Data processing method and system for smart city and storage medium

Abstract

The invention relates to the technical field of data processing, and discloses a data processing method and system for a smart city and a storage medium. The method comprises the following steps: acquiring city multi-source data through a distributed sensor network, and performing data cleaning and standardization processing to obtain standardized city data; establishing a domain concept system for semantic mapping to obtain semantic association data; analyzing the spatial-temporal characteristic and regional interaction relationship to obtain dynamic characteristic data; constructing a multi-level abnormal feature library for event recognition to obtain abnormal mode data; carrying out distributed optimization decision making to obtain management decision data; and establishing an evaluation system for strategy optimization to obtain a city management scheme. According to the method, efficient processing and intelligent analysis of the city data are realized, specifically, through a distributed computing architecture and a multi-level data processing method, the problems of low data processing efficiency, insufficient data association analysis, low computing resource utilization rate and the like are solved, and thus the overall performance of smart city data processing is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a data processing method, system and storage medium for smart cities. Background Art

[0002] In the process of smart city construction, urban data processing technology is developing rapidly. The existing urban data processing methods mainly adopt a centralized architecture, which deploys large-scale data centers to uniformly collect and process data from various fields in the city. These data include information on transportation, environment, energy, security and other aspects. The data processing process involves multiple links such as data cleaning, data standardization, data analysis, and data mining. The current processing solution has realized basic data collection and analysis functions, and can provide certain data support for urban management.

[0003] However, existing urban data processing methods have some obvious shortcomings. First, the centralized architecture is difficult to cope with the rapid growth of data volume and the diversification of processing requirements, and is prone to system response delays and resource bottlenecks when facing emergencies; second, urban data from different sources lack unified semantic understanding and association analysis, which makes it difficult to accurately grasp the relationship between data and reduces the accuracy of data analysis; third, existing methods lack flexible task scheduling and resource allocation mechanisms when dealing with complex scenarios, and cannot fully utilize distributed computing resources; finally, the evaluation of system performance and service quality lacks a comprehensive indicator system, which makes it difficult to accurately reflect the system operation status and service effects. Summary of the invention

[0004] The present application provides a data processing method, system and storage medium for smart cities, which are used to achieve efficient processing and intelligent analysis of urban data. Specifically, it solves problems such as low data processing efficiency, insufficient data correlation analysis, and low computing resource utilization through a distributed computing architecture and a multi-level data processing method, thereby improving the overall performance of smart city data processing.

[0005] In a first aspect, the present application provides a data processing method for a smart city, the data processing method for a smart city comprising: collecting multi-source urban data through a distributed Internet of Things sensor network, cleaning the original data according to data quality assessment rules, and performing format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data; establishing an urban domain concept system based on the standardized urban data, mapping and matching data entities from different sources, combining multimodal information for knowledge representation and reasoning to obtain semantically associated data; using the semantically associated data to perform periodicity and trend analysis on time series, combining spatial location The regional features are extracted based on the location information, and a model of interaction relationships between regions is established to obtain urban dynamic feature data; based on the urban dynamic feature data, point-level, sequence-level and group-level abnormal feature libraries are constructed, and abnormal events are identified through multi-dimensional feature comparison and comprehensive evaluation to obtain abnormal pattern data; based on the abnormal pattern data, computing resources are allocated under the edge computing architecture, and distributed optimization decisions are made through multi-node collaborative computing to obtain urban management decision data; based on the urban management decision data, an evaluation index system for system performance, service quality and user satisfaction is established, and the decision-making strategy is dynamically adjusted and continuously optimized to obtain an optimized urban system management plan.

[0006] In a second aspect, the present application provides a data processing system for a smart city, the data processing system for a smart city comprising:

[0007] The acquisition module is used to collect multi-source urban data through a distributed IoT sensor network, clean the original data according to data quality assessment rules, and perform format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data;

[0008] A matching module is used to establish a city domain concept system based on the standardized city data, map and match data entities from different sources, and perform knowledge representation and reasoning in combination with multimodal information to obtain semantically related data;

[0009] An analysis module is used to use the semantic association data to perform periodicity and trend analysis on the time series, extract regional features in combination with spatial location information, establish an interaction relationship model between regions, and obtain urban dynamic feature data;

[0010] An identification module is used to construct point-level, sequence-level and group-level abnormal feature libraries based on the urban dynamic feature data, identify abnormal events through multi-dimensional feature comparison and comprehensive evaluation, and obtain abnormal pattern data;

[0011] An allocation module is used to allocate computing resources under the edge computing architecture according to the abnormal mode data, and to make distributed optimization decisions through multi-node collaborative computing to obtain urban management decision data;

[0012] The optimization module is used to establish an evaluation index system for system performance, service quality and user satisfaction based on the urban management decision data, dynamically adjust and continuously optimize the decision-making strategy, and obtain an optimized urban system management plan.

[0013] A third aspect of the present application provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned data processing method for a smart city.

[0014] In the technical solution provided by this application, the city's multi-source data is collected through a distributed Internet of Things sensor network, and data cleaning and standardization are carried out in combination with data quality assessment rules, which effectively solves the real-time and reliability problems of data collection, and by establishing an urban domain concept system and a data entity mapping and matching mechanism, the semantic fusion of data from different sources is realized, and the data association analysis capability is enhanced. In the data analysis process, the time series is analyzed for periodicity and trend, and regional features are extracted in combination with spatial location information, and a complete model of regional interaction relationships is constructed, which improves the ability to grasp the dynamic characteristics of the city. By constructing an abnormal feature library at the point level, sequence level and group level, and performing multi-dimensional feature comparison and comprehensive evaluation, the recognition accuracy of abnormal events is enhanced. Computing resource allocation and multi-node collaborative computing are carried out under the edge computing architecture, which improves the efficiency of distributed optimization decision-making. At the same time, by establishing an evaluation index system for system performance, service quality and user satisfaction, the decision-making strategy is dynamically adjusted and continuously optimized to ensure the adaptability and effectiveness of the management solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0016] Figure 1 This is a schematic diagram of an embodiment of a data processing method for a smart city in an embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of space identification in an embodiment of the present application;

[0018] Figure 3This is a schematic diagram of the transaction classification and performance monitoring framework in the embodiment of the present application;

[0019] Figure 4 This is a schematic diagram of an embodiment of a data processing system for a smart city in an embodiment of the present application. DETAILED DESCRIPTION

[0020] Embodiments of the present application provide a data processing method, system and storage medium for a smart city. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of the data processing method for smart city in the embodiment of the present application includes:

[0022] Step S101: collect multi-source urban data through a distributed IoT sensor network, clean the original data according to data quality assessment rules, and perform format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data;

[0023] Step S102: Establish an urban domain concept system based on standardized urban data, map and match data entities from different sources, and perform knowledge representation and reasoning in combination with multimodal information to obtain semantically related data;

[0024] Step S103: using semantic association data, perform periodicity and trend analysis on the time series, extract regional features in combination with spatial location information, establish an interaction relationship model between regions, and obtain urban dynamic feature data;

[0025] Step S104: construct point-level, sequence-level and group-level abnormal feature libraries based on the city dynamic feature data, identify abnormal events through multi-dimensional feature comparison and comprehensive evaluation, and obtain abnormal pattern data;

[0026] Step S105: Based on the abnormal mode data, computing resources are allocated under the edge computing architecture, and distributed optimization decisions are made through multi-node collaborative computing to obtain urban management decision data;

[0027] Step S106: Based on the urban management decision data, an evaluation index system for system performance, service quality and user satisfaction is established, and the decision-making strategy is dynamically adjusted and continuously optimized to obtain an optimized urban system management plan.

