Urban emergency event intelligent response method and device, electronic equipment and storage medium

Abstract

The invention discloses a city emergency event intelligent response method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring multi-modal data of a current city emergency event; performing semantic modeling and event classification processing on the multi-modal data; deducing the current city emergency event according to a pre-constructed knowledge graph and an event classification processing result, and evaluating the risk level of the current city emergency event; and generating a task chain according to a risk level assessment result, and scheduling urban emergency resources according to the task chain. By adopting the scheme, semantic modeling and event classification are carried out by fusing the multi-modal data, and the current city emergency event is deduced according to the pre-constructed knowledge graph, risk assessment is carried out and the task chain is generated, so that the current city emergency event is intelligently responded; automatic sensing, research and judgment, deduction and response scheduling of urban emergency events are realized, and the intelligent level of urban emergency management is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent emergency management technology, and in particular to an intelligent response method, device, electronic equipment and storage medium for urban emergency events. Background Technology

[0002] With the acceleration of urbanization, the frequency and severity of urban emergencies (such as fires, earthquakes, chemical leaks, and large-scale traffic accidents) are increasing. Therefore, the development of efficient and intelligent urban fire emergency plans has become an urgent need.

[0003] Urban emergency response systems generally suffer from pain points such as information fragmentation, cumbersome decision-making processes, and slow resource allocation and response. Existing emergency management platforms mostly adopt centralized display of multi-source data, provide standard preset response plans, and allow manual triggering of tasks or contact with contingency groups.

[0004] Most existing emergency management platforms are information integration products, but current urban management faces the following problems in responding to emergencies: information lag, with event discovery heavily reliant on manual reporting; data isolation and inconsistent responses between various emergency systems (119, 120, transportation, urban management, etc.); inability to form the optimal response path based on data in the first instance; and lack of drills and automatic reasoning mechanisms, making it difficult to adapt to complex and new types of disasters. Summary of the Invention

[0005] This invention provides an intelligent response method, device, electronic device, and storage medium for urban emergency events, in order to solve the problems of high labor costs, inconsistent emergency resource response, and lack of simulation of urban emergency events in the prior art.

[0006] According to one aspect of the present invention, an intelligent response method for urban emergency events is provided, the method comprising:

[0007] Acquire multimodal data on current urban emergency events;

[0008] The multimodal data is subjected to semantic modeling and event classification processing; wherein, the event classification processing includes urban emergency event type classification processing, urban emergency event scenario classification processing, and urban emergency event severity classification processing.

[0009] Based on the pre-constructed knowledge graph and the results of event classification and processing, the current urban emergency events are simulated, and the risk level of the current urban emergency events is assessed.

[0010] A task chain is generated based on the risk level assessment results, and urban emergency resources are dispatched based on the task chain.

[0011] According to another aspect of the present invention, an intelligent response device for urban emergency events is provided, the device comprising:

[0012] The multimodal data acquisition module is used to acquire multimodal data of current urban emergency events;

[0013] The data processing module is used to perform semantic modeling and event classification processing on the multimodal data; wherein, the event classification processing includes urban emergency event type classification processing, urban emergency event scenario classification processing, and urban emergency event severity classification processing.

[0014] The event simulation and risk assessment module is used to simulate current urban emergency events based on a pre-built knowledge graph and the results of event classification and processing, and to assess the risk level of current urban emergency events.

[0015] The intelligent response module is used to generate a task chain based on the risk level assessment results, and to dispatch urban emergency resources according to the task chain.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] At least one processor; and

[0018] A memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to execute the intelligent response method for urban emergency events according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the intelligent response method for urban emergency events according to any embodiment of the present invention.

[0021] According to another aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements the intelligent response method for urban emergency events as described in any embodiment of the present invention.

[0022] The technical solution of this invention acquires multimodal data of current urban emergency events; performs semantic modeling and event classification processing on the multimodal data; based on a pre-constructed knowledge graph and the event classification results, it simulates the current urban emergency events and assesses their risk levels; generates a task chain based on the risk level assessment results, and schedules urban emergency resources according to the task chain. This solves the problems of high labor costs, inconsistent emergency resource responses, and lack of simulation of urban emergency events in existing technologies; and achieves the beneficial effect of realizing automatic perception, analysis, simulation, and response scheduling of urban emergency events, thereby improving the intelligence level of urban emergency management.

[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of an intelligent response method for urban emergency events provided in Embodiment 1 of the present invention;

[0026] Figure 2 This is a flowchart of an intelligent response method for urban emergency events provided in Embodiment 2 of the present invention;

[0027] Figure 3 This is a schematic diagram of the structure of an intelligent response device for urban emergency events provided in Embodiment 3 of the present invention;

[0028] Figure 4 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] The acquisition, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. It should be noted that the terms "first," "second," "target," and "original," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising," "etc.," and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] Example 1

[0032] Figure 1 This is a flowchart of an intelligent response method for urban emergency events provided in Embodiment 1 of the present invention. This embodiment is applicable to situations where multimodal data is integrated for semantic modeling and event classification, and the current urban emergency event is simulated based on a pre-constructed knowledge graph to perform risk assessment and generate a task chain for intelligent response. This method can be executed by an intelligent urban emergency response device, which can be implemented in hardware and / or software and can be configured in any electronic device with network communication capabilities. Figure 1 As shown, the method includes:

[0033] S110. Obtain multimodal data on current urban emergency events.

