A smart city processing method and system based on a multi-modal large model

Abstract

The application discloses a kind of wisdom city processing method and system based on multi-modal large model, and it is related to wisdom city technical field.In the method, target facility is obtained from target area;First prediction object is determined according to target facility;A plurality of acquisition devices are determined according to the first prediction object, and a plurality of monitoring data collected by the plurality of acquisition devices are obtained, and the first position corresponding to the target area is obtained;The plurality of monitoring data and the first position are input into the multi-modal large model for processing to obtain an analysis result;Predicted safety accidents are determined according to the analysis result, and the first emergency treatment scheme is determined based on the predicted safety accidents;Safety level is determined according to the processing duration and the predicted safety accidents, and the first emergency treatment scheme is sent to the first department based on the safety level, so that the personnel of the first department can view the first emergency treatment scheme.The implementation of the technical solution provided can significantly solve the problem of low processing efficiency in the traditional analysis framework.

Description

Technical Field

[0001] This application relates to the field of smart city technology, specifically to a smart city processing method and system based on a multimodal large model. Background Technology

[0002] Smart cities, as a new urban development model based on advanced information and communication technologies (ICT), aim to improve urban operational efficiency, enhance residents' quality of life, and achieve sustainable development goals through cross-domain data fusion, system integration, and technological innovation.

[0003] Smart cities leverage technologies such as the Internet of Things (IoT), cloud computing, and big data to integrate operational data from urban infrastructure sectors like transportation, healthcare, education, and energy, constructing a unified data governance framework. Building upon this, artificial intelligence algorithms, including machine learning and deep learning, are used to analyze massive amounts of heterogeneous data in real time, uncovering potential patterns and risks, and providing intelligent decision support for urban governance. For example, intelligent transportation systems dynamically generate traffic situation maps by integrating data from road condition sensors, vehicle terminals, and mobile devices, enabling congestion prediction and route optimization. While data fusion brings convenience to urban governance, existing analysis processes rely on domain expert experience for anomaly detection. As data volume grows exponentially, traditional analysis frameworks become increasingly time-consuming, resulting in low analysis efficiency.

[0004] Therefore, there is an urgent need for a smart city processing method and system based on a multimodal large model that can solve the above-mentioned technical problems. Summary of the Invention

[0005] This application provides a smart city processing method and system based on a multimodal large model. This method inputs monitoring data automatically acquired by the acquisition device into a multimodal large model for correlation analysis and generates analysis results in real time. It can significantly solve the problems of reliance on expert experience and low processing efficiency in traditional analysis frameworks.

[0006] Firstly, this application provides a smart city processing method based on a multimodal large model. The method includes: acquiring target facilities from a target area, where the target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities, and industrial building facilities, and the target area is any area corresponding to a city; determining a first prediction object based on the target facilities, where the first prediction object includes traffic accidents, medical accidents, natural disaster accidents, construction accidents, public safety accidents, and park safety accidents; determining multiple acquisition devices based on the first prediction object, acquiring multiple monitoring data collected by the multiple acquisition devices, and acquiring a first location corresponding to the target area; inputting the multiple monitoring data and the first location into the multimodal large model for processing to obtain analysis results; determining predicted safety accidents based on the analysis results, and determining a first emergency response plan based on the predicted safety accidents; determining a safety level based on the processing time and the predicted safety accidents, and sending the first emergency response plan to a first department based on the safety level so that the personnel in the first department can view the first emergency response plan, where the processing time is the total time corresponding to processing the first emergency response plan.

[0007] By adopting the above technical solution, the first prediction target is determined based on the type of target facility in the target area. Based on the first prediction target, multiple acquisition devices are used to automatically acquire monitoring data to cover all relevant information. The acquired monitoring data and the first location are then input into a multimodal large model for extraction and correlation analysis without manual intervention. The analysis results are then used to predict safety accidents. Based on the predicted safety accidents, corresponding emergency response plans are quickly generated. The predicted safety accidents and their handling time are automatically classified into safety levels. Emergency resources are dynamically allocated based on the safety levels, and relevant departments are notified in real time via message push. Using a multimodal large model for prediction can replace expert experience, solving the problem of inefficiency caused by the traditional reliance on domain experts to manually label abnormal data and analyze multi-source data one by one.

[0008] Optionally, after determining the first prediction object based on the target facility, the method further includes: obtaining the target factors currently corresponding to the target area, including time factors and weather factors; determining the second prediction object based on the target factors and the target facility; determining whether the second prediction object exists among the first prediction objects; and when the second prediction object exists among the first prediction objects, prioritizing monitoring the target area according to the second prediction object.

[0009] By adopting the above technical solution, the priority of the predicted objects is dynamically adjusted according to the current time and weather. Traditional solutions, which fix the monitoring of preset accident types, cannot cope with sudden environmental changes and lead to waste of resources. By integrating multi-dimensional data of time, weather and facility type, more accurate risk prediction can be achieved, high-risk objects can be monitored first, early warning can be achieved, and the accuracy of accident prediction can be significantly improved.

[0010] Optionally, before acquiring multiple monitoring data collected by multiple acquisition devices and obtaining the first location corresponding to the target area, the method further includes: determining the data type based on the first prediction object, including data types such as electricity consumption data, traffic data, personnel location data, environmental data, and construction data; inputting the data type into a preset frequency database for matching to obtain the target acquisition frequency; determining the acquisition device based on the data type and sending the target acquisition frequency to each acquisition device so that the acquisition device can collect data according to the target acquisition frequency.

[0011] By adopting the above technical solution, the collection frequency is dynamically adjusted according to the data sensitivity of different prediction objects. Traditional solutions use a fixed collection frequency, which leads to insufficient collection of highly sensitive data or redundancy of low-sensitivity data, increasing storage and computing costs. The system automatically allocates collection devices according to data type and dynamically adjusts collection resources based on scenario requirements, which significantly improves the efficiency, accuracy and resource utilization of data collection.

[0012] Optionally, the safety level can be determined based on the processing time and predicted safety incidents, specifically including:

[0013] Extract core data from the first emergency response plan; determine a preset sensitive data table based on the predicted safety incident, determine whether core data exists in the preset sensitive data table, and determine whether the processing time is less than or equal to the preset time; if core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, determine that the first emergency response plan is at a high-risk level, and output the high-risk level as the safety level; if core data exists in the preset sensitive data table and the processing time is greater than the preset time, determine that the first emergency response plan corresponds to a medium-risk level, and output the medium-risk level as the safety level.

[0014] By adopting the above technical solution, the risk level is automatically determined based on the preset sensitive data table and processing time, avoiding the subjectivity and delay of manual assessment. By using the preset sensitive data table, risk assessment is only performed on key data to avoid interference from irrelevant data. Emergency resource allocation is dynamically adjusted according to the risk level. Traditional solutions may lead to resource waste or insufficient resources due to misjudgment of risk level, which significantly improves the assessment accuracy and response efficiency of emergency response solutions.

