Visualization-based emergency command method and system

Abstract

The invention relates to the technical field of emergency command, in particular to an emergency command method and system based on visualization, and the method comprises the following steps: S1, data collection and preprocessing; s2, data fusion and feature extraction; s3, carrying out three-dimensional visual modeling; s4, performing decision deduction and path planning; and S5, dynamic feedback and optimization are performed, interference of data noise and abnormal values is reduced by using a sliding window Gaussian filtering algorithm and artificial data verification based on a hidden Markov model, a risk assessment model is constructed for risk assessment, weights are determined through an analytic hierarchy process, multiple factors are comprehensively considered, risk assessment is more accurate, and the risk assessment efficiency is improved. An accurate decision-making basis is provided for commanders, and resource waste and rescue delay caused by blind decision making are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of emergency command, in particular to an emergency command method and system based on visualization. BACKGROUND

[0002] With the acceleration of global climate change and urbanization, the frequency and complexity of various emergencies such as natural disasters, public health events and man-made accidents are increasing. These events are often sudden, destructive and uncertain, posing a great threat to social safety and people's lives and property. In this context, the emergency command system has become the core and key to responding to emergencies and ensuring public safety.

[0003] However, the existing traditional emergency command system, in the data processing link, multi-source data such as Internet of Things sensors, unmanned aerial vehicle images and manual reporting, due to format heterogeneity, lack of unified space-time reference and effective cleaning mechanism. Although it can achieve the second-level targeted release of early warning information, it cannot properly solve the problems of sensor data noise interference and manual data verification, resulting in a big discount in data quality. Secondly, the lack of visualization is also a big problem. Most systems can only provide static three-dimensional scenes or two-dimensional heat maps, making it difficult to simulate the disaster diffusion process and the response of building structures in disasters in real time. In addition, the intelligence level of decision support is weak, and path planning and resource scheduling mostly rely on experience rules, lacking the ability to dynamically optimize according to actual conditions. It is impossible to find the best balance between response time, resource consumption and safety risks. Therefore, we propose an emergency command method and system based on visualization. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides an emergency command method and system based on visualization, thereby solving the technical problems mentioned in the background art.

[0005] To achieve the above purpose, the present application is realized by the following technical scheme:

[0006] An emergency command method based on visualization, comprising the following steps:

[0007] S1 data acquisition and preprocessing: deploy Internet of Things sensors, unmanned aerial vehicle aerial photography equipment, mobile terminals and manual reporting platforms, collect multiple types of data, set the collection frequency for different data types, use sliding window filtering algorithm to clean sensor data, introduce confidence check to clean manual reporting data, and transmit the cleaned data to edge computing nodes;

[0008] S2 Data fusion and feature extraction: The heterogeneous data is uniformly mapped to the GIS geographic coordinate system, the device positioning error is compensated by affine transformation, the multi-dimensional feature vector is constructed, and the environment, personnel and resource features are fused by entropy weight method and normalization processing;

[0009] S3 Three-dimensional visualization modeling: Import DEM data and building BIM data to generate three-dimensional terrain grid, superimpose real-time heat map layer, color mapping according to temperature value, solve partial differential equation by finite element method to simulate disaster spread, realize dynamic entity rendering, and design interactive function;

[0010] S4 Decision deduction and path planning: Construct risk assessment model, determine weight by AHP, improve Dijkstra algorithm, introduce cost function for path planning, and display planning results with different color paths and mark key point real-time data;

[0011] S5 Dynamic feedback and optimization: Real-time calculation of position deviation between actual trajectory and planning path, re-planning process is triggered when the deviation exceeds the threshold and lasts for a certain time, define abnormal state and set response mechanism, use stochastic gradient descent method to optimize risk assessment model, adjust path planning cost function coefficient by particle swarm optimization algorithm, and continuously optimize the system according to historical data and actual scene changes.