[0034] Urban emergency events can refer to incidents such as fires, earthquakes, chemical spills, and large-scale traffic accidents. The multimodal data refers to data on urban emergency events acquired from at least two dimensions. This multimodal data includes, but is not limited to, video surveillance anomaly identification data, audio monitoring data, social media scraping data, sensor access data, and weather and geographic access data. Video surveillance anomaly identification data refers to abnormal data on urban emergency events obtained from video surveillance. Based on video surveillance, preliminary detection of current urban emergency events can be performed, such as detecting flames, dense smoke, crowds, traffic congestion, abnormally high temperatures, and smoke-filled areas. Audio monitoring data can determine the presence of explosions or cries for help. Relevant keywords are extracted from social media scraping data. Smoke data, temperature and humidity data, and toxic gas data are obtained from sensor access data. High temperature and strong wind warnings, and earthquake data are obtained from weather and geographic access data.

[0035] S120. Perform semantic modeling and event classification processing on the multimodal data.

[0036] Semantic modeling refers to capturing the deep meaning in data or language through formal methods. In this embodiment of the invention, semantic modeling unifies multimodal data into semantic vectors or event descriptions. For example, semantic modeling of sound monitoring data can identify explosions or cries for help. Semantic modeling of video surveillance anomaly identification data can identify flames, dense smoke, crowds, traffic congestion, abnormally high temperatures, and smoke areas. Semantic modeling of social media scraped data can identify relevant keywords for urban emergency events. Semantic modeling of sensor access data can identify smoke data, temperature and humidity data, and toxic gas data. Semantic modeling of weather and geographic access data can identify high temperature and strong wind warnings, as well as earthquake data.

[0037] Event classification processing refers to classifying current urban emergency events based on multimodal data after semantic modeling. This event classification processing includes classification processing based on urban emergency event type, urban emergency event scenario, and urban emergency event severity.

[0038] S130. Based on the pre-constructed knowledge graph and the results of event classification and processing, the current urban emergency events are simulated, and the risk level of the current urban emergency events is assessed.

[0039] The pre-built knowledge graph can refer to a knowledge graph constructed based on historical urban emergency event databases, expert rule sets, urban static attribute information, IoT semantic mapping, and legal and policy knowledge. Nodes in the knowledge graph represent event types, resources, locations, infrastructure, and participating entities. For example, in a fire event, corresponding resources include fire trucks, ambulances, and emergency supplies; infrastructure includes hospitals and fire stations; and participating entities include traffic police, fire brigades, and volunteers.

[0040] Simulating current urban emergency events can refer to projecting current urban emergency events to determine their potential impacts. For example, if a fire breaks out in building C3, based on a pre-built knowledge graph and event classification results, it can be deduced that a fire may occur in building C4, causing dense smoke that reduces visibility on surrounding roads, leading to traffic congestion and affecting the passage of fire trucks and ambulances.

[0041] Risk level assessment refers to determining the impact of a current urban emergency based on simulations, thereby assessing the risk level of the emergency. For example, the risk level of a current urban emergency can be determined based on factors such as the speed of the event's spread, resource mobilization rate, population exposure rate, and transmission potential.

[0042] S140. Generate a task chain based on the risk level assessment results, and dispatch urban emergency resources based on the task chain.

[0043] In this context, a task chain can refer to the measures that need to be taken based on the risk level assessment results, and a series of these measures can be organized into a task chain. For example, if the risk level assessment result of the current urban emergency is determined to be high risk, requiring an immediate upgrade to the highest response, the corresponding task chain would be: notify the police - control roads - dispatch firefighters - activate the transfer hospital - evacuate via broadcast system.

[0044] After generating the task chain, urban emergency resources are dispatched according to the task chain. These urban emergency resources include, but are not limited to, available fire-fighting resources, ambulance resources, and traffic control resources. For example, based on the generated task chain: notifying the police - controlling roads - dispatching fire-fighting teams - activating transfer hospitals - evacuation via broadcast system, the corresponding urban emergency resources to be dispatched include fire-fighting resources, ambulance resources, traffic control resources, and information notification resources.

[0045] This invention provides an intelligent response method for urban emergency events. The method involves acquiring multimodal data of current urban emergency events; performing semantic modeling and event classification on the multimodal data; using a pre-constructed knowledge graph and the event classification results to simulate the current urban emergency event and assess its risk level; generating a task chain based on the risk level assessment results; and scheduling urban emergency resources according to the task chain. By employing the technical solution of this invention, which integrates multimodal data for semantic modeling and event classification, and uses a pre-constructed knowledge graph to simulate the current urban emergency event, conduct risk assessments, and generate task chains, an intelligent response method for current urban emergency events is achieved. This enables automatic perception, analysis, simulation, and response scheduling of urban emergency events, thereby improving the level of intelligent urban emergency management.