[0028] It is understandable that the execution subject of the present application may be a data processing system for a smart city, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0029] Specifically, multi-source urban data are collected through a distributed IoT sensor network. These data include environmental monitoring data, traffic flow data, crowd density data, and energy consumption data. During the data collection process, each sensor node has the ability to self-organize and dynamically adjust the transmission power according to the signal strength and noise level of the surrounding environment. At the same time, a priority data frame scheduling mechanism is adopted to give priority to the transmission of key data. For the collected raw data, data cleaning is first performed to remove data points with abnormal values ​​and repeated data records, and then converted according to a unified data format, normalized the value type, and uniformly calibrated the timestamp. Finally, spatiotemporal alignment is performed to fill in the data of missing time points, and spatial coordinate mapping and correction are performed to obtain standardized urban data. After obtaining standardized urban data, an urban domain concept system is established. This concept system covers the concepts and relationships of multiple subdomains such as transportation, environment, energy, and safety. For data entities from different sources, a corresponding relationship is established between them and the concepts in the concept system through mapping and matching. In this process, the attribute characteristics of the data entity are first extracted, and then mapped and compared with the concept system to establish a corresponding relationship between the entity and the concept. On this basis, multimodal information such as text, numbers, and images are structured and processed to establish associations between different types of information. Finally, through semantic-level fusion, semantic connections between data entities are constructed to obtain semantically associated data.

[0030] Using semantically associated data, we can analyze the periodicity and trend of time series data. By extracting time attribute information, we can segment the data into time windows, identify the daily, weekly, and monthly characteristics of the data, and extract long-term change trends. At the same time, we can extract regional features by combining spatial location information, identify urban functional areas through spatial clustering analysis, and extract characteristic indicators of each area. On this basis, we can analyze the data interaction intensity between different areas, build the influence relationship between areas, and finally obtain the urban dynamic feature data. Based on the urban dynamic feature data, we can build a multi-level abnormal feature library. We can classify the data according to the point level, sequence level, and group level, analyze the numerical distribution of the point level data, and establish the numerical fluctuation range. We can also perform time association analysis on the sequence level data to extract the sequence change characteristics. We can also perform group behavior analysis on the group level data to identify the distribution law of the group data. Through multi-dimensional feature fusion, we can calculate the correlation between features, assign feature weights, and finally obtain abnormal pattern data through comprehensive evaluation and judgment.

[0031] For abnormal mode data, it is processed under the edge computing architecture. First, it is graded according to the computational complexity and data size, the processing priority is determined, and the computing power and network bandwidth occupancy of each node are counted. Then the computing tasks are divided according to the priority, the computing resources of each node are dynamically allocated, and the data transmission scheme between nodes is determined. After each node processes the data in parallel, the calculation results are summarized and merged, and the urban management decision data is generated through decision analysis. Finally, an evaluation index system is established based on the urban management decision data. The data is classified and analyzed according to system performance indicators, service quality indicators, and user satisfaction indicators. For system performance indicators, the response time and resource utilization are counted; for service quality indicators, the service accuracy and timeliness are calculated; for user satisfaction indicators, user feedback and complaint data are counted. Through comprehensive evaluation, the decision strategy parameters are adjusted, and finally the optimized urban system management plan is obtained.

[0032] For example: the distributed sensor network collects vehicle flow data, pedestrian flow data and signal light status data, and after data cleaning and standardization, establishes a traffic field concept system, maps entities such as vehicles, pedestrians, and signal lights with concepts, extracts the temporal variation law and spatial distribution characteristics of traffic flow, identifies congested sections and abnormal events, formulates traffic control strategies through collaborative calculations of edge nodes, and continuously optimizes management plans based on actual results. The specific data processing process includes: cleaning the original traffic data, eliminating abnormal data caused by sensor failures; converting data in different formats into a standard format; establishing semantic associations between vehicle flow, pedestrian flow, and road network structure; analyzing the peak characteristics of traffic flow; identifying abnormal congestion events; multiple edge nodes collaboratively calculate the optimal signal timing plan; and optimizing control strategies based on congestion relief effects.

[0033] In the embodiment of the present application, the city multi-source data is collected through a distributed Internet of Things sensor network, and data cleaning and standardization are performed in combination with data quality assessment rules, which effectively solves the real-time and reliability problems of data collection and data quality, and by establishing an urban domain concept system and a data entity mapping and matching mechanism, the semantic fusion of data from different sources is realized, and the data association analysis capability is enhanced. In the data analysis process, the time series is analyzed for periodicity and trend, and regional features are extracted in combination with spatial location information, and a complete model of regional interaction relationships is constructed, which improves the ability to grasp the dynamic characteristics of the city. By constructing an abnormal feature library at the point level, sequence level and group level, and performing multi-dimensional feature comparison and comprehensive evaluation, the recognition accuracy of abnormal events is enhanced. Computing resource allocation and multi-node collaborative computing are performed under the edge computing architecture, which improves the efficiency of distributed optimization decision-making. At the same time, by establishing an evaluation index system for system performance, service quality and user satisfaction, the decision-making strategy is dynamically adjusted and continuously optimized to ensure the adaptability and effectiveness of the management solution.

[0034] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0035] (1) Collect environmental monitoring data, traffic flow data, crowd density data, and energy consumption data through a distributed IoT sensor network to obtain multi-source urban data;

[0036] (2) Perform a numerical range check on the multi-source data of the city, remove data points with abnormal fluctuations, and remove duplicate data records to obtain preliminary cleaned data;

[0037] (3) Convert the preliminary cleaned data into a unified data format, normalize the numerical type, and calibrate the timestamp to obtain data in a unified format;

[0038] (4) Time series completion is performed on the unified format data to fill in the data of missing time points, and spatial coordinate mapping and correction are performed to obtain time-space aligned data;

[0039] (5) Verify data integrity based on spatiotemporal alignment data, remove data items that do not meet quality standards, and score the data for reliability to obtain quality qualified data;

[0040] (6) Organize and classify the qualified data according to the urban data standards, establish the correlation index between the data, generate the data quality report, and obtain standardized urban data.

