Disaster emergency disposal method and system based on large event perception

Abstract

The invention discloses a disaster emergency disposal method and system based on large event perception, and relates to the technical field of disaster emergency disposal in the environmental sanitation industry, and the method comprises the steps: collecting multi-source heterogeneous data according to a set collection period; obtaining multi-source heterogeneous data corresponding to a historical large disaster event, extracting disaster abnormal features, and training a large disaster event sensing model based on the disaster abnormal features; constructing a large-scale disaster event knowledge base, and obtaining large-scale disaster event early warning information according to the collected multi-source heterogeneous data, the large-scale disaster event sensing model and the large-scale disaster event knowledge base; matching a corresponding emergency disposal plan from a pre-constructed emergency disposal plan library according to the large-scale disaster event early warning information, re-planning a sanitation vehicle route, and obtaining a corresponding emergency disposal scheme according to the re-planned sanitation vehicle route and the corresponding emergency disposal plan; and monitoring the corresponding emergency disposal scheme in real time and generating feedback information to update the corresponding emergency disposal scheme.

Description

Technical Field

[0001] This invention relates to the field of disaster emergency response technology in the sanitation industry, specifically to a disaster emergency response method and system based on large-scale event perception. Background Technology

[0002] With the acceleration of urbanization and the increasing complexity of urban spatial structures, urban meteorological disasters, public safety accidents, and environmental pollution incidents are characterized by diversification, strong regional coupling, and rapid evolution. As the underlying support system for urban operations, the urban sanitation system not only undertakes basic tasks such as daily road cleaning, garbage collection, and facility maintenance, but is also a key force in maintaining basic environmental safety and public health protection in the face of natural disasters, public emergencies, and extreme weather conditions. In scenarios such as rainstorms and floods, typhoons, blizzards and cold waves, chemical leaks, building collapses, major traffic accidents, and the surge in environmental load caused by large-scale festivals, the sanitation system bears the critical responsibility of "controlling pollution, preventing its spread, and ensuring traffic flow" in the first instance.

[0003] Existing technologies suffer from the following problems: a lack of real-time and comprehensive data acquisition and sensing capabilities, making it difficult to obtain timely information on the dynamic changes of disaster events and to accurately predict and assess their impact; existing emergency plans and resource allocation often rely on human experience, lacking intelligent sensing and decision support, resulting in slow responses to complex and sudden disasters; and inadequate feedback mechanisms, lacking real-time monitoring and dynamic optimization of the effectiveness of emergency response plans, hindering continuous optimization and improvement, and reducing the efficiency and effectiveness of emergency response. Therefore, there is an urgent need for disaster emergency response methods and systems based on large-scale event sensing to address the problems existing in current technologies. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a disaster emergency response method based on large-scale event perception, which includes the following steps: Step 1: Pre-build a multi-source data acquisition network platform, and collect multi-source heterogeneous data of the sanitation operation area in real time according to the set acquisition cycle through deployed multi-source sensors; Step 2: Obtain multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Step 3: Construct a large-scale disaster event knowledge base. Based on the multi-source heterogeneous data collected in real time from sanitation operation areas, the large-scale disaster event perception model, and the large-scale disaster event knowledge base, obtain early warning information for large-scale disaster events. Step 4: Based on the early warning information of large-scale disaster events, match the corresponding emergency response plan from the pre-built emergency response plan library, replan the sanitation vehicle routes, and obtain the corresponding emergency response plan based on the replanned sanitation vehicle routes and the corresponding emergency response plan. Step 5: Monitor the implementation effect of the corresponding emergency response plan and changes in the on-site situation in real time, generate feedback information, and update the corresponding emergency response plan based on the feedback information.

[0005] In a preferred embodiment, the multi-source heterogeneous data specifically includes: meteorological data, road traffic conditions, sanitation vehicle data, environmental monitoring data, garbage recycling station status data, energy consumption monitoring data, and sanitation material inventory status.

[0006] In a preferred embodiment, the multi-source data acquisition network platform specifically includes: urban sanitation IoT sensors, meteorological monitoring systems, video surveillance networks, and social media information.