[0046] Example 2

[0047] Figure 2 This is a flowchart of an intelligent response method for urban emergency events provided in Embodiment 2 of the present invention. The present invention further optimizes the aforementioned embodiments based on the above embodiments, and can be combined with various optional solutions from one or more of the above embodiments. For example... Figure 2 As shown, the method includes:

[0048] S210. Obtain multimodal data on current urban emergency events.

[0049] This involves acquiring multimodal data on current urban emergency events from multiple dimensions. For example, it includes acquiring anomaly identification data from video surveillance, sound monitoring data from audio, social media scraping data from various social media apps, sensor access data from various sensors, and weather and geographic access data from weather forecasting stations.

[0050] The video surveillance anomaly identification data includes, but is not limited to, flame data, smoke data, crowd gathering data, abnormal high temperature point data, vehicle congestion data, and smoke area data; the sound monitoring data includes, but is not limited to, explosions and cries for help; the media capture data includes, but is not limited to, keywords of urban emergency events and public opinion heat; the sensor access data includes, but is not limited to, smoke data, temperature and humidity data, and toxic gas data; the weather and geography access data includes, but is not limited to, high temperature data, strong wind warning data, and earthquake data.

[0051] As an optional but non-limiting implementation, the acquisition of video surveillance anomaly identification data for current urban emergency events includes, but is not limited to, steps A1-A2:

[0052] Step A1: Use an object detection algorithm to detect objects in the video frame and determine the detection results; wherein, the detection results include the location information of the object detection box, the category label of the detected object, the confidence score of the detected object, the timestamp of the current video frame, and the video acquisition point information;

[0053] Step A2: Aggregate the detection results by time window and use a time series model or trajectory rule-based method to make a preliminary judgment on urban emergency events in the video frames in order to identify video surveillance anomaly identification data.

[0054] In this context, video surveillance anomaly identification data can refer to the data obtained after performing target detection and anomaly identification on video frames in video surveillance. For example, in this embodiment of the invention, the YOLO model (e.g., YOLOv5 / YOLOv8) is used to perform target detection and anomaly identification on video frames.

[0055] The YOLO model is used to perform object detection on video frames, and the detection results are determined. The detection results include bounding box (bbox), class, confidence, timestamp, and camera_id. Here, bbox refers to the location information of the object detection bounding box, indicating the position of the detected object in the video frame. class refers to the category label of the detected object identified by the model, indicating the type of the detected object; for example, flame, smoke, and people. Confidence refers to the model's confidence (probability) in the recognition result, indicating the degree of certainty that the detected object belongs to the predicted category; the range is usually [0,1], with higher confidence indicating more reliable results. Timestamp is the timestamp of the current video frame, used to synchronize the video frame with subsequent multimodal data. Camera_id refers to the video capture point information, a unique identifier of the video data source, used to distinguish the capture points of different cameras in the city. It is usually assigned by the city's video surveillance system.

[0056] The detection results are aggregated by time window, and a preliminary judgment is made on urban emergency events in the current video frame using a temporal model or trajectory-based rules. The temporal model includes, but is not limited to, a one-dimensional convolutional neural network + bidirectional long short-term memory network, a temporal displacement model, and a three-dimensional convolutional neural network; the trajectory rules include, but are not limited to, the rate of increase in smoke / personnel / flame area, changes in smoke density, and personnel movement patterns; the preliminary judgment of urban emergency events in the current video frame includes, but is not limited to, initial fire outbreak, fire spread, mass riots, and traffic congestion.

[0057] In one optional embodiment of the present invention, the acquired multimodal data is used to train the system to automatically make a preliminary judgment on urban emergency events in video frames. Specifically, multimodal data (flames, smoke, firelight, etc.) is collected and labeled, and the location information of the target detection boxes is labeled using COCO format. The training set can be augmented using synthetic data (overlay of flame / smoke images). YOLO is first used for transfer learning, and then the temporal model is classified using labeled short videos. Training metrics include detection accuracy (mAP), event detection (F1 / recall), and latency (ms / frame). YOLO is exported to ONNX / TensorRT and inference is performed on an edge GPU (or NVIDIA Jetson); the detection results are sent to a central platform for temporal fusion via Kafka / MQTT. Multi-camera fusion + Multi-Object Tracking (such as DeepSORT) is used to obtain object trajectories to assist in determining the spread direction and speed, in order to make a preliminary judgment on urban emergency events in video frames.

[0058] S220. Perform semantic modeling and event classification processing on the multimodal data.

[0059] The system takes multimodal data as input and transforms it into semantic vectors / event descriptions through unimodal coding, cross-modal fusion, and temporal classification, thereby achieving accurate classification of event type, scenario, and severity.