[0041] Specifically, the distributed IoT sensor network collects environmental monitoring data, traffic flow data, crowd density data and energy consumption data through sensor nodes deployed at key locations in the city. Environmental monitoring data covers environmental parameters such as air quality, noise level, temperature and humidity; traffic flow data includes traffic parameters such as the number of vehicles, driving speed, and traffic status; crowd density data records indicators such as the degree of crowd gathering, moving direction, and crowd flow changes; energy consumption data counts information such as electricity usage, gas consumption, and water resource utilization. Each sensor node uses a layered protocol architecture for data transmission. The physical layer dynamically adjusts the signal strength through adaptive power control technology. The data link layer uses a priority scheduling mechanism to ensure the real-time nature of the data. The network layer optimizes the data transmission path based on the routing algorithm. When performing a numerical range test on the collected multi-source urban data, the valid value range of each type of data is first set. The inspection standards for environmental monitoring data are based on the physical limits of environmental parameters, such as temperature data exceeding a reasonable range or abnormal deviations in humidity values; the inspection rules for traffic flow data are based on road capacity and historical data distribution, identifying data points that do not conform to actual conditions; the inspection method for crowd density data is based on site capacity and crowd distribution patterns, eliminating values ​​that clearly exceed the site's carrying capacity; the inspection of energy consumption data is based on energy-consuming equipment specifications and historical energy consumption patterns, removing unreasonable energy consumption records. At the same time, duplicate data records are identified through timestamp comparison, and redundant data items are deleted.

[0042] During the format conversion process of the preliminary cleansing data, a unified data format specification is formulated for data from different sources. The normalization processing of numerical types adopts the maximum and minimum value normalization method to map data of different dimensions to the same numerical interval; the unified calibration of timestamps is based on the standard time base, unifying the data collection time mark to solve the data misalignment problem caused by the asynchrony of time of different devices. In the time series completion processing of format-unified data, the data of missing time points are supplemented by linear interpolation, adjacent value filling and other methods to ensure the time continuity of the data. Spatial coordinate mapping and correction are based on a unified spatial reference system to establish the geographical location relationship of data points. The integrity verification of spatiotemporal alignment data includes data item integrity check and data value validity check. The data item integrity check verifies whether the key fields are complete, and the data value validity check determines whether the data meets the business logic. The data reliability scoring mechanism comprehensively considers factors such as data accuracy, completeness, and consistency, and obtains the data quality score through weighted calculation to screen out data items that meet the quality standards.

[0043] Qualified data is classified and sorted according to urban data standards, and an association index is established between data. Data classification is based on the business attributes and usage scenarios of the data, such as classifying traffic data according to road grade and vehicle type, and classifying environmental data according to monitoring indicators and regional scope. The data association index records the association between different categories of data, such as the association between traffic flow and environmental quality, and the association between passenger density and energy consumption. The data quality report records statistical information during the data processing process, including indicators such as the amount of original data, the amount of data after cleaning, the data completeness rate, and the proportion of abnormal data.

[0044] For example, at a traffic intersection, multiple sensors collect traffic data at the same time. Sensor A records vehicle data from north to south, and sensor B records vehicle data from east to west. In the original data, sensor A records an abnormally large traffic flow at a certain point in time, which far exceeds the road capacity. The abnormal value is identified and eliminated through the value range test. There is a 5-second time difference between the timestamps of the two sensors. The data is aligned to the same time base through unified timestamp calibration. The data of sensor B is missing in a certain period of time, and it is supplemented by interpolation through the data of the adjacent period. After data quality assessment, the data reliability score of sensor A is higher, and the data of sensor B has certain fluctuations. Finally, the processed data is classified according to vehicle type and driving direction, and an associated index with data such as signal light timing and weather conditions is established to form standardized traffic data.

[0045] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0046] (1) Classify the standardized urban data into the fields of transportation, environment, energy, and safety, build a basic concept library for each field, and form a preliminary concept system;

[0047] (2) hierarchically processing the concepts in the preliminary concept system, determining the subordinate and associated relationships between the concepts, and obtaining the urban domain concept system;

[0048] (3) classifying and labeling data entities in the standardized urban data according to the urban domain concept system, extracting attribute characteristics of the data entities, and obtaining entity feature data;

[0049] (4) mapping and comparing the entity feature data with the urban domain concept system, establishing a corresponding relationship between data entities and concepts, and obtaining entity mapping data;

[0050] (5) Structuring the text, numerical value, and image information in the entity mapping data, establishing association relationships between different types of information, and obtaining multimodal association data;

[0051] (6) The multimodal related data are semantically fused to construct semantic connections between data entities, extract implicit relationships between data, and obtain semantically related data.

[0052] Specifically, standardized urban data are first classified and sorted according to the four main areas of transportation, environment, energy, and safety. The transportation field includes basic concepts such as road networks, vehicle flows, and traffic signals; the environmental field covers environmental elements such as air quality, noise levels, and green coverage; the energy field includes energy concepts such as power supply, gas use, and water resource consumption; and the safety field includes safety elements such as public security prevention and control, emergency response, and facility monitoring. Based on these basic concepts, a basic concept library for each field is established to form a preliminary concept system to support urban management. When hierarchically processing the preliminary concept system, the subordinate relationship between concepts is constructed through the concept classification tree. The subordinate relationship describes the inclusion and inclusion relationship between concepts. For example, in the transportation field, the vehicle concept includes sub-concepts such as private cars, buses, and taxis. The association relationship between concepts describes the mutual influence and dependency between different concepts, such as the association between traffic flow and air quality, and the association between pedestrian density and energy consumption. By establishing these relationships, a concept system for the urban field is formed.

[0053] Based on the urban domain concept system, the entities in the standardized urban data are classified and labeled. Data entities are specific objects in the city, such as a road, a building, a facility, etc. In the classification and labeling process, the category attributes of the entity are first identified and classified into the corresponding concept category, and then the specific attribute characteristics of the entity are extracted. The attribute characteristics include the basic attributes of the entity (such as location, size, number) and business attributes (such as usage status, operating parameters, performance indicators). In the mapping and comparison process between the entity feature data and the urban domain concept system, the corresponding relationship between the data entity and the concept is established. The mapping process includes two aspects: attribute mapping and relationship mapping. Attribute mapping matches the specific attributes of the entity with the attributes defined by the concept, and relationship mapping establishes the correspondence between the relationship between entities and the relationship between concepts. Through this mapping, discrete data entities are incorporated into a unified conceptual framework.

[0054] The structured processing of multimodal information such as text, numerical values, and images involves the standardized expression of different types of data. Text information forms a structured representation through keyword extraction and semantic analysis; numerical information is standardized through data type conversion and unit unification; image information forms a numerical description through feature extraction. On this basis, the association relationship between different types of information is established, such as corresponding text descriptions to numerical indicators and associating image features with attribute parameters. The semantic-level fusion of multimodal related data is the process of establishing deep semantic connections between data entities. First, the direct association between entities is analyzed, such as spatial proximity and functional dependency; then the indirect association between entities is discovered through association transfer; finally, various association relationships are combined to extract the implicit relationship between data and form a semantic association network.