[0007] In a preferred embodiment, the disaster characteristics specifically include: abnormal meteorological characteristics, abnormal traffic characteristics, abnormal environmental pollution characteristics, abnormal energy consumption characteristics, abnormal waste recycling station status characteristics, and abnormal operating materials characteristics.

[0008] In a preferred embodiment, the specific process of training a large-scale disaster event perception model based on disaster anomaly characteristics to obtain a pre-trained large-scale disaster event perception model includes: Acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, perform data preprocessing, and extract disaster anomaly features; A large-scale disaster event perception model is constructed based on machine learning algorithms. The disaster anomaly features are input into the large-scale disaster event perception model for training. The large-scale disaster event perception model takes the corresponding large-scale disaster event as its output. The major disaster events specifically include: freezing rain events, flooding events caused by rainstorms, blizzard events, high temperature events, large-scale events, and public health events; The loss function is used as the optimization objective, and a preset model perception accuracy threshold is set. When the accuracy of the large disaster event predicted by the large disaster event perception model is greater than or equal to the preset model perception accuracy threshold, the model training is stopped. A pre-trained model for perceiving large-scale disaster events was developed.

[0009] In a preferred embodiment, the specific process of obtaining early warning information for large-scale disaster events includes: Standardize and preprocess the multi-source heterogeneous data collected in real time from the sanitation operation area to extract the disaster anomaly characteristics at the current moment; The extracted disaster anomaly features are input into the large-scale disaster event perception model, and the large-scale disaster event perception model outputs the corresponding large-scale disaster event. Based on the corresponding large-scale disaster events and the large-scale disaster event knowledge base, early warning information for large-scale disaster events is obtained; The early warning information for major disaster events specifically includes: event type, scope of impact, severity, and development trend.

[0010] In a preferred embodiment, the specific process of replanning sanitation vehicle routes includes: The current location, vehicle type, load status, remaining fuel consumption, and vehicle operation status of all sanitation vehicles are obtained in real time based on the data of sanitation vehicles. Based on real-time road traffic information, road sections that are impassable, restricted, and obstructed are marked, and the toll cost coefficient for each road section is calculated. Based on early warning information of major disaster events, the locations and areas of emergency operation tasks are obtained, and priority weights and completion time windows are assigned to each emergency operation task location and area; Based on the current vehicle status data, the toll cost coefficient of each road segment, and the priority weights and completion time windows assigned to each emergency operation task location area, a path optimization algorithm is used to achieve the optimal matching between vehicles and each emergency operation task location area. The optimal matching principle specifically includes: the adaptability of vehicle type to task requirements, the distance between the vehicle's current location and the task point, the vehicle's remaining fuel consumption and task execution capability, and multi-dimensional matching constraints between task priority and vehicle response speed.

[0011] In a preferred embodiment, the specific process of real-time monitoring of the implementation effect of the corresponding emergency response plan and changes in the on-site situation, generating feedback information, and updating the corresponding emergency response plan based on the feedback information includes: Establish a full-process monitoring and feedback mechanism for emergency response. During the execution of the emergency response plan, multi-dimensional data is collected in real time through preset monitoring terminals, and the completion status of each task point and the progress of regional response are visualized on the dispatch and monitoring platform to generate real-time execution results. The system compares the real-time execution results with the emergency response plan execution results. When time deviation, route deviation, progress delay, or abnormal resource consumption exceeding the preset threshold is detected, feedback information is generated. Based on the feedback information, the plan update mechanism will be activated to update the corresponding emergency response plan.