[0060] As an optional but non-limiting implementation, the semantic modeling and event classification processing of the multimodal data includes, but is not limited to, steps B1-B3:

[0061] Step B1: Encode the multimodal data to obtain encoded multimodal data;

[0062] Step B2: Use a cross-modal fusion model to fuse the encoded multimodal data into a unified vector, and output the confidence vector and semantic labels;

[0063] Step B3: Input several continuous unified vectors into the time series network to determine the event category and severity of the current urban emergency events.

[0064] Specifically, for data from different modalities, independent encoders are designed to transform the raw input into structured embeddings, laying the foundation for cross-modal fusion. For video surveillance anomaly detection data, input video frames or static images are encoded to obtain a visual embedding (typically 512 / 768 dimensions), containing visual semantic information such as color, shape, and object layout. For audio monitoring data, the input raw audio waveform is encoded to output an audio embedding, including encoded sound frequency, intensity, and temporal dynamics. For social media scraping data, unstructured text input, such as social media posts, alarm records, and news descriptions, is encoded to output a text embedding, containing semantic information such as event-related entities, sentiment, and causal relationships. For sensor-accessed data, structured numerical values, such as smoke concentration, temperature, gas concentration, air pressure, and humidity, are input to output a numerical embedding, encoding the dynamic trends of environmental parameters. For weather and geographic access data, the input is geographic location coordinates and timestamp, and the output is spatio-temporal embedding to supplement the contextual background information of the event.

[0065] A cross-modal fusion engine integrates the embeddings of various modalities into a unified semantic vector while preserving the interpretability of modal contributions. Multi-head attention allows direct interaction between different modal embeddings, learning semantic associations between modalities, such as the association between "flame image in video" and "alarm sound in audio." It also preserves temporal or spatial dependencies within a single modality, such as the temporal continuity of a video frame sequence. The output includes a unified semantic vector, a confidence vector and semantic labels, and attention weights. The unified semantic vector refers to the multimodal fusion features at time step t, containing cross-modal semantic information. The confidence vector and semantic labels are preliminary event description phrases generated through the output layer, such as "suspected fire incident." The attention weights record the contribution of each modality to the fusion result, such as "visual modality weight accounts for 60%" and "IoT smoke concentration weight accounts for 30%" in a fire incident, to support model interpretation.

[0066] Urban emergency events typically exhibit temporal continuity, such as the evolution of a fire from "smoke" to "explosion." Therefore, time-series models are needed to capture dynamic features and output the final event classification result. A unified vector of continuous time steps is input into a time-series network to classify current urban emergency events and determine their severity. This time-series network includes, but is not limited to, Bi-LSTM, Temporal Transformer, and TCN. Bi-LSTM (Bidirectional Long Short-Term Memory) captures past and future dependencies of events and is suitable for small to medium-scale time-series data. Temporal Transformer incorporates temporal location encoding to handle long-sequence dependencies, such as the correlation between 24-hour weather data and events. TCN (Temporal Convolutional Network) captures multi-scale temporal features through convolutional kernel expansion and is suitable for high-frequency sampled data, such as real-time streams from IoT sensors. Event classification includes event type classification and event scene classification. Event type classification includes, but is not limited to, fires, earthquakes, traffic accidents, and extreme weather. Event scene classification assists in determining the event type, such as indoor / outdoor scenes and urban / rural scenes. Event severity is used to quantify the severity of current urban emergency events. It can be combined with numerical characteristics (such as smoke concentration, number of casualties mentioned in the text) and the degree of visual damage for comprehensive judgment.

[0067] In this embodiment of the invention, multimodal data fusion is used to avoid unimodal ambiguity; cross-modal attention weights are used to visualize the contribution of each modality to event judgment; and the evolution process of events is captured by a temporal network to improve the classification accuracy of complex events.

[0068] S230. Based on the pre-constructed knowledge graph and the results of event classification and processing, the current urban emergency events are simulated, and the risk level of the current urban emergency events is assessed.

[0069] Before simulating current urban emergency events and assessing their risk levels based on the pre-built knowledge graph and event classification results, it is necessary to first construct the knowledge graph.

[0070] As an optional but non-limiting implementation, before extrapolating the current urban emergency situation based on the pre-built knowledge graph and the event classification processing results, and before assessing the risk level of the current urban emergency situation, the method further includes constructing a knowledge graph, specifically including but not limited to steps C1-C2:

[0071] Step C1: Based on the historical urban emergency event database, expert rule set, urban static attribute information, IoT semantic mapping, and legal and policy knowledge, define core node types, semantic edge relationships, and dynamic attributes; wherein, the core node types include event types, resource types, location types, sensors, infrastructure, rules, and participating entities; semantic edge relationships include causal relationships, spatial relationships, resource dependencies, impact relationships, and temporal attributes; dynamic attributes include weights and constraints.

[0072] Step C2: Construct a knowledge graph based on core node types, semantic edge relationships, and dynamic attributes.