[0015] Optionally, the first emergency response plan is sent to the first department based on the security level, specifically including: when the first emergency response plan corresponds to a high-risk level, performing a first encryption on the first emergency response plan according to the high-risk level to obtain a second emergency response plan, the first encryption including quantum encryption; obtaining the second department, which is the department involved in the first emergency response plan; screening multiple preset departments that are the same as the second department from the target area, one preset department corresponding to one location; obtaining the target permission level corresponding to the third department, the third department being any one of the multiple preset departments; determining whether the target permission level is consistent with the preset permission level, the preset permission level being the permission level corresponding to the high-risk level; when the target permission level is consistent with the preset permission level, determining to allocate the second emergency response plan to the third department, and outputting the third department as the first department.

[0016] By adopting the above technical solution, when an emergency response plan is determined to be high-risk, quantum encryption is automatically triggered to generate a second emergency response plan. The characteristics of quantum encryption ensure the security of both transmission and storage, while traditional encryption may lead to data leakage due to key leaks or algorithm cracking. Through dynamic comparison of the target permission level with the preset permission level, it ensures that only authorized departments can access the high-risk plan. By screening preset departments that are identical to the second department within the target area and combining this with permission level matching, dynamic allocation of emergency resources is achieved. The entire process is executed automatically by the system, reducing human intervention. Through dynamic matching of preset departments and permission levels, multi-department collaboration and cross-regional linkage are supported, significantly improving the security, execution efficiency, and collaborative capabilities of the emergency response plan.

[0017] Optionally, based on the security level, the first emergency response plan is sent to the first department. Specifically, this includes: when the first emergency response plan corresponds to a medium-risk level, performing a second encryption on the first emergency response plan according to the medium-risk level to obtain a third emergency response plan. The second encryption only encrypts the core words in the first emergency response plan. The response plan consists of core words and non-core words, with the core words being words appearing in a preset sensitive information table; obtaining the second location corresponding to the predicted security incident, which is the specific location of the predicted security incident in the target area; sequentially calculating the distances between multiple preset departments and the second location to obtain multiple distance values; sorting the multiple distance values ​​in ascending order to obtain a target sorting result; obtaining the fourth department, which is the distance value ranked first in the target sorting result, and sending the third emergency response plan to the fourth department.

[0018] By adopting the above technical solution, when the emergency response plan is at a medium risk level, only the core words appearing in the preset sensitive information table in the emergency response plan are encrypted, rather than the entire emergency response plan. Through targeted encryption, the security of sensitive information is ensured while reducing the amount of encryption calculation and improving response efficiency. By calculating the distance from multiple preset departments to the accident location and sorting them from smallest to largest distance, the third emergency response plan is sent to the nearest department first. Emergency resources are dynamically allocated based on the distance sorting results.

[0019] Optionally, after determining the first prediction object based on the target facility, the method further includes: acquiring target activities in the target area within a preset time period, including concert activities, sports events, and exhibition activities; combining the target activities with the target facilities to obtain target combination information; performing simulation analysis on the target combination information to obtain a third prediction object, and prioritizing monitoring the target area according to the third prediction object.

[0020] By adopting the above technical solutions, target activities (such as concerts, sports meets, and exhibitions) are combined with target facilities. Through simulation analysis, potential high-risk scenarios are identified. Traditional solutions may neglect combined risks due to the lack of correlation analysis between activities and facilities, or respond passively after risks occur. Based on the third prediction object, the target area is monitored first. Based on the simulation analysis results, emergency plans for high-risk combinations are formulated in advance. By identifying high-risk combinations in advance and formulating countermeasures, the probability of accidents is reduced.

[0021] A second aspect of this application provides a smart city processing system based on a multimodal large model. The system includes an acquisition unit, a processing unit, and a sending unit. The acquisition unit acquires target facilities from a target area. The target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities, and industrial building facilities. The target area is any area corresponding to a city. The processing unit determines a first prediction object based on the target facilities. The first prediction object includes traffic accidents, medical accidents, natural disaster accidents, construction accidents, public safety accidents, and park safety accidents. Based on the first prediction object, it determines multiple acquisition devices and acquires multiple monitoring data collected by the multiple acquisition devices, and acquires a first location corresponding to the target area. It inputs the multiple monitoring data and the first location into the multimodal large model for processing to obtain analysis results. Based on the analysis results, it determines predicted safety accidents and determines a first emergency response plan based on the predicted safety accidents. The sending unit determines a safety level based on the processing time and the predicted safety accidents, and sends the first emergency response plan to a first department based on the safety level so that the personnel in the first department can view the first emergency response plan. The processing time is the total time corresponding to processing the first emergency response plan.

[0022] Optionally, the acquisition unit is used to acquire the target factors currently corresponding to the target area, including time factors and weather factors; the processing unit is used to determine the second prediction object based on the target factors and target facilities; determine whether the second prediction object exists in the first prediction object; when the second prediction object exists in the first prediction object, the target area is monitored according to the second prediction object first.

[0023] Optionally, the processing unit is used to determine the data type based on the first prediction object, including electricity consumption data type, traffic data type, personnel location data type, environmental data type, and construction data type; input the data type into a preset frequency database for matching to obtain the target acquisition frequency; determine the acquisition device based on the data type, and send the target acquisition frequency to each acquisition device so that the acquisition device can acquire data according to the target acquisition frequency.

[0024] Optionally, the acquisition unit is used to extract core data from the first emergency response plan; the processing unit is used to determine a preset sensitive data table based on the predicted safety accident, determine whether core data exists in the preset sensitive data table, and determine whether the processing time is less than or equal to the preset time; when core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, the first emergency response plan is determined to be at a high-risk level, and the high-risk level is output as a safety level; when core data exists in the preset sensitive data table and the processing time is greater than the preset time, the first emergency response plan is determined to be at a medium-risk level, and the medium-risk level is output as a safety level.

[0025] Optionally, the processing unit is used to perform a first encryption on the first emergency response plan according to the high-risk level when the first emergency response plan corresponds to a high-risk level, to obtain a second emergency response plan, wherein the first encryption includes quantum encryption; the acquisition unit is used to acquire a second department, which is the department involved in the first emergency response plan; the processing unit is used to screen multiple preset departments that are the same as the second department from the target area, with one preset department corresponding to one location; the acquisition unit is used to acquire the target permission level corresponding to the third department, which is any one of the multiple preset departments; the processing unit is used to determine whether the target permission level is consistent with the preset permission level, wherein the preset permission level is the permission level corresponding to the high-risk level; the sending unit is used to determine to allocate the second emergency response plan to the third department when the target permission level is consistent with the preset permission level, and output the third department as the first department.

[0026] Optionally, the processing unit is used to perform a second encryption on the first emergency response plan when the first emergency response plan corresponds to a medium-risk level, to obtain a third emergency response plan. The second encryption is to encrypt only the core words in the first emergency response plan. The processing plan consists of core words and non-core words, and the core words are words that appear in a preset sensitive information table. The acquisition unit is used to acquire the second location corresponding to the predicted safety accident, which is the specific location of the predicted safety accident in the target area. The distances between multiple preset departments and the second location are calculated sequentially to obtain multiple distance values. The multiple distance values ​​are sorted in ascending order to obtain the target sorting result. The sending unit is used to acquire the fourth department, which is the distance value ranked first in the target sorting result, and send the third emergency response plan to the fourth department.