[0055] For example, in the data processing process, the road traffic data and environmental monitoring data are first classified into the traffic field and the environmental field respectively. At the conceptual level, basic concepts such as roads, traffic flow, and air quality are established, and the relationship between them is clarified: roads carry traffic flow, and traffic flow affects air quality. When annotating specific data entities, a certain main road is used as a road entity, with attributes such as the number of lanes and traffic capacity; the traffic data of this section of road is used as a traffic entity, recording characteristics such as traffic flow and speed; the surrounding air monitoring points are used as environmental entities to monitor pollutant concentrations. Through mapping and comparison, the corresponding relationship between road entities and road concepts, traffic flow data and traffic concepts, and monitoring data and environmental concepts is established. In multimodal data processing, the road condition information in text form, the traffic flow data in numerical form, and the monitoring data in image form are structured and associated. Finally, through semantic fusion, the association rules between changes in traffic flow, changes in pollutant concentration, and changes in weather conditions are discovered to form a semantic association network.

[0056] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0057] (1) Extracting temporal attribute information from semantically related data, segmenting the data by time window, and statistically analyzing the data change patterns to obtain basic time series data;

[0058] (2) Perform periodic decomposition on the basic time series data, identify the daily, weekly, and monthly change characteristics of the data, extract the long-term change trend of the data, and obtain periodic analysis data;

[0059] (3) Associating the periodic analysis data with the spatial coordinate information, performing cluster analysis on the spatial distribution characteristics of the data, and obtaining spatial cluster data;

[0060] (4) Identify urban functional areas through spatial clustering data, extract characteristic indicators of each area, establish regional characteristic descriptions, and obtain regional characteristic data;

[0061] (5) Calculate the correlation of regional feature data, analyze the data interaction intensity between different regions, construct the influence relationship between regions, and obtain regional interaction data;

[0062] (6) Integrate regional interaction data with time series variation patterns, analyze the dynamic variation patterns of regional interactions, and obtain urban dynamic characteristic data.

[0063] Specifically, in the process of smart city data processing, extracting time attribute information from semantically associated data first locates the time dimension information of each data point, including time tags such as data collection time and data update time. When segmenting data according to time windows, set fixed time intervals (such as hours, days, weeks, etc.) to divide the continuous time series into multiple time segments. Perform statistical analysis on the data in each time segment, including calculating statistical features such as mean, variance, peak value, etc., recording the change trend and distribution characteristics of the data, and forming time series basic data. In the process of periodic decomposition of time series basic data, extract the change laws of different time scales respectively. First, identify the daily change characteristics of the data, such as data fluctuations during peak hours in the morning and evening; then analyze the weekly change characteristics, such as the data differences between weekdays and weekends; finally summarize the monthly change characteristics, such as the monthly periodic change law. At the same time, by removing short-term fluctuations, extract the long-term change trend of the data, and obtain periodic analysis data.

[0064] When associating periodic analysis data with spatial coordinate information, geographic location information is added to each data point, including spatial identifiers such as longitude and latitude, and area code. Figure 2 As shown, it is a schematic diagram of spatial identification in an embodiment of the present application, wherein the figure shows the process of identifying and interacting with urban functional areas based on spatiotemporal data analysis. The area on the left in the figure is marked as a residential area, and the area on the right is marked as a commercial area, and the regional boundaries are represented by solid rectangles. Multiple data collection points are distributed inside each functional area, marked with solid dots, and these collection points are responsible for collecting spatiotemporal data such as the flow of people and activity intensity in the area. The flow of people between areas is represented by a curve with an arrow, which reflects the two-way interactive relationship between residential and commercial areas. The wave curve at the bottom of the figure shows the flow pattern of the time series, reflecting the dynamic changes of factors such as the flow of people and logistics between regions over time. The data points are grouped by a spatial clustering algorithm, and points with similar spatial positions and similar data characteristics are divided into the same cluster to obtain spatial clustering data. The mathematical expression of the spatial clustering process is as follows:

[0065]

[0066] Among them, F region represents the regional characteristic value, P i represents the regional flow density index, Di Represents the regional building density index, T i represents the regional traffic flow index, α i , β i , γ i They represent the weight coefficients of each indicator, and n represents the number of characteristic indicators.

[0067] When identifying urban functional areas based on spatial clustering data, the data feature distribution of each cluster is analyzed, and key indicators that can characterize regional functions, such as passenger flow characteristics, building density, traffic flow, etc., are extracted to establish regional feature descriptions and form regional feature data. The correlation between regions is calculated using the following formula:

[0068]

[0069] Among them, R interaction represents the inter-region interaction strength, V i and V j Represent the feature vectors of region i and region j respectively, ω ij represents the weight coefficient between regions, Q ij represents the distance attenuation factor between regions, represents a vector operator, and m represents the total number of regions.

[0070] When integrating regional interaction data with the laws of time series changes, we analyze the changes in the interaction relationship between regions in different time periods, including the migration patterns of people, the intensity of traffic connections, the characteristics of resource flows, etc., and finally form a complete data set that describes the dynamic characteristics of the city.

[0071] Take the interactive relationship analysis between commercial and residential areas as an example: first, extract the time attributes of data such as pedestrian flow and consumption behavior from the semantic association data, segment the time window according to the hourly unit, and count the changes in pedestrian flow in each period. Then, perform periodic decomposition on these time series data, and find that residents flow from residential areas to commercial areas during the morning and evening peak hours on weekdays, while weekends show different time distribution characteristics. Combine these periodic data with spatial location information, and identify the main commercial and residential areas through cluster analysis. Extract characteristic indicators for each area, such as store density and business hours distribution in commercial areas, and population density and age structure in residential areas. Through correlation analysis, find the interactive relationship between pedestrian flow, logistics, information flow, etc. between commercial and residential areas, and finally combine these regional interactive characteristics with the time change law to form urban dynamic characteristic data.

[0072] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0073] (1) The urban dynamic characteristic data is divided into single-point data, time series data, and group data according to the data characteristics, and the data characteristics are classified and counted to obtain characteristic classification data;

[0074] (2) Perform numerical distribution analysis on single-point data in feature classification data, establish numerical fluctuation range limits, extract outlier features, and obtain point-level feature data;

[0075] (3) Perform time correlation analysis on the time series data in the point-level feature data, extract the change characteristics of the data sequence, establish a sequence feature description, and obtain sequence-level feature data;

[0076] (4) Perform group behavior analysis on the group data in the sequence-level feature data, identify the distribution pattern of the group data, extract group characteristics, and obtain group-level feature data;

[0077] (5) Perform multi-dimensional feature fusion on the group-level feature data, calculate the correlation between the features, assign feature weights, and obtain feature association data;

[0078] (6) Identify abnormal events in feature-related data, establish abnormal scoring standards, conduct comprehensive evaluation and judgment, and obtain abnormal pattern data.

[0079] Specifically, the urban dynamic feature data is classified and divided into three types: single point data, time series data and group data. Single point data refers to an independent data point at a certain time and location, such as the instantaneous traffic flow at a certain intersection; time series data refers to a data sequence with time continuity, such as continuous monitoring data of traffic flow; group data refers to a data set that describes the overall characteristics of multiple related objects, such as the distribution of traffic flow at all intersections in a region. Feature classification data is formed by statistically analyzing the quantity distribution, value range and change characteristics of different types of data. When processing single point data in feature classification data, the statistical characteristics of the values ​​are first calculated, including the mean, standard deviation, maximum and minimum values, etc., to determine the distribution range of the data. A reasonable boundary of value fluctuation is established based on the statistical characteristics, and data points beyond the boundary are marked as outliers. The feature information of each outlier is extracted, including the time of occurrence, location, degree of deviation, etc., and these feature information are integrated to form point-level feature data.