[0073] This involves constructing a knowledge graph based on a historical urban emergency event database, expert rule sets, urban static attribute information, IoT semantic mapping, and legal and policy knowledge, defining core node types, semantic edge relationships, and dynamic attributes. The expert rule sets include, but are not limited to, SOPs (Standard Operating Procedures) and procedures from fire departments, traffic police, and emergency management departments; the historical urban emergency event database includes, but is not limited to, local / national accident logs, including, but not limited to, event types, evolution paths, response measures, and results; urban static attribute information includes, but is not limited to, building structures, road topology, population density, and key facilities (gas stations, chemical plants); IoT semantic mapping refers to the mapping from sensors to entities (e.g., which sensor belongs to which building); and legal and policy knowledge includes, but is not limited to, evacuation rules and road closure priorities.

[0074] The structure of a knowledge graph includes nodes and edges. Node types include, but are not limited to, event types, resource types, location types, sensors, infrastructure, rules, and participating entities; the semantic relationships of edges include, but are not limited to, causal relationships, spatial relationships, resource dependencies, influence relationships, and temporal attributes; dynamic attributes include, but are not limited to, node and edge weights, historical probabilities, delay distributions, and conditional constraints, such as wind speeds exceeding a preset wind speed threshold.

[0075] As an optional but non-limiting implementation, the method involves extrapolating current urban emergency events based on a pre-constructed knowledge graph and event classification processing results, and assessing the risk level of current urban emergency events, including but not limited to steps D1-D3:

[0076] Step D1: Based on the event classification processing results, convert the multimodal data into semantic triples, and perform static attribute matching on the current urban emergency events based on the pre-constructed knowledge graph; wherein, the static attributes include building structure, road topology, population density, and key facilities;

[0077] Step D2: Perform path search on the pre-built knowledge graph to determine at least two candidate paths; wherein the candidate paths consist of spatial sprawl paths, secondary event paths, and resource impact paths;

[0078] Step D3: Determine the evolution probability and expected evolution time of the at least two candidate paths, determine the evolution curves of the at least two candidate paths based on the evolution probability and expected evolution time, and construct an event development trend diagram;

[0079] Step D4: Based on the event development trend map, assess the risk level of the current urban emergency event from at least two dimensions; wherein, the at least two dimensions include event status value, resource coverage, population exposure, and spread trend.

[0080] The acquired multimodal data is converted into semantic triples, and static attribute matching is performed on the current urban emergency events based on a pre-constructed knowledge graph. For example, the semantic triples include "Event: Fire in Building C3, Location: XX Street, Intensity: Level 3, Weather: Wind speed 8m / s, Temperature 25℃", and the static attribute matching includes matching static attributes such as "Building C3 - Fire resistance rating B" and "Gas station within 50 meters".

[0081] Starting from the event node, the semantic edges in the knowledge graph are traversed using breadth-first search (BFS) or a limited depth search (e.g., a maximum depth of 5 layers to avoid path explosion) to search for potential impact paths. Examples include: spatial spread paths such as "fire in building C3 - adjacent to building C4 - causing fire in building C4"; secondary event paths such as "fire - generating dense smoke - affecting visibility of surrounding roads - causing road congestion - affecting the passage of rescue vehicles"; and resource impact paths such as "road congestion - traffic police intervention - affecting traffic control resources in other areas".

[0082] For each candidate path, the evolution probability (P) and expected evolution time (T) are calculated by combining dynamic attributes from the knowledge graph. The evolution probability is calculated by integrating historical statistical probabilities (e.g., "the historical probability of a similar fire spreading to an adjacent building is 0.3"), rule constraints (e.g., "wind speed of 8 m / s increases the spread probability to 0.4"), and entity attributes (e.g., building C4 is made of "flammable materials," increasing the probability by 0.1). P is obtained through a Bayesian network or weighted summation. The expected evolution time is calculated by accumulating the path time based on the edge attribute "time_delay," such as "the fire spreads to building C4 10 minutes after it occurs (T=10 min), and road congestion is triggered 5 minutes later (T=15 min)." The multi-path evolution curve is output in the form of "time axis (T) + probability (P)" to construct an event development trend graph, and key nodes with high probability (e.g., P>0.5) and short time (e.g., T<30 min) are marked, such as "high risk of road congestion after 15 minutes, traffic police should be dispatched first."

[0083] Based on the aforementioned event development trend map, the risk level of the current urban emergency event is assessed from at least two dimensions. These at least two dimensions include, but are not limited to, event status value, resource coverage, population exposure, and propagation trend. Specifically, the event status value (E) refers to the escalation probability or fire spread rate derived from the knowledge graph, ranging from [0,1]; resource coverage (R) refers to the number of resources available for mobilization within the maximum permissible response time / the ideal number of resources, ranging from [0,1], with higher values ​​indicating lower risk; population exposure (Q) refers to the standardized population density within the affected area, ranging from [0,1]; and propagation potential (S) refers to the rate of public opinion dissemination and the potential for secondary incidents, ranging from [0,1]. The corresponding risk value can be expressed as: ,in, , , as well as Indicates weight, It can be obtained through historical data, expert opinions, or Bayesian optimization learning.