[0027] Optionally, the acquisition unit is used to acquire target activities in the target area within a preset time period, including concert activities, sports events, and exhibition activities; the processing unit is used to combine the target activities with the target facilities to obtain target combination information; to perform simulation analysis on the target combination information to obtain a third prediction object, and to prioritize monitoring the target area according to the third prediction object.

[0028] In a third aspect, this application provides an electronic device including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory, causing the electronic device to perform any of the methods described above in this application.

[0029] In a fourth aspect, this application provides a computer-readable storage medium storing instructions that, when executed, perform any of the methods described above in this application.

[0030] In summary, one or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0031] 1. The primary prediction target is determined based on the type of target facility in the target area. Multiple acquisition devices are then used to automatically acquire monitoring data, covering all relevant information. The acquired monitoring data and the primary location are input into a multimodal large-scale model for extraction and correlation analysis without manual intervention. Analysis results are then used to predict safety incidents and quickly generate corresponding emergency response plans. The predicted safety incidents and their response times are automatically categorized into safety levels, and emergency resources are dynamically allocated based on these levels. Real-time notifications to relevant departments are sent via push notifications. Using a multimodal large-scale model for prediction can replace expert experience, addressing the inefficiency of traditional methods that rely on domain experts manually labeling abnormal data and analyzing multiple data sources one by one.

[0032] 2. The priority of predicted objects is dynamically adjusted according to the current time and weather. Traditional methods monitor fixed accident types, which cannot cope with sudden environmental changes and lead to waste of resources. By integrating multi-dimensional data of time, weather and facility type, more accurate risk prediction can be achieved. High-risk objects are monitored first to achieve early warning and significantly improve the accuracy of accident prediction. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating a smart city processing method based on a multimodal large model provided in an embodiment of this application;

[0034] Figure 2 This is a schematic diagram of the structure of a smart city processing system based on a multimodal large model provided in an embodiment of this application;

[0035] Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0036] Explanation of reference numerals in the attached drawings: 201, acquisition unit; 202, processing unit; 203, transmission unit; 300, electronic device; 301, processor; 302, memory; 303, user interface; 304, network interface; 305, communication bus. Detailed Implementation

[0037] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0038] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0039] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0040] Smart cities, as a new urban development model based on advanced information and communication technologies (ICT), aim to improve urban operational efficiency, enhance residents' quality of life, and achieve sustainable development goals through cross-domain data fusion, system integration, and technological innovation.

[0041] Smart cities leverage technologies such as the Internet of Things (IoT), cloud computing, and big data to integrate operational data from urban infrastructure sectors like transportation, healthcare, education, and energy, constructing a unified data governance framework. Building upon this, artificial intelligence algorithms, including machine learning and deep learning, are used to analyze massive amounts of heterogeneous data in real time, uncovering potential patterns and risks, and providing intelligent decision support for urban governance. For example, intelligent transportation systems dynamically generate traffic situation maps by integrating data from road condition sensors, vehicle terminals, and mobile devices, enabling congestion prediction and route optimization. Relying on a unified government service platform, smart cities promote the full-process digitalization of government services, achieving "one-stop online processing." Citizens can complete identity authentication, document submission, and progress tracking online through mobile terminals or self-service terminals, breaking down time and space limitations. For instance, electronic certificate systems based on blockchain technology can achieve cross-departmental data sharing and business collaboration, significantly shortening approval cycles and improving the accessibility of public services.

[0042] While data fusion has brought convenience to urban governance, the detection of data anomalies in the existing analysis process relies on the experience of domain experts. As the scale of data grows exponentially, traditional analysis frameworks take a long time, resulting in low analysis efficiency.

[0043] Therefore, how to solve the problems of reliance on expert experience and low processing efficiency in traditional analysis frameworks? This application provides a smart city processing method based on a multimodal large model, applied in a server. The server in this application is a platform providing smart city services. Figure 1 This is a flowchart illustrating a smart city processing method based on a multimodal large model provided in an embodiment of this application. (Refer to...) Figure 1 The method includes the following steps S101-S106.

[0044] S101: Obtain the target facility from the target area.

[0045] In the aforementioned S101, when monitoring the real-time situation of various cities, due to the differences in the internal buildings and geographical locations of different cities, it is necessary to combine the city's buildings and geographical locations to determine potential safety accidents and make timely predictions when predicting various situations in each city. This application uses the monitoring of various data of a certain city as an example. First, the monitoring area is determined, i.e., the target area, which refers to any city or a certain area within a city. Then, through urban geographic information systems (GIS), architectural drawings, satellite imagery, Internet of Things sensors, etc., information on target facilities within the target area is obtained. Target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities, and industrial building facilities. Based on the function of the facilities, various buildings in real life are divided into five categories: public service facilities include schools, hospitals, libraries, fire stations, etc.; commercial facilities include shopping malls, office buildings, hotels, restaurants, etc.; residential facilities include residential communities, apartments, villas, etc.; productive building facilities include factories, warehouses, logistics centers, etc.; and industrial building facilities include chemical plants, power plants, oil refineries, etc. Spatial analysis can be performed using GIS platforms (such as ArcGIS and QGIS), and facility attributes can be extracted by combining them with Building Information Modeling (BIM).

[0046] For example, the target area is District A of a city. Then, we need to find out what target facilities exist in District A. If District A includes a people's hospital, a primary school, a residential community, and a chemical plant, then we can categorize the facilities into five major facility categories based on their functions.

[0047] S102: Determine the first prediction target based on the target facility. The first prediction target includes traffic accidents, medical accidents, natural disaster accidents, construction accidents, public safety accidents, and park safety accidents.

[0048] In step S102 above, after obtaining the target facility from the target area, the types of safety accidents it may cause can be analyzed according to the facility type. Here, the type of safety accident refers to the first predicted object. For example, public service facilities may cause medical accidents or natural disasters, while commercial facilities may cause fire accidents or public safety accidents. Residential facilities may cause gas leak accidents or falling objects from heights, while production buildings may cause construction accidents or equipment failure accidents, and industrial buildings may cause chemical leak accidents or explosion accidents. For example, if the target facility is a chemical plant, the first predicted object is a chemical leak accident or an explosion accident.