[0080] For the time series data in the point-level feature data, perform time correlation analysis. Time align the data series to ensure that the time intervals of the data points are consistent. Then analyze the time correlation of the data, including autocorrelation (the relationship between different time points in the same data series) and cross-correlation (the time relationship between different data series). Based on the correlation analysis results, extract the change characteristics of the data series, such as trend, periodicity, mutation, etc., establish a sequence feature description, and obtain sequence-level feature data.

[0081] When analyzing group behavior of group data in sequence-level feature data, the following mathematical model is used:

[0082]

[0083] Among them, G behavior Represents the characteristic value of group behavior, M k Represents the group mobility pattern index, N k Represents the group aggregation index, H k represents the group activity intensity index, λ k , μ k , σ k Respectively represent the weight coefficient of each indicator, θ k represents the time attenuation factor, t represents the time parameter, and l represents the number of feature dimensions.

[0084] When multi-dimensional feature fusion is performed on group-level feature data, a feature vector is first established, which includes multiple dimensions such as spatial distribution features, time evolution features, and behavior pattern features. The correlation coefficient matrix between different features is calculated to identify the dependency relationship between features. Feature weights are assigned according to the importance and discrimination of the features to form feature-related data. To identify abnormal events in feature-related data, the abnormality scoring criteria are first defined, including single-point abnormality scores (degree of numerical deviation), sequence abnormality scores (degree of abnormal trend of change), and group abnormality scores (degree of abnormal distribution pattern). These scores are weighted and combined to obtain a comprehensive abnormality score. Based on the abnormality score and the preset threshold, judgment is made, and finally abnormal events are identified to form abnormal pattern data.

[0085] Take traffic congestion analysis as an example: First, traffic data is divided into single-point data (instantaneous traffic flow at the intersection), time series data (sequence of traffic flow changes over time) and group data (overall traffic flow distribution of the regional road network). Analysis of single-point data found that the traffic flow at a certain intersection was significantly beyond the normal range and was marked as an outlier. Through time series data analysis, it was found that the traffic flow at this intersection continued to rise and showed a significant correlation with the traffic flow changes at surrounding intersections. Group data analysis reveals the traffic flow pattern and congestion diffusion law of the entire region. Through multi-dimensional feature fusion, the characteristics of traffic flow, speed, occupancy rate, etc. are comprehensively analyzed, and finally the formation, development and dissipation process of congestion events are identified, generating abnormal pattern data. These data reflect the spatial scope, temporal evolution and impact of congestion events.

[0086] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0087] (1) The abnormal pattern data is classified according to the computational complexity and data size, the processing priority of data at different levels is determined, and the graded processing data is obtained;

[0088] (2) Perform load calculation on the hierarchical processing data, count the computing power and network bandwidth usage of each node, and obtain resource status data;

[0089] (3) Divide the computing tasks in the resource status data according to priority, dynamically allocate the computing resources of each node, and obtain task allocation data;

[0090] (4) Perform data exchange and task coordination between nodes for task allocation data, determine the data transmission scheme between nodes, and obtain collaborative computing data;

[0091] (5) Process the collaborative computing data in parallel on each node, summarize the computing results of each node, merge the data, and obtain distributed processing data;

[0092] (6) Conduct decision analysis on distributed processing data, integrate the processing results of each node, generate management decision plans, and obtain urban management decision data.

[0093] Specifically, in the distributed computing process of smart cities, abnormal pattern data is classified and graded. The data is graded according to its computational complexity (such as computational time complexity and spatial complexity) and data scale (such as data volume and data dimension). The computational complexity is divided into real-time computing level, near real-time computing level and offline computing level; the data scale is divided into small scale (KB level), medium scale (MB level) and large scale (GB level and above). By combining computational complexity and data scale, the priority of data processing is determined to form hierarchical processing data. When performing load calculation on hierarchical processing data, the following calculation model is used:

[0094]

[0095] Among them, L node represents the node load value, C j Indicates the CPU usage indicator, B j Indicates the bandwidth usage index, U j Represents the memory usage index, x j ,y j 、z j Respectively represent the weight coefficient of each resource indicator, w j represents the distance attenuation factor, d represents the distance between nodes, and p represents the resource indicator dimension. The model is used to count the resource status of each computing node, including indicators such as CPU utilization, memory usage, and network bandwidth usage, to obtain resource status data.

[0096] Prioritize the computing tasks in the resource status data based on the urgency of the task, the amount of resource required, and the expected execution time. Dynamically evaluate the available resources of each computing node, and assign tasks to the most suitable nodes based on the computing power and current load status of the node. The task allocation process takes load balancing factors into consideration to avoid overloading a single node and form task allocation data. When coordinating between nodes for task allocation data, design a data transmission scheme. First, determine the data dependency between nodes and build a directed acyclic graph for task execution. Then plan the data transmission path and select links with sufficient network bandwidth for data transmission. At the same time, establish a task synchronization mechanism between nodes to ensure the execution order of dependent tasks and form collaborative computing data.

[0097] The collaborative computing data is processed in parallel on the edge nodes, and each node performs a specific computing process according to the assigned tasks. The nodes synchronize their states and exchange intermediate results through a message passing mechanism. After the calculation is completed, the calculation results of each node are collected, and the data is merged and sorted to form distributed processing data. Decision analysis is performed on the distributed processing data, and the processing results of each node are checked for consistency to handle data conflicts and inconsistencies. Then, the processing results are evaluated and screened according to business rules to extract key decision information. Finally, a management decision plan that meets business needs is generated to form urban management decision data.

[0098] For example, traffic congestion occurs in a certain area, generating a large amount of abnormal pattern data. These data are graded according to the urgency of processing. Real-time traffic data is the highest priority and needs to be processed immediately; traffic flow prediction is the second highest priority and requires near real-time processing; historical data analysis is a normal priority and allows delayed processing. The system evaluates the resource status of edge nodes and finds that the CPU load of some nodes is high. It assigns high-priority tasks to nodes with lower loads. A data transmission channel is established between nodes, and the traffic flow data of upstream intersections is transmitted to downstream nodes to support collaborative computing at the road network level. Each node processes data in parallel, with some nodes responsible for real-time traffic analysis and other nodes for traffic flow prediction. The calculation results of each node are summarized to generate urban management decision data including traffic control plans.