[0084] Risk levels can be divided into four levels, for example, they can be represented as:

[0085] Level I: This indicates a high risk and requires an immediate upgrade to a high-risk response.

[0086] Level II: Medium to high risk, mobilize large amounts of resources as soon as possible;

[0087] Level III: Medium risk; monitor and prepare for response.

[0088] Level IV: Low risk, standard procedure.

[0089] This invention, through structured modeling and rule-based reasoning using knowledge graphs, enables the transformation of emergency response from passive to proactive prediction. It can be widely applied to scenarios such as urban fire spread prediction, traffic accident secondary disaster early warning, and hazardous chemical leakage and diffusion simulation, providing emergency command with a three-in-one decision-making basis of "time-probability-resources", significantly improving the accuracy and timeliness of emergency response.

[0090] S240. Generate a task chain based on the risk level assessment results, and dispatch urban emergency resources based on the task chain.

[0091] Among them, after conducting a risk level assessment of the current urban emergency, the resources that need to be dispatched are determined based on the risk level assessment results, and a task chain is generated to dispatch urban emergency resources.

[0092] As an optional but non-limiting implementation, the generation of a task chain based on the risk level assessment results, and the dispatching of urban emergency resources based on the task chain, includes, but is not limited to, steps E1-E2:

[0093] Step E1: Map urban emergency resources based on the location information and risk level assessment results of the current urban emergency events, and generate task chains; wherein, the task chains include event reporting, traffic control, fire dispatch, rescue, and evacuation; urban emergency resources include available fire resources, rescue resources, and traffic control resources;

[0094] Step E2: Based on the task chain and the city emergency resource mapping results, dispatch city emergency resources and automatically generate response record reports.

[0095] The urban emergency resource mapping refers to mapping specific urban emergency resources, such as fire-fighting resources, ambulance resources, and traffic control resources, based on the location information and risk level assessment results of the current urban emergency event; and estimating the estimated arrival time of the urban emergency resources. Optionally, if emergency resources are lacking in the current area, emergency resources in adjacent areas can be dispatched or air support can be requested. A task chain is generated based on the mapped urban emergency resources and the estimated arrival time.

[0096] Optionally, the task chain is initialized as a series of tasks, each containing a duration, earliest start time, latest start time, and required resources. Optimization is performed using a constraint solver, employing integer linear programming (MILP) or approximate algorithms (such as heuristics, greedy algorithms, or genetic algorithms) to find the optimal execution sequence, thereby minimizing the objective function to determine the final task chain (e.g., maximizing the protected population, minimizing response time, or reducing losses). The objective function can be designed as a multi-objective, which can be handled through parallel solving or weighted composition methods. For example, concurrent execution of traffic police road closures (5 minutes) and fire department dispatch (8 minutes) can effectively shorten the total time.

[0097] The dispatching of urban emergency resources includes, but is not limited to, traffic control, information dissemination, and public opinion monitoring. Traffic control refers to real-time planning of evacuation routes (avoiding congestion points) and linkage with intelligent traffic lights (mandating green light waves); information dissemination refers to tiered early warning pushes (government / media / public) and multi-channel coverage (SMS / radio / APP pop-ups); public opinion monitoring refers to real-time crawling of social media platforms, identifying rumors, and automatically debunking warnings.

[0098] In this embodiment of the invention, for critical actions such as road closures or forced evacuations, resource scheduling can be performed through automatic and manual dual-signature methods, or automatically authorized based on preset scenarios. After resource scheduling, all scheduling requests are logged, and response record reports are automatically generated for post-event traceability.

[0099] In one optional embodiment of the invention, the method further includes simulating a virtual city scene and training an intelligent response based on historical urban emergency data to deduce the optimal resource scheduling strategy under different events. For example, the "physical / social" system is modeled as a digital twin / simulator, and the fire propagation model, urban road network / traffic model, and crowd evacuation model are coupled to simulate the execution and effects of each strategy. A large number of candidate resource scheduling strategy combinations are run in the simulator, and the optimal resource scheduling strategy combination is selected through an evaluation function. Specifically, this includes constructing a simulation engine, a strategy space, an evaluation function, and an optimization method. The simulation engine includes a fire model, traffic simulation, crowd behavior, and a coupler; the fire model can use a probability propagation model (based on the flame spread rate) or incorporate a more accurate physical model; the traffic simulation uses microscopic traffic simulation to simulate the impact of road vehicle flow and traffic control; crowd behavior is determined based on an intelligent agent model to determine movement speed, evacuation behavior, and congestion response; the coupler is used to schedule the simulation time step and synchronize fire changes and traffic conditions at each step, for example, a fire blocking a main road will affect the estimated arrival time of rescue vehicles. The strategy space refers to the combination of resource scheduling strategies, which include, but are not limited to, road closure / opening, prioritizing firefighting / prioritizing medical aid, designating evacuation routes, setting up temporary medical points, and diversion measures. The combination space is large, requiring search / optimization methods. The evaluation function can refer to the evaluation function that assigns weights to each emergency resource to determine the combination of resource scheduling strategies. Optimization methods include sample search, black-box optimization, reinforcement learning, and heuristic / rule-guided search. For example, the current scenario is identified, an initial simulation state is constructed, the simulator is invoked to generate at least two candidate resource scheduling strategies, and parallel simulation is performed to calculate the result of each strategy, thereby determining the target strategy. The target strategy is then distributed for resource scheduling.