[0012] In one possible implementation, the sliding window filtering algorithm cleans the sensor data, and the formula is where D(t) represents the filtered data value at time t, C L (i) represents the original sensor data at time i, the sliding window half-width k = 5, and the Gaussian kernel parameter σ = 1.5.

[0013] In one possible implementation, the color mapping function is: where T represents the temperature, and the high-temperature region edge is processed by Gaussian blur to simulate the transition effect of fire spread.

[0014] In one possible implementation, the finite element method for solving partial differential equation is where represents the rate of change of variable φ with respect to time t, is the Hamiltonian operator, D is the diffusion coefficient, u is the wind speed vector, and S represents the source term.

[0015] In one possible implementation, the cost function formula is: C(n) = ∑ i (k1d i +k2ρ i +k3|Δh i |+k4T i ), where di L is the length of the road segment, p i | is the density of the flow, |Δh i | is the change in elevation, T i is the temperature influence factor, k1, k2, k3, k4 are 0.6, 0.3, 0.08 and 0.02 respectively.

[0016] In one possible implementation, a visual-based emergency command system includes a data acquisition and preprocessing module, a visualization module, a decision deduction module, a command and dispatch module, and an evaluation module.

[0017] The data acquisition and preprocessing module is used to acquire and preliminarily process multi-source heterogeneous data, including a perception unit and an edge computing unit; the perception unit collects emergency-related data by unifying data formats through protocol conversion middleware; and the edge computing unit preliminarily processes the collected data.

[0018] The visualization module is used to realize visual presentation and physical process simulation of an emergency scene, including a scene construction unit and a physical simulation unit; the scene construction unit is used to load and render three-dimensional geographic information and dynamic entities, import digital elevation model (DEM) and building BIM data to construct a realistic scene; and the physical simulation unit is used to simulate disaster diffusion and building structure response, solve partial differential equations to simulate disaster diffusion based on the finite element method, and simulate building structure stress using Hooke's law.

[0019] The decision deduction module is used to perform risk assessment and path planning based on visual data, including a risk assessment unit and a path planning unit; the risk assessment unit is used to calculate a multi-objective comprehensive risk value and construct a risk assessment model; and the path planning unit is used to generate optimal evacuation and rescue paths and set cost function parameters according to different emergency scenarios and requirements.

[0020] The command and dispatch module is used to realize dispatching of emergency resources and conveying of instructions, including a resource allocation unit and a communication management unit; the resource allocation unit is used to optimize material and personnel dispatching schemes; and the communication management unit is used to ensure instruction transmission.

[0021] The evaluation module is used to quantitatively evaluate the operation effect of the emergency command system, and uses a response efficiency evaluation formula to evaluate resource utilization and compare multiple evaluation indexes.

[0022] Compared with the prior art, the beneficial effects are:

[0023] 1. In this scheme, the sliding window Gaussian filtering algorithm and artificial data checking based on hidden Markov model are used to reduce the interference of data noise and abnormal values, and a risk assessment model is constructed for risk assessment, and the weights are determined by the analytic hierarchy process, and various factors are considered comprehensively, so that the risk assessment is more accurate, and accurate decision basis is provided for command personnel, and resource waste and rescue delay caused by blind decision are avoided;

[0024] 2. In this scheme, DEM data and building BIM model are combined to construct a more realistic emergency scene visualization scene. At the same time, the convection-diffusion equation is solved based on the finite element method to simulate disaster spread, and Hooke's law is used to simulate building structure response, so that the command personnel can intuitively understand the disaster development trend and building safety condition, and the interactive function design supports zooming, rotating and roaming of the three-dimensional scene, and the detailed attribute window can be popped up by clicking the entity, so that the command personnel can obtain real-time information, and the intuitiveness and interactivity of emergency command are improved, which is helpful for making quick decisions;

[0025] 3. In this scheme, by monitoring the deviation between the actual trajectory and the planned path in real time, the system can timely find and handle various problems in the emergency process, and the random gradient descent method is used to optimize the risk assessment model, and the particle swarm algorithm is used to adjust the path planning cost function coefficient, so that the system can continuously optimize according to historical data and actual scene changes. Ensure that the rescue and emergency disposal work is carried out efficiently, and comprehensively improve the adaptability and execution efficiency of the emergency command system in different environments. BRIEF DESCRIPTION OF DRAWINGS

[0026] The above description is only a summary of the technical scheme of the present application. In order to more clearly understand the technical means of the present application and can be implemented according to the content of the description, the following will be described in detail with the preferred embodiment of the present application and with the help of the drawings.