[0049] In addition to directly analyzing potential safety incidents based on facility type, safety incidents can be dynamically adjusted based on time and weather within the target area. Priority should be given to monitoring more urgent incidents for timely detection and early warning. Specifically, this involves: acquiring the current target factors for the target area, including time and weather factors; determining a second prediction target based on the target factors and target facilities; determining whether a second prediction target exists within the first prediction target; and prioritizing monitoring the target area according to the second prediction target when one exists within the first. Specifically, the current time within the target area is first acquired to determine if it is during peak hours, holidays, or special periods. Then, real-time weather data for the target area is acquired, including heavy rain, high temperatures, strong winds, fog, and sunny weather. The current time can be obtained through the system clock or combined with historical data to determine special periods (such as the Spring Festival travel rush or peak tourist season). Peak tourist seasons can be analyzed by combining data with tourist attractions in the target area. Real-time weather data can also be obtained through the meteorological bureau's API. For example, the target area is District A of a city, the current time is 18:00 on October 1st, 202x (National Day holiday), and the current weather is heavy rain (50mm / hour). Combining the target factors (time and weather) with the target facilities, we analyze the types of safety accidents that may occur. For example, if the current time is during a holiday, the shopping mall will be crowded, potentially causing stampedes; if the current time is during the evening rush hour, potentially causing traffic accidents. When the weather in the target area is heavy rain, low-lying areas may experience flooding, chemical plants may cause chemical leaks, and when the weather is hot, open-air construction areas may cause heatstroke. Based on the correlation analysis results, we determine the possible types of safety accidents. The target facilities are a plaza and a chemical plant. Given the target factors of a holiday and heavy rain, the second predicted accident for the plaza is a stampede, and for the chemical plant, it is a chemical leak. Since the first predicted accident (such as traffic accidents or medical accidents) has already been determined based on the target facilities, we then compare the first and second predicted accidents to determine if there is any overlap. If there is overlap, the second predicted accident is prioritized for monitoring. When there is no overlap, monitoring can continue according to the first predicted object. For example, if the target facility is a square, the corresponding first predicted object is a public safety accident; if the target facility is a chemical plant, the corresponding first predicted object is a chemical leak accident. Introducing time and weather factors, and then performing correlation analysis between time and weather factors and the target facility, when the target facility is a square, the second predicted object is a stampede accident; when the target facility is a chemical plant, the second predicted object is a chemical leak accident. The first and second predicted objects are then compared. When the target facility is a square, since the first predicted object (public safety accident) includes the second predicted object (stampede accident), the second predicted object should be monitored first.When the target facility is a chemical plant, the first and second predicted targets are the same, and monitoring continues. If the second predicted target exists, the priority of monitoring equipment is adjusted, focusing on areas affected by time and weather factors. This avoids the situation where areas with potential safety incidents cannot be detected in a timely manner without prioritization, ensuring that high-risk predicted targets receive priority monitoring and improving emergency response efficiency. For example, during heavy rains like the National Day holiday, chemical leaks at chemical plants can be detected and warned of in a timely manner by increasing the monitoring frequency of gas sensors.

[0050] Furthermore, spatial association can be established between activities taking place in the target area and target facilities to avoid overlooking high-risk areas during prediction. This includes: acquiring target activities in the target area within a preset timeframe, including concerts, sporting events, and exhibitions; combining target activities with target facilities to obtain target combination information; and performing simulation analysis on the target combination information to obtain a third prediction object, prioritizing monitoring of the target area based on this third prediction object. Specifically, first, the current primary location corresponding to the target area is determined. Then, web crawling technology is used to retrieve activity information taking place within this primary location within a preset timeframe from relevant platforms. This information can be obtained from official event approval records from the government or venues, ticketing platforms, and promotional information from social media platforms. The preset timeframe can be one week, one month, or a specific period (such as holidays). After acquiring the activity data, duplicate and invalid activity information is removed. Activities include concerts, sporting events, and exhibitions. Next, acquire the target facilities within the target area. Commercial facilities include venues (such as stadiums), while public service facilities include roads (roads surrounding stadiums), parking lots, and public transportation stations. Associate each activity with its corresponding facility to form an activity-facility combination, where the corresponding facility refers to the location where the activity is held. Geographic Information System (GIS) can be used to match activity locations with facility locations. For example, if the activity is named a science and technology exhibition and the location is a science and technology museum, the target facility is the science and technology museum. Combining the activity information with the target facility yields a science and technology exhibition-science and technology museum combination. Then, predict the potential risks or impacts of each activity-facility combination to generate a third prediction object (such as pedestrian flow, traffic congestion, and safety hazards). Data from similar past activities (such as pedestrian flow and accident records) can be referenced. Traffic simulation software (such as VISSIM) can also be used to simulate pedestrian and vehicle flow during the event. Since the activity in this case is a science and technology exhibition, the equipment of the science and technology exhibition and the science and technology museum are correlated to analyze potential safety accidents. Therefore, the third prediction object identifies risks such as facility aging and blocked fire lanes. Once the third predicted event is obtained, the target area can be monitored based on this event, and fire lanes and power facilities can be inspected in real time. For example, during a celebrity concert, if the current third predicted event is determined to be a traffic accident, then traffic and pedestrian flow on the roads surrounding the concert will be predicted first. Based on the predicted peak pedestrian flow, security measures will be automatically adjusted to ensure the safety of the event.

[0051] S103: Determine multiple acquisition devices based on the first prediction object, acquire multiple monitoring data collected by the multiple acquisition devices, and acquire the first location corresponding to the target area.

[0052] In step S103 above, after determining the first predicted object based on the actual situation of the target area, appropriate monitoring equipment is selected according to the first predicted object. For example, when the first predicted object is a chemical leak accident, the equipment for collecting data on chemical leak accidents includes gas sensors, liquid level sensors, etc. Then, the monitoring data collected by the collection equipment is obtained in real time through an IoT platform or by direct connection to the equipment. When collecting monitoring data, it is necessary to obtain the location of the first predicted object in the target area, and then establish a connection with the collection equipment installed at that location.

[0053] Furthermore, before using data acquisition equipment to obtain monitoring data, the acquisition frequency affects the subsequent data analysis results. Therefore, after determining the data acquisition equipment, it is necessary to adjust the acquisition frequency of each equipment to ensure that the acquired monitoring data can be used for subsequent analysis and to avoid situations where there is insufficient data for analysis. Specifically, this includes: determining the data type based on the first prediction object, including data types such as electricity consumption, traffic, personnel location, environment, and construction; inputting the data types into a preset frequency database for matching to obtain the target acquisition frequency; determining the acquisition equipment based on the data type and sending the target acquisition frequency to each acquisition equipment so that the acquisition equipment can collect data according to the target acquisition frequency. Specifically, based on the target facilities within the target area, the types of safety accidents that may occur are predicted, i.e., the first prediction object. The first prediction object is associated with the data types to be monitored: traffic accidents require traffic data types (traffic flow, vehicle speed, road congestion index); fires require environmental data types (temperature, smoke concentration, oxygen content); chemical leaks require environmental data types (gas concentration, liquid level, pressure); and risks in densely populated areas require personnel location data types (population density, dwell time). For example, the first predicted event might be a chemical leak in an industrial park. The data types leading to this leak include environmental and construction-related data. Environmental data includes gas concentration, liquid level, and pressure, while construction-related data includes pipeline pressure and valve status. A pre-built frequency database is constructed, containing the correspondence between each data type and its acquisition frequency. For instance, for environmental data, gas concentration can be set to 1 time / minute or 1 time / hour, and temperature can be set to 1 time / 5 minutes; for traffic data, traffic flow can be set to 1 time / second or 1 time / 10 seconds. Different acquisition frequencies are selected based on different data types. The corresponding acquisition frequency is selected from the database based on the first predicted event, ensuring that the collected monitoring data is available for subsequent data analysis. The first predicted event can also be classified by risk level. For example, a chemical leak could be classified as high-risk, allowing the acquisition frequency to be set to 1 time / minute. Then, corresponding hardware devices are selected based on the data type. For environmental data, gas sensors and temperature sensors are used. For traffic data, cameras and geomagnetic sensors are used. For personnel location data, Wi-Fi probes and Bluetooth beacons are used. The target sampling frequency is sent to the acquisition device via IoT protocols (such as MQTT, CoAP) or API interfaces. The acquisition device adjusts its sampling period based on the received frequency. For example, a gas sensor might adjust its sampling frequency from once per hour to once per minute.