[0012] Another aspect of the present invention discloses a disaster emergency response system based on large-scale event perception, which includes the following modules: a data acquisition module, a model training module, an early warning information generation module, an emergency plan generation module, and a feedback and update module; Data acquisition module; pre-built multi-source data acquisition network platform, and real-time acquisition of multi-source heterogeneous data of sanitation operation area through deployed multi-source sensors according to the set acquisition cycle; Model training module: acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Early warning information generation module; construct a large-scale disaster event knowledge base, and obtain large-scale disaster event early warning information based on real-time collected multi-source heterogeneous data of sanitation operation areas, large-scale disaster event perception model and large-scale disaster event knowledge base; Emergency response plan generation module: Based on the early warning information of large-scale disaster events, it matches the corresponding emergency response plan from the pre-built emergency response plan library, re-plans the sanitation vehicle routes, and obtains the corresponding emergency response plan based on the re-planned sanitation vehicle routes and the corresponding emergency response plan. Feedback and update module: Real-time monitoring of the implementation effect of the corresponding emergency response plan and changes in the on-site situation, generating feedback information, and updating the corresponding emergency response plan based on the feedback information.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention pre-constructs a multi-source data acquisition network platform and uses deployed multi-source sensors to collect multi-source heterogeneous data of sanitation operation areas in real time according to a set acquisition cycle; it acquires multi-source heterogeneous data corresponding to historical large-scale disaster events and extracts disaster anomaly features, trains a large-scale disaster event perception model based on the disaster anomaly features, and obtains a pre-trained large-scale disaster event perception model; it constructs a large-scale disaster event knowledge base, and obtains large-scale disaster event early warning information based on the real-time collected multi-source heterogeneous data of sanitation operation areas, the large-scale disaster event perception model, and the large-scale disaster event knowledge base; it matches the corresponding emergency response plan from the pre-constructed emergency response plan library based on the large-scale disaster event early warning information, and re-plans sanitation vehicle routes; based on the re-planned sanitation vehicle routes and the corresponding emergency response plan, it obtains the corresponding emergency response scheme; it monitors the execution effect and on-site situation changes of the corresponding emergency response scheme in real time, generates feedback information, and updates the corresponding emergency response scheme based on the feedback information. Attached Figure Description

[0014] Figure 1 This is a schematic flowchart illustrating the steps of a disaster emergency response method based on large-scale event perception, according to an embodiment of this application.

[0015] Figure 2 This is a schematic diagram showing the connection of various modules in the disaster emergency response system based on large-scale event awareness, as described in an embodiment of this application. Detailed Implementation

[0016] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0017] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly. Example

