obtain the impact risk index, the impact risk index is the abnormal coefficient; where R represents the impact risk index; W i Indicates the data type of the i-th abnormal state; S i Represents the impact risk coefficient.

5. A grid-based city comprehensive management method according to claim 4, characterized in that: The information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; Different anomaly levels are marked and sorted, including: Based on the weight comparison results of people, feelings, events, places and objects, W1>W2>W3>W4>W5 are obtained; the abnormal level of the basic abnormal data is set, and the abnormal level of the basic abnormal data is , where represents the weight factor; Indicates the number of abnormal judgment items, ranging from [1,5], when W1=5; W2, W3, W4 and W5 decrease in sequence and are all integers; Based on the base abnormal data set, the base abnormal data set includes a plurality of base abnormal data and their corresponding abnormal levels G; the abnormal level G includes five level identifiers of red solid dots, orange solid dots, yellow solid dots, green solid dots and blue solid dots assigned by the system; Based on the base abnormal data and its corresponding level identification; using the level identification, the base abnormal data corresponding to the red solid circle is marked as high abnormality, the orange solid circle is marked as medium-high abnormality, the yellow solid circle is marked as medium abnormality, the green solid circle is marked as medium-low abnormality, and the blue solid circle is marked as low abnormality; According to the abnormal state of the level identification, the base abnormal data with the red level identification is prioritized, and the corresponding base abnormal data are sorted in turn according to the orange, yellow, green and blue level identifications.

6. A grid-based city comprehensive management method according to claim 5, characterized in that: The processed base abnormal data is compared with the unprocessed base abnormal data to obtain comparison results, including: Process the corresponding base-level abnormal data in sequence according to the level identifier; and re-mark the processed base-level abnormal data as processed; Re-extract abnormal keywords from the processed basic abnormal data; and retrieve abnormal scene images with the basic abnormal data marked as processed time nodes; Compare the abnormal scene image at the processed time node with the abnormal scene image at the previously unprocessed time node to obtain distinguishing feature points between the two images; The processing result is obtained based on the number of distinguishing feature points combined with the anomaly coefficient.

7. A grid-based city comprehensive management method according to claim 6, characterized in that: Evaluate the comparison results to obtain a public satisfaction value; Based on the public satisfaction value, the abnormal data at the grassroots level can be managed in an orderly and meticulous manner, including: Based on the processing results, a processing threshold is set, and the number of distinguishing feature points is combined with the abnormal coefficient through: ; To obtain the public satisfaction value; where S represents the evaluation value; α represents the satisfaction coefficient; C1 and C2 represent the scene distinguishing feature weights; t represents the time interval between processed and unprocessed; N represents the number of distinguishing feature points; M represents the processing threshold; A satisfaction threshold is preset. When the evaluation value S is greater than the satisfaction threshold, the public satisfaction value is lower; when the evaluation value S is less than or equal to the satisfaction threshold, the public satisfaction value is higher, so that abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

8. A grid-based city integrated management system, characterized in that: The system is provided with an electronic device including a memory, a processor, and a grid-based city comprehensive management method program stored in the memory and executable on the processor. When the grid-based city comprehensive management method program is executed by the processor, the steps of the grid-based city comprehensive management method according to any one of claims 1 to 7 are implemented. The system includes: Data collection module: It is used to collect various basic information data in urban areas to obtain community grassroots data sets; Data classification module: It is used to obtain the historical community grassroots data set to train and build an information classification model to output the classification results of the community grassroots data; obtain the grassroots abnormal data set and input it into the information classification model, and the information classification model outputs the category results of each grassroots abnormal data item in the grassroots abnormal data set; Data management module: used to obtain an abnormality coefficient according to the classified basic abnormal data items; set an information control mechanism, the information control mechanism obtains the abnormality level of the basic abnormal data according to the abnormality coefficient; mark different abnormality levels, and sort the marked abnormality levels, based on the sorting results, give priority to processing high abnormal basic data; compare the processed basic abnormal data with the unprocessed basic abnormal data to obtain the comparison result; Data verification module: It is used to evaluate the comparison results to obtain the public satisfaction value; based on the public satisfaction value, the abnormal data at the grassroots level can be managed in an orderly and meticulous manner.

Images