[0054] After determining the acquisition frequency, the data acquisition equipment acquires relevant monitoring data according to the frequency and sends the acquired data to the server. It also needs to obtain the first location corresponding to the target area. Here, the first location refers to the location of the facility corresponding to the first predicted object within the target area. For example, if the target facility in the target area is a chemical plant, and the first predicted object is determined to be a chemical plant leak accident, the data acquisition equipment is determined based on the chemical plant leak accident. The data acquisition equipment includes gas sensors (monitoring chemical leaks) and pressure sensors (monitoring explosion risks). The gas sensor is then used to acquire the ammonia concentration of 15 ppm in a certain area, and the pressure sensor is used to acquire the pressure of a storage tank of 2.5 MPa. The acquired monitoring data are then transmitted. At this point, it is also necessary to obtain the specific location of the chemical plant within the target area, i.e., the first location.

[0055] S104: Input multiple monitoring data and the first location into the multimodal large model for processing to obtain the analysis results.

[0056] In step S104 above, after receiving multiple monitoring data points from the acquisition device, before inputting the multiple monitoring data points and the first location information into the multimodal large-scale model for processing, a multimodal large-scale model needs to be constructed. This model uses a large-scale model such as BERT, combining text (monitoring data), images (facility images), and geographic locations (GIS data) for comprehensive analysis. Initially, various historical monitoring data, historical accidents, and geographic locations need to be collected to obtain a data sample set. Then, an initial model is used to train the data sample set. When the loss value between the training result and the actual result meets the convergence condition, training ends, and the initial model at the end of training is output as the multimodal large-scale model. Next, the multiple monitoring data points and the first location information are converted into a format understandable by the multimodal large-scale model. Then, key features (such as ammonia concentration and pressure values) are extracted from the data. Based on these features, the possible types and probabilities of safety accidents are predicted; here, the probability of occurrence refers to both the type and probability of the safety accident. For example, if the input data shows an ammonia concentration of 15 ppm and abnormal tank pressure, the layout and location of the chemical plant are also input into the multimodal large model for processing. The analysis results show that the chemical plant has a chemical leakage risk with a probability of 70%.

[0057] S105: Based on the analysis results, determine the predicted safety accidents, and based on the predicted safety accidents, determine the first emergency response plan.

[0058] In step S105 above, the accident type and probability of occurrence are obtained from the analysis results. The accident type and probability of occurrence are then compared with a preset probability of occurrence. If the probability of occurrence is less than the preset probability, the probability of the accident is determined to be low, and monitoring of the area continues, with the prediction of no safety accidents. If the probability of occurrence is equal to or greater than the preset probability, the accident type in the analysis results is determined to be a predicted safety accident. In this case, a predicted safety accident refers to a safety accident that is highly likely to occur. In the example above, the chemical plant has a chemical leak risk with a probability of 70%, and the predicted safety accident is a chemical leak accident. Based on the predicted safety accident type, a corresponding emergency response plan is formulated. When the predicted safety accident is a chemical leak accident, the first emergency response plan includes evacuating surrounding residents; activating the emergency plan and notifying fire, environmental protection, and other departments; shutting off the leak source and carrying out emergency repairs. When the predicted safety accident is an explosion accident, the first emergency response plan includes activating the emergency evacuation procedure; notifying medical, fire, and other departments to stand by; and cutting off the power supply to prevent secondary accidents.

[0059] S106: Determine the safety level based on the processing time and predicted safety incidents, and send the first emergency response plan to the first department based on the safety level so that the corresponding personnel in the first department can view the first emergency response plan.

[0060] In S106 above, the processing time is the total time corresponding to the processing of the first emergency response plan. After determining the first emergency response plan based on the predicted safety accident, it is necessary to obtain the response time of the first emergency response plan, that is, how long each department needs to complete the response plan. The safety level can be divided according to the severity and impact of the predicted safety accident, or it can be determined according to the processing time and the predicted safety accident. Specifically, this includes: extracting core data from the first emergency response plan; determining a preset sensitive data table based on the predicted safety accident, determining whether the core data exists in the preset sensitive data table, and determining whether the processing time is less than or equal to the preset time; when the core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, the first emergency response plan is determined to be at a high-risk level, and the high-risk level is output as the safety level; when the core data exists in the preset sensitive data table and the processing time is greater than the preset time, the first emergency response plan is determined to be at a medium-risk level, and the medium-risk level is output as the safety level.

[0061] Specifically, the process begins by obtaining the first emergency response plan, which includes the issuing department, monitoring time, data collection frequency, probability of occurrence, and predicted time of occurrence. Then, Natural Language Processing (NLP) technology is used to extract structured data from the emergency response plan text, match key fields, identify high-risk keywords, and summarize the key information from the first emergency response plan to obtain core data. This core data includes processing time, handling measures, affected areas, and resource requirements. Processing time refers to the time required to complete the emergency response plan, such as 30 minutes; handling measures refer to specific emergency measures, such as evacuating personnel; affected areas refer to the scope of the accident's impact, such as a specific floor; and resource requirements refer to the required manpower and materials, such as 3 firefighters and 2 fire extinguishers. After obtaining the core data, a sensitive field table is created based on predicted safety accidents. This table is a database linking each historical accident to its risk level, identifying which fields correspond to high risk and which do not. The core data is then matched against the preset sensitive data table to check if the core data contains any of the fields listed in the table. For example, core data includes a processing time of 45 minutes, valve closure, personnel evacuation, and a specific workshop. The pre-set sensitive data database includes gas concentration, personnel density, and areas involving hazardous zones. In this case, the pre-set sensitive data table contains the core data. The extracted processing time is then compared with the pre-set time. The pre-set time is the permissible processing time for a normal chemical leak. When the processing time is 45 minutes and the pre-set time is 60 minutes, the processing time is less than or equal to the pre-set time, indicating a high-risk first emergency response plan. When the processing time is longer than the pre-set time, the first emergency response plan is identified as medium-risk. When the core data is not present in the pre-set sensitive data table, and the processing time is longer than the pre-set time, the first emergency response plan is identified as low-risk. When the core data is not present in the pre-set sensitive data table, and the processing time is less than or equal to the pre-set time, the first emergency response plan is identified as medium-risk.