[0018] Please see Figure 1 As shown, this application provides a disaster emergency response method based on large-scale event awareness, mainly applied to the sanitation industry. The specific method includes the following steps: Step 1: Pre-build a multi-source data acquisition network platform, and collect multi-source heterogeneous data of the sanitation operation area in real time according to the set acquisition cycle through deployed multi-source sensors; Based on the above embodiments, the multi-source data acquisition network platform specifically includes: urban sanitation IoT sensors, meteorological monitoring systems, video surveillance networks, and social media information; Based on the above embodiments, the multi-source heterogeneous data specifically includes: meteorological data, road traffic conditions, sanitation vehicle data, environmental monitoring data, garbage recycling station status data, energy consumption monitoring data, and sanitation material inventory status, etc. Specifically, meteorological data includes: real-time monitoring values ​​of temperature, air pressure, precipitation, wind force, visibility, etc., as well as forecast data for the next week; road traffic conditions include: traffic flow data: hourly and minute traffic flow on roads, including vehicle throughput, speed, lane occupancy, etc.; real-time road condition information: reflecting the traffic status of major roads, including whether there is traffic congestion, accidents, road construction, traffic control, and obstacles, etc.; sanitation vehicle data includes: GPS positioning data: real-time monitoring of the location information of sanitation vehicles, including... Latitude, vehicle speed, and route; load data, vehicle operating status (vehicle start / stop status), fuel consumption data, etc.; environmental monitoring data specifically includes: concentration data of pollutants such as PM2.5, PM10, sulfur dioxide, nitrogen oxides, carbon monoxide, and ozone; waste recycling station status data specifically includes: the filling level of waste bins, the capacity of waste recycling stations, and remaining storage capacity; energy consumption monitoring data specifically includes: electricity consumption data and water resource consumption data; sanitation operation material inventory status specifically includes: material inventory data and material consumption data, etc. Step 2: Obtain multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Based on the above embodiments, the disaster characteristics specifically include: abnormal meteorological characteristics, abnormal traffic characteristics, abnormal environmental pollution characteristics, abnormal energy consumption characteristics, abnormal garbage recycling station status characteristics, and abnormal operating materials characteristics; Specifically, meteorological anomalies may include: abnormal precipitation intensity and wind force caused by extreme weather events such as rainstorms and typhoons; which may easily trigger disasters such as floods, landslides, and house collapses; by extracting abnormal precipitation intensity and wind force as meteorological anomalies, we can reflect the characteristics of disaster occurrence, essential laws, and development trends, and provide a foundation for subsequent training of large-scale disaster event perception models. For example, abnormal energy consumption characteristics may include: a burst water pipe or a broken circuit causing significant anomalies in water and electricity usage data. By extracting abnormal water and electricity characteristics, we can reflect the occurrence of water and electricity disasters. Sanitation workers can arrive at the scene to clean up and protect the environment based on the perceived water and electricity disasters, and at the same time, we can provide a basis for training large-scale disaster event perception models. Based on the above embodiments, the specific process of training a large-scale disaster event perception model based on disaster anomaly characteristics to obtain a pre-trained large-scale disaster event perception model includes: Acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, perform data preprocessing, and extract disaster anomaly features; A large-scale disaster event perception model is constructed based on machine learning algorithms. The disaster anomaly features are input into the large-scale disaster event perception model for training. The large-scale disaster event perception model takes the corresponding large-scale disaster event as its output. The major disaster events specifically include: freezing rain events, flooding events caused by rainstorms, blizzard events, high temperature events, large-scale events, public health events, etc. The loss function is used as the optimization objective, and a preset model perception accuracy threshold is set. When the accuracy of the large disaster event predicted by the large disaster event perception model is greater than or equal to the preset model perception accuracy threshold, the model training is stopped. The model was pre-trained to detect large-scale disaster events. Specifically, large-scale events include events such as large-scale population movements during holidays, concerts, and exhibitions. During holidays, the surge in population movement and tourism activities may lead to a surge in urban waste. Sanitation workers need to predict densely populated areas in advance and increase manpower and resources to cope with the centralized treatment of waste. Large-scale events such as concerts and exhibitions usually generate a lot of waste. Sanitation departments need to make arrangements for waste sorting, cleaning and transportation in advance, especially during peak periods after the event, to ensure environmental cleanliness and timely waste removal. Specifically, a loss function is used as the optimization objective, which includes: mean squared error, mean absolute error, cross-entropy loss, etc.; a large-scale disaster event perception model is constructed based on machine learning algorithms, which include: decision tree, random forest, neural network, etc. Step 3: Construct a large-scale disaster event knowledge base. Based on the multi-source heterogeneous data collected in real time from sanitation operation areas, the large-scale disaster event perception model, and the large-scale disaster event knowledge base, obtain early warning information for large-scale disaster events. Based on the above embodiments, the specific process of obtaining early warning information for large-scale disaster events includes: Standardize and preprocess the multi-source heterogeneous data collected in real time from the sanitation operation area to extract the disaster anomaly characteristics at the current moment; The extracted disaster anomaly features are input into the large-scale disaster event perception model, and the large-scale disaster event perception model outputs the corresponding large-scale disaster