[0100] This invention, based on multimodal data fusion semantic modeling, constructs a knowledge graph and infers events, generates response task chains based on large model reasoning, and coordinates heterogeneous systems such as transportation, government affairs, public security, and media for scheduling, thereby realizing the automatic perception, analysis, inference, and response scheduling of urban emergency events and improving the level of intelligent urban emergency management.

[0101] Example 3

[0102] Figure 3 This is a schematic diagram of the structure of an intelligent response device for urban emergency events provided in Embodiment 3 of the present invention. Figure 3 As shown, the device includes:

[0103] The multimodal data acquisition module 310 is used to acquire multimodal data of current urban emergency events;

[0104] The data processing module 320 is used to perform semantic modeling and event classification processing on the multimodal data; wherein, the event classification processing includes urban emergency event type classification processing, urban emergency event scenario classification processing, and urban emergency event severity classification processing.

[0105] The event simulation and risk assessment module 330 is used to simulate current urban emergency events based on a pre-built knowledge graph and event classification and processing results, and to assess the risk level of current urban emergency events.

[0106] The intelligent response module 340 is used to generate a task chain based on the risk level assessment results and to dispatch urban emergency resources based on the task chain.

[0107] Optionally, the multimodal data includes: video surveillance anomaly identification data, audio monitoring data, social media capture data, sensor access data, and weather and geographic access data.

[0108] Optional, a multimodal data acquisition module, specifically used for:

[0109] A target detection algorithm is used to detect targets in video frames and the detection results are determined. The detection results include the location information of the target detection box, the category label of the detected object, the confidence score of the detected object, the timestamp of the current video frame, and the video acquisition point information.

[0110] The detection results are aggregated according to time windows, and a time series model or trajectory-based rule is used to make a preliminary judgment on urban emergency events in video frames in order to identify video surveillance anomaly identification data.

[0111] Optional, data processing module, specifically used for:

[0112] The multimodal data is encoded to obtain the encoded multimodal data;

[0113] A cross-modal fusion model is used to fuse the encoded multimodal data into a unified vector, and output the confidence vector and semantic labels.

[0114] Several continuous unified vectors are input into a time series network to classify current urban emergency events and determine their severity.

[0115] Optional, the event simulation and risk assessment module is specifically used for:

[0116] Based on the event classification results, multimodal data is converted into semantic triples, and static attribute matching of current urban emergency events is performed based on a pre-constructed knowledge graph; wherein, the static attributes include building structure, road topology, population density, and key facilities;

[0117] A path search is performed on a pre-built knowledge graph to identify at least two candidate paths; wherein the candidate paths consist of spatial sprawl paths, secondary event paths, and resource impact paths.

[0118] The evolution probability and expected evolution time of the at least two candidate paths are determined, the evolution curves of the at least two candidate paths are determined based on the evolution probability and expected evolution time, and an event development trend diagram is constructed.

[0119] The risk level of the current urban emergency event is assessed from at least two dimensions based on the event development trend map; wherein the at least two dimensions include event status value, resource coverage, population exposure, and spread trend.

[0120] Optionally, before simulating the current urban emergency based on the pre-built knowledge graph and the event classification processing results, and before assessing the risk level of the current urban emergency, the device further includes a knowledge graph construction module, specifically including:

[0121] Based on historical urban emergency event databases, expert rule sets, urban static attribute information, IoT semantic mapping, and legal and policy knowledge, core node types, semantic edge relationships, and dynamic attributes are defined. The core node types include event types, resource types, location types, sensors, infrastructure, rules, and participating entities. Semantic edge relationships include causal relationships, spatial relationships, resource dependencies, impact relationships, and temporal attributes. Dynamic attributes include weights and constraints.

[0122] A knowledge graph is constructed based on the core node type, semantic edge relationships, and dynamic attributes.

[0123] Optional, intelligent response module, specifically used for:

[0124] Based on the location information and risk level assessment results of current urban emergency events, urban emergency resources are mapped and task chains are generated; wherein, the task chains include event reporting, traffic control, fire dispatch, rescue, and evacuation; urban emergency resources include available fire resources, rescue resources, and traffic control resources;

[0125] Based on the task chain and the mapping results of urban emergency resources, urban emergency resources are dispatched and response record reports are automatically generated.

[0126] The intelligent urban emergency response device provided in the embodiments of the present invention can execute the intelligent urban emergency response method provided in any of the embodiments of the present invention, and has the corresponding functions and beneficial effects of executing the intelligent urban emergency response method. For details, please refer to the relevant operations of the intelligent urban emergency response method in the foregoing embodiments.