[0099] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0100] (1) Classify urban management decision data according to system performance indicators, service quality indicators, and user satisfaction indicators, conduct statistical analysis on the data, and obtain classification evaluation data;

[0101] (2) Calculate the system performance indicators in the classified evaluation data, count the system response time and resource utilization, generate performance statistics, and obtain system performance evaluation data;

[0102] (3) Quantify the service quality indicators in the system performance evaluation data, calculate the service accuracy and service timeliness, integrate the service evaluation data, and obtain the service quality evaluation data;

[0103] (4) Analyze the user satisfaction index in the service quality evaluation data, collect user feedback data and complaint data, organize the data, and obtain user satisfaction evaluation data;

[0104] (5) Comprehensively calculate the user satisfaction evaluation data according to the evaluation index weights, evaluate and score the decision-making strategy, and obtain the strategy evaluation data;

[0105] (6) Optimize and analyze the strategy evaluation data, adjust the decision parameters, generate management strategy plans, and obtain the optimized urban system management plan.

[0106] Specifically, in the management evaluation process of smart cities, urban management decision data is first classified according to three dimensions: system performance indicators describe the operating status of infrastructure and computing resources, service quality indicators reflect the effectiveness of management services, and user satisfaction indicators reflect the experience of service recipients. Statistical analysis is performed on these three types of data to calculate the distribution characteristics, change trends, and interrelationships of various indicators to form classified evaluation data.

[0107] The following model is used to calculate the system performance indicators in the classified evaluation data:

[0108]

[0109] Among them, P system represents the system performance score, R i Represents the response time indicator, E i Represents the resource utilization index, S i Indicates the server load indicator, a i 、b i 、c i Respectively represent the weight coefficient of each indicator, k i represents the time impact factor, t represents the running time, and n represents the performance indicator dimension. The system performance evaluation data is obtained through this model.

[0110] Quantify the service quality indicators in the system performance evaluation data and count the processing results of service requests, including the correctness, completeness and timeliness of service responses. Calculating service accuracy requires counting the success rate and error rate of service requests, and service timeliness is measured by calculating the time delay of service responses. Integrate these evaluation indicators to form service quality evaluation data. Analyze the user satisfaction indicators in the service quality evaluation data, and process user feedback data and complaint data. User feedback data includes information such as service evaluation, usage experience, and improvement suggestions. Complaint data records problems encountered during the service process and aspects that users are dissatisfied with. Through data sorting, classification and summary, user satisfaction evaluation data is obtained.

[0111] The comprehensive calculation of user satisfaction evaluation data adopts the following model:

[0112]

[0113] Among them, S strategy represents the strategy evaluation score, F m represents the user feedback index, C m Indicates the complaint handling index, I m represents the service improvement indicator, u m 、v m 、w m Represent the weight coefficient of each indicator, represents the time decay function, and q represents the evaluation indicator dimension. The strategy evaluation data is obtained through this model.

[0114] Optimize and analyze the strategy evaluation data, identify indicators with low scores, and analyze the causes of the problems. Then adjust the decision parameters according to the evaluation results, including resource allocation strategy, service response strategy, user interaction strategy, etc. Finally, generate a new management strategy plan to form an optimized urban system management plan.

[0115] For example: collect management decision data related to bus services, including vehicle dispatching strategies, route planning schemes, and station layout optimization. From the perspective of system performance, calculate indicators such as the response time and server load of the bus dispatching system; from the perspective of service quality, calculate indicators such as departure punctuality and vehicle utilization; analyze passenger evaluation and complaint information from the perspective of user satisfaction. In the data processing process, these data are first classified according to three dimensions, and then various indicators are calculated separately. System performance calculations show that dispatching response delays in certain periods, service quality statistics find that the departure intervals of some lines are unreasonable, and user feedback data reflect that the problem of vehicle congestion during peak hours is prominent. Through comprehensive evaluation, the scores of various indicators are calculated to find the links that need to be optimized. Finally, an optimization plan is generated, including adjusting the departure frequency, optimizing route settings, and improving vehicle dispatching strategies.

[0116] In a specific embodiment, the process of calculating the system performance index in the classified evaluation data may specifically include the following steps:

[0117] (1) Grouping the system response time in the classified evaluation data by transaction type, counting different types of response delays, summarizing the average response time, and obtaining response time data;

[0118] (2) sampling the processor usage in the response time data, counting the peak and average CPU usage, recording usage fluctuations, and obtaining processor statistics;

[0119] (3) Monitor the memory usage in the processor statistics, count the memory occupancy rate and release rate, record the memory allocation status, and obtain memory statistics;

[0120] (4) Analyze the storage space usage in the memory statistics, count the data writing speed and reading speed, record the storage capacity changes, and obtain storage statistics;

[0121] (5) Record the network transmission conditions in the stored statistical data, calculate the statistical data throughput and network bandwidth occupancy, calculate the network load conditions, and obtain network statistical data;

[0122] (6) Conduct a comprehensive performance analysis of network statistical data, calculate the scores of various performance indicators, generate a performance evaluation report, and obtain system performance evaluation data.

[0123] Specifically, the classified evaluation data is deeply analyzed and processed. The response time data in the classified evaluation data is processed to classify the transaction types into real-time transactions (such as traffic signal control), quasi-real-time transactions (such as bus scheduling) and non-real-time transactions (such as data statistical analysis). Figure 3As shown, it is a schematic diagram of the transaction classification and performance monitoring framework in the embodiment of the present application. The figure is divided into three main parts: the upper part shows the transaction classification framework, the middle part is the system performance monitoring index, and the lower part is the performance evaluation report. In the transaction classification part, transactions are divided into three categories: real-time transactions, quasi-real-time transactions, and non-real-time transactions. Each type of transaction is marked with its typical application scenario, response time requirement, and priority level. The system performance monitoring part includes four dimensions: processor monitoring, memory monitoring, storage monitoring, and network monitoring. Each dimension lists specific monitoring indicators. The performance evaluation report part is used to summarize and analyze the above monitoring data. The figure uses a one-way arrow to connect the transaction classification and performance monitoring parts to indicate the connection relationship between the data flow and the analysis process. The overall hierarchical layout of rectangular nesting is adopted to highlight the logical relationship between the various components of the system. For each transaction type, the request initiation time and the response completion time are recorded, and the time difference is calculated to obtain the response delay. The response delay of each type of transaction is statistically analyzed, including the maximum delay, the minimum delay, and the average delay, to form the response time data. In the process of processing the response time data, the processor usage is synchronously collected. The sampling frequency is set to sample once per second to record the real-time utilization rate of the central processing unit. The peak and average levels of processor usage are calculated through continuous sampling data, and the changing trends of processor usage, including sudden high loads and continuous loads, are recorded. These data are sorted and summarized to form processor statistics.

[0124] Based on the processor statistics, memory usage is monitored. Memory monitoring includes physical memory and virtual memory, recording the total amount of memory, used memory, and available memory. When calculating the memory occupancy rate, the used memory is compared with the total memory; when calculating the memory release rate, the amount of memory released per unit time is counted. At the same time, the memory allocation process, including the status information of memory application, allocation, and release, is recorded and compiled into memory statistics. Based on the memory statistics, the usage of storage space is further analyzed. Storage analysis includes two levels: temporary storage and persistent storage, focusing on the read and write performance of data. By recording the execution time of data write operations and read operations, the data write speed and read speed are calculated. At the same time, the capacity changes of storage space are monitored, the growth trend of data storage volume is recorded, and storage statistics are compiled.