[0027] Figure 1 The flowchart of the emergency command method of the present application is shown in the figure;

[0028] Figure 2 The framework diagram of the emergency command system of the present application is shown in the figure. DETAILED DESCRIPTION

[0029] The preferred embodiment of the present application will be described in detail with reference to the accompanying drawings However, the present application can be realized in various different forms, therefore the present application is not limited to the embodiments described below, and in order to more clearly describe the present application, the components not connected with the invention will be omitted from the drawings;

[0030] The technical scheme in the embodiment of the present application is to solve the problems in the above background art, and the general idea is as follows:

[0031] Example 1

[0032] The embodiment introduces an emergency command method based on visualization, including the following steps:

[0033] S1 data acquisition and preprocessing

[0034] In the emergency command method, the quality problem of heterogeneous data is solved by constructing a multi-dimensional data acquisition network and targeted cleaning.

[0035] S11 data acquisition

[0036] In order to cover all information dimensions of emergency scenes, four types of data acquisition terminals are deployed to form a three-dimensional data acquisition system of physical environment, personnel state and resource distribution.

[0037] S111 physical environment data acquisition

[0038] Deploy an array of Internet of Things sensors, including:

[0039] Temperature and humidity sensor, measurement accuracy ±0.5℃ / ±2%RH, used to monitor real-time changes in environmental temperature and humidity, and capture temperature spikes in fire scenes;

[0040] Gas sensor, detection accuracy ±1ppm, used to real-time warning of toxic gas concentration in chemical leakage scenes;

[0041] Three-axis vibration sensor, sensitivity 100mV / g, used to collect building structure vibration data in earthquake monitoring, frequency response range 0.1-1000Hz.

[0042] The unmanned aerial vehicle cluster carries a visible light camera (resolution 4K) and an infrared thermal imager (temperature measurement range -20℃-500℃, accuracy ±2℃), and circulates aerial photography according to the preset route (height 200m, speed 15m / s), generating disaster area orthographic image and thermal distribution data.

[0043] S112 personnel state data acquisition

[0044] The personnel positioning terminal uses a UWB high-precision positioning bracelet (positioning accuracy ±0.3m), integrates an acceleration sensor and a heart rate module, and uploads real-time three-dimensional coordinates (X, Y, Z), motion state (still / walking / running) and physiological parameters (heart rate, oxygen saturation).

[0045] The rescue vehicle is installed with a vehicle terminal, which transmits real-time position, speed, fuel consumption and load data through Beidou / GPS dual-mode positioning.

[0046] S113 resource distribution data acquisition

[0047] Mobile APP is developed for manual reporting platform, which supports real-time information reporting by text, voice and pictures. The platform has built-in intelligent verification function, which automatically associates sensor data within 500m of the reported location to reduce duplicate reporting.

[0048] Differentiated collection frequency is set according to the dynamic characteristics of data:

[0049] High-frequency collection (1Hz): temperature, gas concentration, vibration data in fire scenarios to ensure capturing transient characteristics of disaster evolution;

[0050] Medium-frequency collection (0.2Hz): personnel location and vehicle trajectory data in earthquake evacuation scenarios to balance data density and transmission energy consumption;

[0051] Event-triggered collection: manual reporting data is uploaded in real time only when emergency button trigger or sensor threshold alarm (specifically, temperature > 70℃) is detected.

[0052] S12 data cleaning

[0053] A hierarchical cleaning strategy is adopted according to the noise characteristics of different data types.