[0127] Example 4

[0128] Figure 4 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0129] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0130] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0131] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as intelligent response methods for urban emergency events.

[0132] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0133] In some embodiments, the urban emergency response method can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the urban emergency response method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to execute the urban emergency response method by any other suitable means (e.g., by means of firmware).

[0134] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transferring data and instructions to the storage system, the at least one input device, and the at least one output device.

[0135] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0136] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0137] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0138] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0139] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0140] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0141] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A smart response method for urban emergency events, characterized in that, The method includes: Acquire multimodal data on current urban emergency events; The multimodal data is subjected to semantic modeling and event classification processing; wherein, the event classification processing includes urban emergency event type classification processing, urban emergency event scenario classification processing, and urban emergency event severity classification processing. Based on the pre-constructed knowledge graph and the results of event classification and processing, the current urban emergency events are simulated, and the risk level of the current urban emergency events is assessed. A task chain is generated based on the risk level assessment results, and urban emergency resources are dispatched according to the task chain.

2. The method according to claim 1, characterized in that, The multimodal data includes: video surveillance anomaly identification data, audio monitoring data, social media data, sensor access data, and weather and geographic access data.

3. The method according to claim 2, characterized in that, Obtain anomaly identification data from video surveillance of current urban emergency events, including: A target detection algorithm is used to detect targets in video frames and the detection results are determined. The detection results include the location information of the target detection box, the category label of the detected object, the confidence score of the detected object, the timestamp of the current video frame, and the video acquisition point information. The detection results are aggregated according to time windows, and a time series model or trajectory-based rule is used to make a preliminary judgment on urban emergency events in video frames in order to identify video surveillance anomaly identification data.

4. The method according to claim 1, characterized in that, The process of semantic modeling and event classification of the multimodal data includes: The multimodal data is encoded to obtain the encoded multimodal data; A cross-modal fusion model is used to fuse the encoded multimodal data into a unified vector, and output the confidence vector and semantic labels. Several continuous unified vectors are input into a time series network to classify current urban emergency events and determine their severity.

5. The method according to claim 1, characterized in that, Based on a pre-constructed knowledge graph and event classification processing results, the system simulates current urban emergency events and assesses their risk levels, including: Based on the event classification results, multimodal data is converted into semantic triples, and static attribute matching of current urban emergency events is performed based on a pre-constructed knowledge graph; wherein, the static attributes include building structure, road topology, population density, and key facilities; A path search is performed on a pre-built knowledge graph to identify at least two candidate paths; wherein the candidate paths consist of spatial sprawl paths, secondary event paths, and resource impact paths. The evolution probability and expected evolution time of the at least two candidate paths are determined, the evolution curves of the at least two candidate paths are determined based on the evolution probability and expected evolution time, and an event development trend diagram is constructed. The risk level of the current urban emergency event is assessed from at least two dimensions based on the event development trend map; wherein the at least two dimensions include event status value, resource coverage, population exposure, and spread trend.

6. The method according to claim 1, characterized in that, Before extrapolating current urban emergency events based on a pre-constructed knowledge graph and event classification processing results, and before assessing the risk level of current urban emergency events, the method further includes constructing a knowledge graph, specifically including: Based on historical urban emergency event databases, expert rule sets, urban static attribute information, IoT semantic mapping, and legal and policy knowledge, core node types, semantic edge relationships, and dynamic attributes are defined. The core node types include event types, resource types, location types, sensors, infrastructure, rules, and participating entities. Semantic edge relationships include causal relationships, spatial relationships, resource dependencies, impact relationships, and temporal attributes. Dynamic attributes include weights and constraints. A knowledge graph is constructed based on the core node type, semantic edge relationships, and dynamic attributes.

7. The method according to claim 1, characterized in that, The process of generating a task chain based on the risk level assessment results and dispatching urban emergency resources based on the task chain includes: Based on the location information and risk level assessment results of current urban emergency events, urban emergency resources are mapped and task chains are generated; wherein, the task chains include event reporting, traffic control, fire dispatch, rescue, and evacuation; urban emergency resources include available fire resources, rescue resources, and traffic control resources; Based on the task chain and the mapping results of urban emergency resources, urban emergency resources are dispatched and response record reports are automatically generated.

8. An intelligent response device for urban emergency events, characterized in that, The device includes: The multimodal data acquisition module is used to acquire multimodal data of current urban emergency events; The data processing module is used to perform semantic modeling and event classification processing on the multimodal data; wherein, the event classification processing includes urban emergency event type classification processing, urban emergency event scenario classification processing, and urban emergency event severity classification processing. The event simulation and risk assessment module is used to simulate current urban emergency events based on a pre-built knowledge graph and the results of event classification and processing, and to assess the risk level of current urban emergency events. The intelligent response module is used to generate a task chain based on the risk level assessment results, and to dispatch urban emergency resources according to the task chain.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the urban emergency response method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the intelligent response method for urban emergency events as described in any one of claims 1-7.

Images