[0062] Furthermore, after determining the security level, a suitable transmission method is selected based on the security level to send the first emergency response plan to the relevant departments so that personnel in those departments can take corresponding actions in a timely manner. Specifically, this includes: sending the first emergency response plan to the first department based on the security level; when the first emergency response plan corresponds to a high-risk level, performing a first encryption on the first emergency response plan according to the high-risk level to obtain a second emergency response plan; the first encryption includes quantum encryption; obtaining the second department, which is the department involved in the first emergency response plan; identifying multiple preset departments identical to the second department from the target area, with each preset department corresponding to a location; obtaining the target permission level corresponding to the third department, which is any one of the multiple preset departments; determining whether the target permission level matches the preset permission level, which is the permission level corresponding to the high-risk level; when the target permission level matches the preset permission level, determining to allocate the second emergency response plan to the third department, and outputting the third department as the first department. Specifically, by evaluating the first emergency response plan, the security level corresponding to the first emergency response plan is determined; when the first emergency response plan corresponds to a high-risk level, the encryption process is initiated. For example, a chemical plant experiences a leak. If the estimated leak volume exceeds 5 tons and there are residential areas within a 5-kilometer radius, it is classified as a high-risk area. The first emergency response plan can be encrypted using a quantum key distribution system to generate a random key, leveraging the properties of quantum states (such as photon polarization) to ensure the key cannot be eavesdropped. The first emergency response plan is then converted into a digital format and encrypted using a symmetric encryption algorithm (such as AES) combined with quantum key distribution. Ciphertext and a corresponding quantum key identifier (used for decryption matching) are generated. This generated ciphertext can then be used as the second emergency response plan. The key is securely transmitted beforehand via a quantum channel to the relevant departments mentioned in the emergency response plan. The list of responsible departments in the first emergency response plan (such as the fire department, environmental protection bureau, and medical team) is parsed. This list is stored as a "second department" list for subsequent access control. For example, the first emergency response plan might specify that the fire department is responsible for firefighting, the environmental protection bureau for pollution monitoring, and the medical team for treating the injured. A geographical area can be defined based on the accident location (e.g., all departments within a 10-kilometer radius). The system iterates through the database of all departments within the target area, filtering out departments whose names or functions match those in the second confidentiality list. Each department is then pre-assigned a permission level (e.g., Fire Department - High, Environmental Protection Bureau - Medium). The permission level is then determined based on the high risk associated with a pre-defined safety incident; here, the permission level refers to the qualifications or scale required for the department to handle the incident. Next, a third department is selected one by one from the pre-defined department list. The permission level of this department is then queried and compared with the pre-defined permission levels (high risk corresponds to high permission).The system compares the access level of the third department with the preset access level. If the third department's access level matches the preset level, it is determined that the third department can handle the incident, and the second emergency response plan can be allocated to the third department. The encrypted second emergency response plan is transmitted to the qualified department, achieving secure encryption, precise allocation, and access control for high-risk emergency plans, ensuring the efficiency and security of the emergency response. If the access level of the third department does not match the preset access level, other departments must be selected from the preset departments, and their access levels compared with the preset access levels to find a qualified department to handle the incident. After receiving the second emergency response plan, the third department decrypts it using a key to obtain the decrypted first emergency response plan, and then executes the corresponding handling measures according to the decrypted first emergency response plan. The entire allocation process is recorded in real time for easy auditing and traceability.

[0063] Furthermore, when the first emergency response plan corresponds to a medium-risk level, a second encryption is performed on the first emergency response plan based on the medium-risk level to obtain a third emergency response plan. The second encryption only encrypts the core words in the first emergency response plan. The plan consists of core words and non-core words, with core words being those appearing in a preset sensitive information table. The second location corresponding to the predicted safety accident is obtained; the second location is the specific location of the predicted safety accident in the target area. The distances from multiple preset departments to the second location are calculated sequentially, resulting in multiple distance values. These distance values ​​are then sorted from smallest to largest to obtain the target ranking result. A fourth department is obtained; the fourth department is the first distance value in the target ranking result, and the third emergency response plan is sent to the fourth department. Specifically, if the risk level of the first emergency response plan is determined to be medium-risk, the second encryption process is triggered. First, words appearing in the preset sensitive information table are extracted from the first emergency response plan. For example, if the preset sensitive information table includes "leak," "fire," and "explosion," and the first emergency response plan includes "a gas leak has occurred; immediately evacuate personnel and close the valve," then the core word is "leak." The second encryption process only encrypts the core words, leaving the non-core words unchanged. The system adapts to keyword matching algorithms to identify core words, then replaces and encrypts these core words (e.g., by replacing them with hash values) to obtain a third emergency response plan. Next, it obtains the second location corresponding to the predicted safety accident. If the first predicted accident is a chemical leak, then the chemical plant in the target area needs to be monitored. The specific location of the chemical plant in the target area, i.e., the second location, can then be obtained. The accident location can be acquired in real time via IoT sensors. Then, multiple pre-defined departments are determined from the first emergency response plan. These pre-defined departments refer to relevant departments that need to participate in the emergency response plan, including fire departments, hospitals, environmental protection bureaus, safety supervision bureaus, and traffic management departments. For example, if the first emergency response plan mentions a fire department, all fire departments in the target area are acquired, resulting in multiple pre-defined departments. The locations of each pre-defined department are then obtained sequentially, and the distances from each pre-defined department to the second location are calculated sequentially. Straight-line distances can be calculated using latitude and longitude (e.g., the Haversine formula), or map APIs (e.g., X-Map, X-Degree Map) can be used to obtain actual road distances. For each pre-defined department, its distance to the second location is calculated, resulting in multiple distance values. Sort multiple distance values ​​in ascending order to obtain the target sorting result. Use a sorting algorithm (such as quicksort or mergesort) to sort the distance values. Extract the department at the top of the target sorting result (i.e., the closest department), designate it as the fourth department, and automatically push the second emergency response plan through the emergency command system. Send the third emergency response plan to the head of the fourth department.If the emergency response plan mentions multiple departments, the relevant department closest to the pre-set safety incident can be selected sequentially according to the above content, and the encrypted third emergency response plan can then be sent to these departments. After receiving the third emergency response plan, the fourth department decrypts the encrypted data involved in the third emergency response plan to obtain the decrypted emergency response plan, and then implements the corresponding handling measures according to the decrypted emergency response plan.

[0064] When the first emergency response plan corresponds to a low-risk level, such as when the first emergency response plan is for abnormal weather information and requires warning people in a certain area to pay attention to safety when traveling, and this first emergency response plan does not involve sensitive information, it can be directly sent to the meteorological department and traffic management department so that their personnel can carry out subsequent operations according to the plan. In addition to the aforementioned safety incidents, it can also provide users with real-time traffic congestion information so that users can adjust their travel time or routes in a timely manner. It can also combine historical sites or buildings in the target area to provide travel arrangements for users or other tourists. Furthermore, it can conduct daily monitoring of safety incidents in various factories or industrial parks, providing early warnings to reduce the probability of equipment failure and ensure worker safety.