[0125] For storage statistics, supplement the monitoring records of network transmission performance. Network performance monitoring includes two dimensions: data transmission efficiency and network resource occupancy. Data transmission efficiency is measured by calculating data throughput, that is, the amount of data successfully transmitted per unit time; network resource occupancy is evaluated by statistical bandwidth usage. Based on these monitoring data, the network load level is calculated and compiled to form network statistics. Finally, a comprehensive analysis of network statistics is conducted to uniformly evaluate the overall performance of the system. A performance evaluation indicator system is set, including response time indicators, resource utilization indicators, and service capability indicators. Quantitative calculations and scoring of various indicators are performed, and finally a performance evaluation report is generated to form system performance evaluation data.

[0126] For example, in the process of traffic signal control, the response time of different types of transactions is first recorded. Real-time transactions such as the response time of traffic light control instructions, quasi-real-time transactions such as the processing time of road condition information updates, and non-real-time transactions such as the calculation time of traffic flow statistics. At the same time, the processor usage is monitored, and the processor load during peak signal optimization calculations is recorded. Memory monitoring shows the changes in memory usage during the execution of the signal optimization algorithm, and storage analysis records the performance of continuous writing and query operations of traffic data. Network monitoring focuses on the transmission efficiency of signal control instructions and the synchronous bandwidth usage of road condition data. Through the comprehensive analysis of these data, the performance status of the entire traffic control system is evaluated to provide data support for system optimization.

[0127] The above describes the data processing method for smart city in the embodiment of the present application. The following describes the data processing system for smart city in the embodiment of the present application. Figure 4 , an embodiment of the data processing system for smart city in the embodiment of the present application includes:

[0128] The acquisition module 201 is used to collect multi-source urban data through a distributed IoT sensor network, clean the original data according to data quality assessment rules, and perform format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data;

[0129] The matching module 202 is used to establish a city domain concept system based on standardized city data, map and match data entities from different sources, and perform knowledge representation and reasoning in combination with multimodal information to obtain semantically related data;

[0130] The analysis module 203 is used to use the semantic association data to perform periodicity and trend analysis on the time series, extract regional features in combination with spatial location information, establish an interaction relationship model between regions, and obtain urban dynamic feature data;

[0131] The identification module 204 is used to construct point-level, sequence-level and group-level abnormal feature libraries based on the urban dynamic feature data, identify abnormal events through multi-dimensional feature comparison and comprehensive evaluation, and obtain abnormal pattern data;

[0132] Allocation module 205, used to allocate computing resources under the edge computing architecture according to the abnormal pattern data, and make distributed optimization decisions through multi-node collaborative computing to obtain urban management decision data;

[0133] The optimization module 206 is used to establish an evaluation index system of system performance, service quality and user satisfaction based on the urban management decision data, dynamically adjust and continuously optimize the decision strategy, and obtain an optimized urban system management plan.

[0134] Through the collaboration of the above components, the distributed IoT sensor network collects multi-source urban data, and combines data quality assessment rules to clean and standardize data, effectively solving the problems of real-time data collection and reliability of data quality. By establishing the urban domain concept system and data entity mapping and matching mechanism, the semantic fusion of data from different sources is realized, and the data association analysis capability is enhanced. In the data analysis process, the periodicity and trend analysis of the time series is carried out, and the regional features are extracted in combination with the spatial location information. A complete model of regional interaction relationship is constructed, which improves the ability to grasp the dynamic characteristics of the city. By constructing anomaly feature libraries at the point level, sequence level and group level, and performing multi-dimensional feature comparison and comprehensive evaluation, the accuracy of abnormal event identification is enhanced. Computing resource allocation and multi-node collaborative computing under the edge computing architecture improve the efficiency of distributed optimization decision-making. At the same time, by establishing an evaluation index system for system performance, service quality and user satisfaction, the decision-making strategy is dynamically adjusted and continuously optimized to ensure the adaptability and effectiveness of the management plan.

[0135] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the data processing method for a smart city.

[0136] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0138] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A data processing method for a smart city, characterized in that: The data processing method for smart city includes: Collect multi-source urban data through a distributed IoT sensor network, clean the raw data according to data quality assessment rules, convert the format and perform spatiotemporal alignment on the cleaned data according to preset data standards to obtain standardized urban data; Establishing an urban domain concept system based on the standardized urban data, mapping and matching data entities from different sources, combining multimodal information for knowledge representation and reasoning, and obtaining semantically related data; Utilizing the semantic association data, periodicity and trend analysis is performed on the time series, regional features are extracted in combination with spatial location information, and an interaction relationship model between regions is established to obtain urban dynamic feature data; Based on the urban dynamic feature data, point-level, sequence-level and group-level abnormal feature libraries are constructed, and abnormal events are identified through multi-dimensional feature comparison and comprehensive evaluation to obtain abnormal pattern data; Based on the abnormal pattern data, computing resources are allocated under the edge computing architecture, and distributed optimization decisions are made through multi-node collaborative computing to obtain urban management decision data; Based on the urban management decision data, an evaluation index system for system performance, service quality and user satisfaction is established, and the decision-making strategy is dynamically adjusted and continuously optimized to obtain an optimized urban system management plan.

2. The data processing method for smart city according to claim 1, characterized in that: The method collects multi-source urban data through a distributed IoT sensor network, cleans the original data according to data quality assessment rules, and performs format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data, including: Collect environmental monitoring data, traffic flow data, crowd density data and energy consumption data through distributed IoT sensor networks to obtain multi-source urban data; Performing a numerical range check on the multi-source data of the city, eliminating data points with abnormal fluctuations, and removing duplicated data records to obtain preliminary cleaned data; The preliminary cleansed data is converted into a unified data format, the numerical type is normalized, and the timestamp is uniformly calibrated to obtain data in a unified format; Performing time series completion on the data in the unified format, filling in the data of missing time points, and performing spatial coordinate mapping and correction to obtain time-space aligned data; Performing data integrity verification based on the spatiotemporal alignment data, eliminating data items that do not meet quality standards, and scoring the data for reliability to obtain quality qualified data; The qualified data are sorted and classified according to the urban data standards, a correlation index between the data is established, a data quality report is generated, and standardized urban data is obtained.

3. The data processing method for smart city according to claim 1, characterized in that: The method of establishing a city domain concept system based on the standardized city data, mapping and matching data entities from different sources, combining multimodal information for knowledge representation and reasoning, and obtaining semantically related data includes: Classify the standardized urban data into fields such as transportation, environment, energy, and safety, build a basic concept library in each field, and form a preliminary concept system; The concepts in the preliminary concept system are processed hierarchically, the subordinate and associated relationships between the concepts are determined, and the concept system of the urban field is obtained; Classify and annotate data entities in the standardized urban data according to the urban domain concept system, extract attribute features of the data entities, and obtain entity feature data; Mapping and comparing the entity feature data with the urban domain concept system, establishing a corresponding relationship between data entities and concepts, and obtaining entity mapping data; Structuring the text, numerical value, and image information in the entity mapping data, establishing association relationships between different types of information, and obtaining multimodal association data; The multimodal associated data are semantically fused to construct semantic connections between data entities, extract implicit relationships between data, and obtain semantically associated data.