event. Based on the corresponding large-scale disaster events and the large-scale disaster event knowledge base, early warning information for large-scale disaster events is obtained; The early warning information for major disaster events specifically includes: event type, scope of impact, severity, and development trend; Specifically, the large-scale disaster event knowledge base includes: cleaning, preprocessing, and format standardization of multi-source heterogeneous data collected from sanitation operation areas to ensure data accuracy and consistency; structuring the core concepts, attributes, and interrelationships of large-scale disaster events to form a large-scale disaster event ontology and storing it in a knowledge graph or vector database; using natural language processing (NLP) technology to extract entities and relationships from the knowledge graph or vector database; providing disaster perception and decision support through an inference engine and intelligent query system; and being able to update in real time and automatically expand and optimize data based on newly occurring disaster events to support subsequent intelligent retrieval, inference, and decision support. Step 4: Based on the early warning information of large-scale disaster events, match the corresponding emergency response plan from the pre-built emergency response plan library, replan the sanitation vehicle routes, and obtain the corresponding emergency response plan based on the replanned sanitation vehicle routes and the corresponding emergency response plan. Based on the above embodiments, the pre-constructed emergency response plan library specifically includes: personnel estimation model plan, temporary facility addition plan, enhanced control measures and rehearsal plan for densely populated areas, drainage and dredging plan, garbage anti-floating measures plan, disinfection and epidemic prevention and rehearsal plan, snow and ice removal plan, etc.; each plan includes: resource requirement list, response process, personnel allocation, equipment deployment, personnel task division, operation process time nodes, material consumption prediction and equipment scheduling path, etc. The specific process of replanning sanitation vehicle routes includes: The current location, vehicle type (sweeper, water truck, garbage truck, snowplow, drainage truck, etc.), load status, remaining fuel consumption, and vehicle operation status (working / idle / returning) of all sanitation vehicles are obtained in real time based on the sanitation vehicle data. Based on real-time road traffic information, road sections that are impassable, restricted, and obstructed are marked, and the toll cost coefficient for each road section is calculated. Specifically, the toll cost coefficients for each road segment (including time cost, safety risk, congestion index, etc.); Based on early warning information of major disaster events, the locations and areas of emergency operation tasks are obtained, and priority weights and completion time windows are assigned to each emergency operation task location and area; Based on the current vehicle status data, the toll cost coefficient of each road segment, and the priority weights and completion time windows assigned to each emergency operation task location area, a path optimization algorithm is used to achieve the optimal matching between vehicles and each emergency operation task location area. The matching principles specifically include: the adaptability of vehicle type to task requirements, the distance between the vehicle's current location and the task point, the vehicle's remaining fuel consumption and task execution capability, and multi-dimensional matching constraints between task priority and vehicle response speed. Specifically, path optimization algorithms include one or more combinations of A* algorithm, Hungarian algorithm, genetic algorithm, or heuristic algorithm. Based on the replanned sanitation vehicle routes and the corresponding emergency response plans, corresponding emergency response schemes are obtained. The corresponding emergency response schemes specifically include: the types of large-scale disaster events, as well as the corresponding vehicle route planning schemes, personnel allocation schemes, material allocation schemes, execution schedules, work procedures, and alternative schemes. Step 5: Monitor the implementation effect of the corresponding emergency response plan and changes in the on-site situation in real time, generate feedback information, and update the corresponding emergency response plan based on the feedback information; Based on the above embodiments, the specific process of real-time monitoring of the implementation effect of the corresponding emergency response plan and changes in the on-site situation, generating feedback information, and updating the corresponding emergency response plan based on the feedback information includes: Establish a full-process monitoring and feedback mechanism for emergency response. During the execution of the emergency response plan, multi-dimensional data is collected in real time through preset monitoring terminals, and the completion status of each task point and the progress of regional response are visualized on the dispatch and monitoring platform to generate real-time execution results. The system compares the real-time execution results with the emergency response plan execution results. When time deviation, route deviation, progress delay, or abnormal resource consumption exceeding the preset threshold is detected, feedback information is generated. Based on the feedback information, the plan update mechanism will be activated to update the corresponding emergency response plan; Specifically, the pre-set monitoring terminals include: vehicle-mounted GPS system, vehicle-mounted sensors, personnel positioning system, Internet of Things sensors, material management system, etc. The multi-dimensional data specifically includes: the location coordinates, speed, trajectory, and arrival time at the task point of sanitation vehicles; changes in vehicle load, fuel consumption, and operating status of equipment (such as the working status of sprinklers, sweepers, and snowplows); location information, on-duty status, and working hours of sanitation workers; video footage of the work site; environmental change data of the work area (such as changes in water depth, road icing conditions, and garbage collection progress); real-time consumption and remaining inventory of materials; ensuring the comprehensiveness and real-time nature of the monitoring data. Specifically, the feedback information may include, for example, if the feedback indicates that the garbage collection progress is lagging behind, initiating a plan update mechanism by adjusting vehicle scheduling and increasing the number of workers; to ensure that the garbage collection task is completed on time and to avoid affecting subsequent sanitation work or the sanitation of the public environment. Example