[0065] By employing the above method, the first prediction target is determined based on the type of target facility in the target area. Based on this first prediction target, multiple acquisition devices are used to automatically acquire monitoring data, covering all relevant information. The acquired monitoring data and the first location are then input into a multimodal large-scale model for extraction and correlation analysis. No manual intervention is required to obtain the analysis results. Based on these results, predicted safety incidents are determined, and corresponding emergency response plans are quickly generated. The predicted safety incidents and their handling time are automatically classified into safety levels. Emergency resources are dynamically allocated based on these safety levels, and relevant departments are notified in real time via push notifications. Using a multimodal large-scale model for prediction can replace expert experience, solving the problem of inefficiency caused by traditional methods that rely on domain experts manually labeling abnormal data and analyzing multiple data sources one by one.

[0066] This application also provides a smart city processing system based on a multimodal large model. Figure 2 This is a schematic diagram of the structure of a smart city processing system based on a multimodal large model provided in an embodiment of this application. (Refer to...) Figure 2 The system includes an acquisition unit 201, a processing unit 202, and a sending unit 203.

[0067] Acquisition unit 201 acquires target facilities from the target area. The target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities, and industrial building facilities. The target area is any area corresponding to a city.

[0068] Processing unit 202 determines a first prediction target based on the target facility. The first prediction target includes traffic accidents, medical accidents, natural disasters, construction accidents, public safety accidents, and park safety accidents. Based on the first prediction target, it determines multiple data acquisition devices and acquires multiple monitoring data collected by the multiple data acquisition devices, and acquires the first location corresponding to the target area. It inputs the multiple monitoring data and the first location into a multimodal large model for processing to obtain analysis results. Based on the analysis results, it determines the predicted safety accident and determines a first emergency response plan based on the predicted safety accident.

[0069] The sending unit 203 determines the safety level based on the processing time and the predicted safety incident, and sends the first emergency response plan to the first department based on the safety level so that the corresponding personnel in the first department can view the first emergency response plan. The processing time is the total time corresponding to processing the first emergency response plan.

[0070] In one possible implementation, the acquisition unit 201 is used to acquire the target factors currently corresponding to the target area, including time factors and weather factors; the processing unit 202 is used to determine the second prediction object based on the target factors and target facilities; determine whether the second prediction object exists in the first prediction object; when the second prediction object exists in the first prediction object, monitor the target area according to the second prediction object first.

[0071] In one possible implementation, the processing unit 202 is used to determine the data type based on the first prediction object, including electricity data type, traffic data type, personnel location data type, environmental data type, and construction data type; input the data type into a preset frequency database for matching to obtain the target acquisition frequency; determine the acquisition device based on the data type, and send the target acquisition frequency to each acquisition device so that the acquisition device can acquire data according to the target acquisition frequency.

[0072] In one possible implementation, the acquisition unit 201 is used to extract core data from the first emergency response plan; the processing unit 202 is used to determine a preset sensitive data table based on the predicted safety accident, determine whether core data exists in the preset sensitive data table, and determine whether the processing time is less than or equal to a preset time; when core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, the first emergency response plan is determined to be at a high-risk level, and the high-risk level is output as a safety level; when core data exists in the preset sensitive data table and the processing time is greater than the preset time, the first emergency response plan is determined to be at a medium-risk level, and the medium-risk level is output as a safety level.

[0073] In one possible implementation, the processing unit 202 is used to perform a first encryption on the first emergency response plan according to the high-risk level when the first emergency response plan corresponds to a high-risk level, to obtain a second emergency response plan, wherein the first encryption includes quantum encryption; the acquisition unit 201 is used to acquire a second department, which is the department involved in the first emergency response plan; the processing unit 202 is used to screen multiple preset departments that are the same as the second department from the target area, with each preset department corresponding to a location; the acquisition unit 201 is used to acquire the target permission level corresponding to a third department, which is any one of the multiple preset departments; the processing unit 202 is used to determine whether the target permission level is consistent with the preset permission level, wherein the preset permission level is the permission level corresponding to the high-risk level; the sending unit is used to determine to allocate the second emergency response plan to the third department when the target permission level is consistent with the preset permission level, and output the third department as the first department.

[0074] In one possible implementation, the processing unit 202 is used to perform a second encryption on the first emergency response plan when the first emergency response plan corresponds to a medium-risk level, based on the medium-risk level, to obtain a third emergency response plan. The second encryption is to encrypt only the core words in the first emergency response plan. The processing plan consists of core words and non-core words, and the core words are words that appear in a preset sensitive information table. The acquisition unit 201 is used to acquire the second location corresponding to the predicted safety accident, which is the specific location of the predicted safety accident in the target area. The distances between multiple preset departments and the second location are calculated sequentially to obtain multiple distance values. The multiple distance values ​​are sorted in ascending order to obtain a target sorting result. The sending unit 203 is used to acquire the fourth department, which is the distance value ranked first in the target sorting result, and send the third emergency response plan to the fourth department.

[0075] In one possible implementation, the acquisition unit 201 is used to acquire target activities in the target area within a preset time period, including concert activities, sports events, and exhibition activities; the processing unit 202 is used to combine the target activities with the target facilities to obtain target combination information; perform simulation analysis on the target combination information to obtain a third prediction object, and prioritize monitoring the target area according to the third prediction object.

[0076] It should be noted that the system provided in the above embodiments is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0077] This application also discloses an electronic device. (See reference...) Figure 3 , Figure 3 This application provides a schematic diagram of the structure of an electronic device. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 302, and at least one communication bus 305.

[0078] The communication bus 305 is used to enable communication between these components.

[0079] The user interface 303 may include a display screen and a camera. Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0080] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0081] The processor 301 may include one or more processing cores. The processor 301 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 302, and by calling data stored in memory 302. Optionally, the processor 301 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 301 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and application requests; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 301 and may be implemented as a separate chip.

[0082] The memory 302 may include random access memory (RAM) or read-only memory. Optionally, the memory 302 may include a non-transitory computer-readable storage medium. The memory 302 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 302 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch functionality, sound playback functionality, image playback functionality, etc.), instructions for implementing the various method embodiments described above, etc. The data storage area may store data involved in the various method embodiments described above. Optionally, the memory 302 may also be at least one storage device located remotely from the aforementioned processor 301.

[0083] like Figure 3 As shown, the memory 302, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for smart city processing based on a multimodal large model.

[0084] exist Figure 3 In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 301 can be used to call the application program for smart city processing based on multimodal large model stored in the memory 302. When executed by one or more processors, the electronic device performs one or more of the methods described in the above embodiments.

[0085] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0086] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0087] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some service interfaces; indirect couplings or communication connections between devices or units may be electrical or other forms.

[0088] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0089] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0090] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, 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. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0091] The above description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and the disclosure of practical truths. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure.