[0054] S121 sensor data noise reduction

[0055] Sliding window Gaussian filter algorithm is used to process time series data, with the formula: Where D(t) represents the filtered data value at time t, C L (i) represents the original sensor data at time i, sliding window half-width k = 5 (covering 11 time points), Gaussian kernel parameter σ = 1.5, which can reduce sensor noise peak by 60%-80% according to actual measurement. Taking temperature data as an example, the root mean square error (RMSE) of the filtered waveform decreases from 3.2℃ of the original data to 1.1℃, effectively preserving the trend component.

[0056] S122 manual data verification

[0057] A behavior pattern library based on Hidden Markov Model (HMM) is constructed, with training data including 200,000 historical valid reporting records. Specific rules include:

[0058] Coordinate mutation detection: if the displacement of the adjacent two reports is > 200m and no transportation tool is matched, it is determined as invalid data;

[0059] Content consistency verification: combined with sensor data, if the manual report is "safe area" but the temperature within 50m is > 60℃ and the CO concentration is > 500ppm, it is marked as conflict data, triggering manual review process.

[0060] The cleaned data is transmitted to the edge computing node through a message queue (specifically, Kafka).

[0061] S2 data fusion and feature extraction

[0062] To solve the problem of spatial and temporal reference differences of heterogeneous data, spatial coordinate unification and multi-dimensional feature fusion are used to construct a standardized data set suitable for visual modeling.

[0063] S21 spatial alignment

[0064] A two-step method is used to achieve geographic coordinate unification of full-factor data.

[0065] S211 coordinate system conversion

[0066] Convert the heterogeneous coordinates such as UAV image pixel coordinates (pixel coordinate system) and bracelet UWB coordinates (local coordinate system) to WGS84 geographic coordinate system, and the core steps are as follows:

[0067] UAV image: Through POS data for aerial triangulation, combined with ground control points GCP to solve the exterior orientation elements, realize the mapping of image to GIS coordinates.

[0068] Positioning bracelet: Use reference tags with known positions (deployed at an interval of 100m) for coordinate calibration, eliminate non-line-of-sight errors NLoS of UWB positioning through weighted least squares method, and improve the accuracy after calibration.

[0069] S212 affine transformation compensation

[0070] For devices with systematic positioning deviation (specifically early vehicle-mounted GPS), a two-dimensional affine transformation model is used: Where (X, Y) is the original coordinate, (X', Y') is the transformed coordinate, matrix a, b, c, d represents rotation and scaling, e, f represents translation, and at least 3 ground calibration points (specifically road intersections) are used to fit the transformation parameters, so that the root mean square residual RMS is less than 0.3m, meeting the accuracy requirements of three-dimensional visualization.

[0071] S22 feature fusion

[0072] Construct a three-dimensional feature vector F = [f env ,f human ,f resource ] containing environment, personnel, and resources, and each component is weighted and normalized by entropy weight method:

[0073] Environmental feature f env : Fusion of temperature T, humidity H, gas concentration C, vibration acceleration A, etc., the calculation formula is: Where w iWeight representing the i-th environmental parameter. The weight is calculated by entropy weight method, which reflects the relative importance of each environmental parameter in the disaster scene, v i Actual measurement value representing the i-th environmental parameter, v i,min Minimum value representing the i-th environmental parameter, v i,max Maximum value representing the i-th environmental parameter. Specifically, in the fire scene, the temperature weight w T = 0.45, the CO concentration weight w C = 0.35, which is trained based on historical data of 100 fires.

[0074] Personnel feature f human : Personnel density p, moving speed v, and aggregation degree C h (calculated by DBSCAN algorithm, value 0-1), where N is the number of people in the area, S is the area, and the DBSCAN parameter is set to neighborhood radius 5m and minimum point number 5.

[0075] Resource feature f resource : Integrated material reserve M, vehicle available time t avail , and rescue team response distance d resp , through the normalization formula: Where M max is the maximum reserve of materials, t max is the maximum available time of the vehicle, and d min is the distance of the nearest available resource point.