[0019] Please see Figure 2As shown, in another embodiment of the present invention, the present invention also discloses a disaster emergency response system based on large-scale event perception, the system comprising the following modules: a data acquisition module, a model training module, an early warning information generation module, an emergency plan generation module, and a feedback and update module; The modules described above are connected via wired and / or wireless means to enable data transmission between them. Data acquisition module; pre-built multi-source data acquisition network platform, and real-time acquisition of multi-source heterogeneous data of sanitation operation area through deployed multi-source sensors according to the set acquisition cycle; Model training module: acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Early warning information generation module; construct a large-scale disaster event knowledge base, and obtain large-scale disaster event early warning information based on real-time collected multi-source heterogeneous data of sanitation operation areas, large-scale disaster event perception model and large-scale disaster event knowledge base; Emergency response plan generation module: Based on the early warning information of large-scale disaster events, it matches the corresponding emergency response plan from the pre-built emergency response plan library, re-plans the sanitation vehicle routes, and obtains the corresponding emergency response plan based on the re-planned sanitation vehicle routes and the corresponding emergency response plan. Feedback and update module: Real-time monitoring of the implementation effect of the corresponding emergency response plan and changes in the on-site situation, generating feedback information, and updating the corresponding emergency response plan based on the feedback information.

[0020] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0021] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A disaster emergency response method based on large-scale event perception, characterized in that, The method includes the following steps: Step 1: Pre-build a multi-source data acquisition network platform, and collect multi-source heterogeneous data of the sanitation operation area in real time according to the set acquisition cycle through deployed multi-source sensors; Step 2: Obtain multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Step 3: Construct a large-scale disaster event knowledge base. Based on the multi-source heterogeneous data collected in real time from sanitation operation areas, the large-scale disaster event perception model, and the large-scale disaster event knowledge base, obtain early warning information for large-scale disaster events. Step 4: Based on the early warning information of large-scale disaster events, match the corresponding emergency response plan from the pre-built emergency response plan library, replan the sanitation vehicle routes, and obtain the corresponding emergency response plan based on the replanned sanitation vehicle routes and the corresponding emergency response plan. Step 5: Monitor the implementation effect of the corresponding emergency response plan and changes in the on-site situation in real time, generate feedback information, and update the corresponding emergency response plan based on the feedback information.

2. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The multi-source heterogeneous data specifically includes: meteorological data, road traffic conditions, sanitation vehicle data, environmental monitoring data, garbage recycling station status data, energy consumption monitoring data, and sanitation material inventory status.

3. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The multi-source data acquisition network platform specifically includes: urban sanitation IoT sensors, meteorological monitoring systems, video surveillance networks, and social media information.

4. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The specific characteristics of the disaster include: abnormal meteorological characteristics, abnormal traffic characteristics, abnormal environmental pollution characteristics, abnormal energy consumption characteristics, abnormal garbage recycling station status characteristics, and abnormal operating materials characteristics.

5. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The specific process of training a large-scale disaster event perception model based on disaster anomaly features to obtain a pre-trained large-scale disaster event perception model includes: Acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, perform data preprocessing, and extract disaster anomaly features; A large-scale disaster event perception model is constructed based on machine learning algorithms. The disaster anomaly features are input into the large-scale disaster event perception model for training. The large-scale disaster event perception model takes the corresponding large-scale disaster event as its output. The major disaster events specifically include: freezing rain events, flooding events caused by rainstorms, blizzard events, high temperature events, large-scale events, and public health events; The loss function is used as the optimization objective, and a preset model perception accuracy threshold is set. When the accuracy of the large disaster event predicted by the large disaster event perception model is greater than or equal to the preset model perception accuracy threshold, the model training is stopped. A pre-trained model for perceiving large-scale disaster events was developed.

6. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The specific process of obtaining early warning information for major disaster events includes: Standardize and preprocess the multi-source heterogeneous data collected in real time from the sanitation operation area to extract the disaster anomaly characteristics at the current moment; The extracted disaster anomaly features are input into the large-scale disaster event perception model, and the large-scale disaster event perception model outputs the corresponding large-scale disaster event. Based on the corresponding large-scale disaster events and the large-scale disaster event knowledge base, early warning information for large-scale disaster events is obtained; The early warning information for major disaster events specifically includes: event type, scope of impact, severity, and development trend.

7. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The specific process of replanning sanitation vehicle routes includes: The current location, vehicle type, load status, remaining fuel consumption, and vehicle operation status of all sanitation vehicles are obtained in real time based on the data of sanitation vehicles. Based on real-time road traffic information, road sections that are impassable, restricted, and obstructed are marked, and the toll cost coefficient for each road section is calculated. Based on early warning information of major disaster events, the locations and areas of emergency operation tasks are obtained, and priority weights and completion time windows are assigned to each emergency operation task location and area; Based on the current vehicle status data, the toll cost coefficient of each road segment, and the priority weights and completion time windows assigned to each emergency operation task location area, a path optimization algorithm is used to achieve the optimal matching between vehicles and each emergency operation task location area. The optimal matching principle specifically includes: the adaptability of vehicle type to task requirements, the distance between the vehicle's current location and the task point, the vehicle's remaining fuel consumption and task execution capability, and multi-dimensional matching constraints between task priority and vehicle response speed.

8. The disaster emergency response method based on large-scale event perception according to claim 1, characterized in that, The specific process of real-time monitoring of the implementation effectiveness of the corresponding emergency response plan and changes in the on-site situation, generating feedback information, and updating the corresponding emergency response plan based on the feedback information includes: Establish a full-process monitoring and feedback mechanism for emergency response. During the execution of the emergency response plan, multi-dimensional data is collected in real time through preset monitoring terminals, and the completion status of each task point and the progress of regional response are visualized on the dispatch and monitoring platform to generate real-time execution results. The system compares the real-time execution results with the emergency response plan execution results. When time deviation, route deviation, progress delay, or abnormal resource consumption exceeding the preset threshold is detected, feedback information is generated. Based on the feedback information, the plan update mechanism will be activated to update the corresponding emergency response plan.

9. A disaster emergency response system based on large-scale event perception, employing the disaster emergency response method based on large-scale event perception as described in any one of claims 1-8, characterized in that, The system includes the following modules: data acquisition module, model training module, early warning information generation module, emergency response plan generation module, and feedback and update module; Data acquisition module; A multi-source data acquisition network platform is pre-built, and multi-source heterogeneous data of the sanitation operation area is collected in real time according to the set acquisition cycle through deployed multi-source sensors. Model training module: acquire multi-source heterogeneous data corresponding to historical large-scale disaster events, extract disaster anomaly features, train a large-scale disaster event perception model based on disaster anomaly features, and obtain a pre-trained large-scale disaster event perception model; Early warning information generation module; A knowledge base for large-scale disaster events is constructed. Based on multi-source heterogeneous data collected in real time from sanitation operation areas, a large-scale disaster event perception model, and the knowledge base for large-scale disaster events, early warning information for large-scale disaster events is obtained. Emergency response plan generation module; Based on the early warning information of major disaster events, the corresponding emergency response plan is matched from the pre-built emergency response plan database, and the sanitation vehicle routes are replanned. Based on the replanned sanitation vehicle routes and the corresponding emergency response plan, the corresponding emergency response plan is obtained. Feedback and update module; The system monitors the implementation effectiveness of corresponding emergency response plans and changes in the on-site situation in real time, generates feedback information, and updates the corresponding emergency response plans based on the feedback information.

Images