4. The data processing method for smart city according to claim 1, characterized in that: The semantic association data is used to analyze the periodicity and trend of the time series, and the regional features are extracted in combination with the spatial location information to establish an interaction relationship model between regions to obtain the city dynamic feature data, including: Extracting time attribute information from the semantically associated data, segmenting the data by time window, and performing statistical analysis on the data change rules to obtain time series basic data; Performing periodic decomposition on the time series basic data, identifying the daily, weekly and monthly change characteristics of the data, extracting the long-term change trend of the data, and obtaining periodic analysis data; Associating the periodic analysis data with the spatial coordinate information, performing cluster analysis on the spatial distribution characteristics of the data, and obtaining spatial cluster data; Identify urban functional areas through the spatial clustering data, extract characteristic indicators of each area, establish regional characteristic descriptions, and obtain regional characteristic data; Perform correlation calculation on the regional feature data, analyze the data interaction intensity between different regions, construct the influence relationship between regions, and obtain regional interaction data; The regional interaction data is integrated with the time series variation law, the dynamic variation law of the interaction between regions is analyzed, and the urban dynamic characteristic data is obtained.

5. The data processing method for smart city according to claim 1, characterized in that: According to the urban dynamic feature data, point-level, sequence-level and group-level abnormal feature libraries are constructed, and abnormal events are identified through multi-dimensional feature comparison and comprehensive evaluation to obtain abnormal pattern data, including: The urban dynamic characteristic data is divided into single point data, time series data and group data according to data characteristics, and the data characteristics are classified and counted to obtain characteristic classification data; Performing numerical distribution analysis on single-point data in the feature classification data, establishing numerical fluctuation range limits, extracting outlier features, and obtaining point-level feature data; Performing time correlation analysis on the time series data in the point-level feature data, extracting the change characteristics of the data sequence, establishing a sequence feature description, and obtaining sequence-level feature data; Performing group behavior analysis on the group data in the sequence-level feature data, identifying the distribution pattern of the group data, extracting group features, and obtaining group-level feature data; The group-level feature data is subjected to multi-dimensional feature fusion, the correlation between the features is calculated, and feature weights are assigned to obtain feature association data; Abnormal events are identified on the feature-related data, an abnormal scoring standard is established, and a comprehensive evaluation and judgment are performed to obtain abnormal pattern data.

6. The data processing method for smart city according to claim 1, characterized in that: According to the abnormal mode data, computing resources are allocated under the edge computing architecture, and distributed optimization decisions are made through multi-node collaborative computing to obtain urban management decision data, including: The abnormal pattern data is graded according to the computational complexity and the data size, and the processing priorities of data at different levels are determined to obtain graded processing data; Perform load calculation on the hierarchical processing data, count the computing power and network bandwidth occupancy of each node, and obtain resource status data; Dividing the computing tasks in the resource status data according to priority, dynamically allocating computing resources of each node, and obtaining task allocation data; Performing data exchange and task coordination between nodes on the task allocation data, determining a data transmission scheme between nodes, and obtaining collaborative computing data; Processing the collaborative computing data in parallel on each node, aggregating the computing results of each node, merging the data, and obtaining distributed processing data; Decision analysis is performed on the distributed processing data, processing results of each node are integrated, a management decision plan is generated, and urban management decision data is obtained.

7. The data processing method for smart city according to claim 1, characterized in that: Based on the urban management decision data, an evaluation index system of system performance, service quality and user satisfaction is established, and the decision strategy is dynamically adjusted and continuously optimized to obtain an optimized urban system management plan, including: Classifying the urban management decision data according to system performance indicators, service quality indicators and user satisfaction indicators, and performing statistical analysis on the data to obtain classification evaluation data; Calculate the system performance indicators in the classified evaluation data, count the system response time and resource utilization, generate performance statistics, and obtain system performance evaluation data; Quantifying the service quality indicators in the system performance evaluation data, calculating the service accuracy and service timeliness, integrating the service evaluation data, and obtaining service quality evaluation data; Analyze the user satisfaction index in the service quality evaluation data, collect user feedback data and complaint data, and organize the data to obtain user satisfaction evaluation data; The user satisfaction evaluation data is comprehensively calculated according to the evaluation index weights, and the decision-making strategy is evaluated and scored to obtain strategy evaluation data; The strategy evaluation data is optimized and analyzed, decision parameters are adjusted, a management strategy plan is generated, and an optimized urban system management plan is obtained.

8. The data processing method for smart city according to claim 7, characterized in that: The system performance indicators in the classified evaluation data are calculated, the system response time and resource utilization are counted, and performance statistics are generated to obtain system performance evaluation data, including: The system response time in the classified evaluation data is grouped according to the transaction type, different types of response delays are counted, and the average response time is summarized to obtain response time data; Sampling the processor occupancy in the response time data, counting the peak and average usage of the central processor, recording the usage fluctuation, and obtaining processor statistical data; Monitor the memory usage in the processor statistics, count the memory occupancy rate and release rate, record the memory allocation status, and obtain memory statistics; Analyze the storage space usage in the memory statistical data, calculate the statistical data writing speed and reading speed, record the storage capacity change, and obtain storage statistical data; Record the network transmission conditions in the stored statistical data, calculate the statistical data throughput and network bandwidth occupancy, calculate the network load conditions, and obtain network statistical data; Perform a comprehensive performance analysis on the network statistical data, calculate the scores of various performance indicators, generate a performance evaluation report, and obtain system performance evaluation data.

9. A data processing system for a smart city, used to implement the data processing method for a smart city as described in any one of claims 1 to 8, characterized in that: The data processing system for smart city includes: The acquisition module is used to collect multi-source urban data through a distributed IoT sensor network, clean the original data according to data quality assessment rules, and perform format conversion and spatiotemporal alignment processing on the cleaned data according to preset data standards to obtain standardized urban data; A matching module is used to establish a city domain concept system based on the standardized city data, map and match data entities from different sources, and perform knowledge representation and reasoning in combination with multimodal information to obtain semantically related data; An analysis module is used to use the semantic association data to perform periodicity and trend analysis on the time series, extract regional features in combination with spatial location information, establish an interaction relationship model between regions, and obtain urban dynamic feature data; An identification module is used to construct point-level, sequence-level and group-level abnormal feature libraries based on the urban dynamic feature data, identify abnormal events through multi-dimensional feature comparison and comprehensive evaluation, and obtain abnormal pattern data; An allocation module is used to allocate computing resources under the edge computing architecture according to the abnormal mode data, and to make distributed optimization decisions through multi-node collaborative computing to obtain urban management decision data; The optimization module is used to establish an evaluation index system for system performance, service quality and user satisfaction based on the urban management decision data, dynamically adjust and continuously optimize the decision-making strategy, and obtain an optimized urban system management plan.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the data processing method for a smart city as described in any one of claims 1 to 8 is implemented.

Images