Claims

1. A smart city processing method based on a multimodal large model, characterized in that, The method includes: Obtain target facilities from the target area, wherein the target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities and industrial building facilities, and the target area is an area corresponding to any city; Based on the target facility, a first prediction object is determined, which includes traffic accidents, medical accidents, natural disaster accidents, construction accidents, public safety accidents, and park safety accidents. Based on the first prediction object, multiple acquisition devices are determined, and multiple monitoring data collected by the multiple acquisition devices are obtained, and the first location corresponding to the target area is obtained; The multiple monitoring data and the first location are input into a multimodal large model for processing to obtain analysis results; Based on the analysis results, a predicted safety accident is determined, and a first emergency response plan is determined based on the predicted safety accident. The safety level is determined based on the processing time and the predicted safety incident. Based on the safety level, the first emergency response plan is sent to the first department so that the personnel in the first department can view the first emergency response plan. The processing time is the total time required to process the first emergency response plan. Specifically, determining the safety level based on the processing time and the predicted safety incident includes: extracting core data from the first emergency response plan; determining a preset sensitive data table based on the predicted safety incident, determining whether the core data exists in the preset sensitive data table, and determining whether the processing time is less than or equal to a preset time; when the core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, the first emergency response plan is determined to be at a high-risk level, and the high-risk level is output as the safety level; when the core data exists in the preset sensitive data table and the processing time is greater than the preset time, the first emergency response plan is determined to be at a medium-risk level, and the... The medium-risk level is output as the security level; the step of sending the first emergency response plan to the first department based on the security level specifically includes: when the first emergency response plan corresponds to the high-risk level, performing a first encryption on the first emergency response plan according to the high-risk level to obtain a second emergency response plan, the first encryption including quantum encryption; obtaining a second department, the second department being the department involved in the first emergency response plan; screening multiple preset departments identical to the second department from the target area, one preset department corresponding to one location; obtaining the target permission level corresponding to a third department, the third department being any one of the multiple preset departments; determining whether the target permission level is consistent with the preset permission level, the preset permission level being the permission level corresponding to the high-risk level; when the target permission level is consistent with the preset permission level, determining to allocate the second emergency response plan to the third department, and outputting the third department as the first department.

2. The method according to claim 1, characterized in that, After determining the first prediction target based on the target facility, the method further includes: Obtain the target factors currently corresponding to the target area, including time factors and weather factors; A second prediction target is determined based on the target factors and the target facilities; Determine whether the second prediction object exists in the first prediction object; When the second prediction object exists among the first prediction objects, the target area is monitored preferentially according to the second prediction object.

3. The method according to claim 1, characterized in that, Before acquiring multiple monitoring data collected by the multiple acquisition devices and obtaining the first location corresponding to the target area, the method further includes: The data type is determined based on the first prediction object, and the data type includes electricity data type, traffic data type, personnel location data type, environmental data type, and construction data type; The data type is input into a preset frequency database for matching to obtain the target acquisition frequency; The acquisition device is determined according to the data type, and the target acquisition frequency is sent to each acquisition device so that the acquisition device can acquire data according to the target acquisition frequency.

4. The method according to claim 1, characterized in that, The step of sending the first emergency response plan to the first department based on the security level specifically includes: When the first emergency response plan corresponds to the medium risk level, the first emergency response plan is encrypted a second time according to the medium risk level to obtain a third emergency response plan. The second encryption is to encrypt only the core words in the first emergency response plan. The response plan is composed of the core words and non-core words. The core words are the words that appear in the preset sensitive information table. Obtain the second location corresponding to the predicted safety accident, where the second location is the specific location of the predicted safety accident in the target area; The distances between the preset departments and the second location are calculated sequentially to obtain multiple distance values; The multiple distance values ​​are sorted in ascending order to obtain the target sorting result; Obtain the fourth department, which is the distance value that ranks first in the target sorting result, and send the third emergency response plan to the fourth department.

5. The method according to claim 1, characterized in that, After determining the first prediction target based on the target facility, the method further includes: Acquire target activities in the target area within a preset time period, including concerts, sporting events, and exhibitions; The target activity is combined with the target facility to obtain target combination information; The target combination information is simulated and analyzed to obtain a third prediction object, and the target area is monitored preferentially according to the third prediction object.

6. A smart city processing system based on a multimodal large model, characterized in that, The system includes an acquisition unit (201), a processing unit (202), and a sending unit (203). The acquisition unit (201) acquires target facilities from the target area. The target facilities include any one or more of public service facilities, commercial facilities, residential facilities, productive building facilities, and industrial building facilities. The target area is the area corresponding to any city. The processing unit (202) determines a first prediction object based on the target facility, the first prediction object including traffic accidents, medical accidents, natural disasters, construction accidents, public safety accidents, and park safety accidents; determines multiple acquisition devices based on the first prediction object, acquires multiple monitoring data collected by the multiple acquisition devices, and acquires a first location corresponding to the target area; inputs the multiple monitoring data and the first location into a multimodal large model for processing to obtain analysis results; determines predicted safety accidents based on the analysis results, and determines a first emergency response plan based on the predicted safety accidents; The sending unit (203) determines the security level based on the processing time and the predicted security incident, and sends the first emergency response plan to the first department based on the security level, so that the personnel corresponding to the first department can view the first emergency response plan. The processing time is the total time for processing the first emergency response plan. The step of determining the safety level based on processing time and the predicted safety incident specifically includes: extracting core data from the first emergency response plan; determining a preset sensitive data table based on the predicted safety incident, determining whether the core data exists in the preset sensitive data table, and determining whether the processing time is less than or equal to a preset time; when the core data exists in the preset sensitive data table and the processing time is less than or equal to the preset time, determining that the first emergency response plan is at a high-risk level, and outputting the high-risk level as the safety level; when the core data exists in the preset sensitive data table and the processing time is greater than the preset time, determining that the first emergency response plan corresponds to a medium-risk level, and outputting the medium-risk level as the safety level; and sending the first emergency response plan to the first department based on the safety level, specifically including... The process includes: when the first emergency response plan corresponds to the high-risk level, performing a first encryption on the first emergency response plan according to the high-risk level to obtain a second emergency response plan, wherein the first encryption includes quantum encryption; obtaining a second department, wherein the second department is the department involved in the first emergency response plan; screening multiple preset departments that are the same as the second department from the target area, wherein each preset department corresponds to a location; obtaining a target permission level corresponding to a third department, wherein the third department is any one of the multiple preset departments; determining whether the target permission level is consistent with a preset permission level, wherein the preset permission level is the permission level corresponding to the high-risk level; when the target permission level is consistent with the preset permission level, determining to allocate the second emergency response plan to the third department, and outputting the third department as the first department.

7. An electronic device, characterized in that, The device includes a processor (301), a memory (302), a user interface (303), and a network interface (304). The memory (302) is used to store instructions. The user interface (303) and the network interface (304) are used to communicate with other devices. The processor (301) is used to execute the instructions stored in the memory (302) to cause the electronic device (300) to perform the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, perform the method as described in any one of claims 1-5.

Images