[0076] S3 Three-dimensional visualization modeling

[0077] Through scene modeling and physical process simulation, abstract data is converted into intuitive and interactive three-dimensional visualization interface, which is convenient for intuitive and visual understanding of the situation for command and decision-making.

[0078] S31 Scene construction

[0079] Basic three-dimensional scene generation: Import DEM data and building BIM model (including beam column structure, door and window position, fire resistance rating, etc.) at 1:500 scale, generate terrain grid M = {V, E} through triangulation algorithm, where the vertex V coordinate precision reaches centimeter level, and the edge relationship E supports collision detection and path planning.

[0080] Integrate city component data (roads, bridges, fire hydrant locations), build emergency thematic layers containing multiple elements, support hierarchical display (such as only showing fire access, refuge sites).

[0081] Thermal rendering: Use piecewise linear interpolation method to generate temperature thermal map, and the color mapping function is defined as:

[0082] High temperature area edge is processed by Gaussian blur, simulating the transition effect of fire spread.

[0083] Trajectory visualization: personnel and vehicle trajectories are displayed with timestamped dynamic trails, trail colors change with speed (green ≤5m / s, yellow 5-15m / s, red >15m / s), supporting backtracking past 30 minutes of trajectory.

[0084] S32 dynamic rendering

[0085] Disaster spread simulation: based on the finite element method to solve the convection-diffusion equation, the model control equation is: Where φt represents the rate of change of variable φ (i.e. the concentration of diffusing substances such as smoke, toxic gases, etc.) over time t, is the diffusion term, where is the Hamiltonian operator, representing the change in space, and D is the diffusion coefficient, which is related to the properties of the medium, and different media have different diffusion coefficients. represents the diffusion flux caused by the concentration gradient, while represents the divergence of the diffusion flux in space, reflecting the diffusion trend of substances at different points in space due to uneven concentration; is the convection term, u is the wind speed vector, describing the flow speed and direction of the fluid in space, uφ represents the flux caused by the flow of the fluid carrying the substance, and the negative sign in front indicates that the direction of the flux is opposite to the direction of the flow speed. This term reflects the influence of fluid flow on the distribution of substance concentration, i.e. the migration phenomenon of substances with the flow of fluid; S represents the source term, representing the generation or disappearance rate of substances. Tetrahedral mesh partitioning (average mesh size 2m) is adopted, and implicit difference method is used for solving, with time step set to 1s, ensuring the balance between simulation accuracy and real-time performance.

[0086] Interactive function design: support zooming (0.1-10 times), rotating (6DOF) and roaming in 3D scene, click on entity to pop up attribute window (specifically, click on fire truck to display: license plate number, carrying capacity 3 tons, remaining fuel 40%, estimated arrival time 8 minutes).

[0087] S4 decision-making and path planning

[0088] Based on the real-time data provided by the visualization scene, executable command strategies are generated through multi-objective optimization algorithm, realizing the transformation from data visualization to intelligent decision-making.

[0089] S41 risk assessment

[0090] A three-element risk assessment model including time, cost and safety is constructed, and the objective function is: min(αtresp + βC resource + γR collapse ), where t resp is the rescue response time, defined as the time from decision making to the arrival of resources at the core area of the disaster, pre-calculated by path planning algorithm, C resource is the resource consumption cost, including human cost, vehicle wear and tear, and material consumption, R collapse building collapse risk (0-1), calculated by Bayesian network based on vibration acceleration, temperature stress, and BIM structure parameters.

[0091] The weights are determined by the analytic hierarchy process (AHP), specifically, in the fire scenario, α = 0.5 (response time priority), β = 0.3, and γ = 0.2; in the earthquake scenario, α = 0.3, β = 0.1, and γ = 0.6 (structure safety priority), and the weight matrix consistency ratio CR < 0.1.

[0092] S42 path planning

[0093] The Dijkstra algorithm is improved by introducing a cost function that includes terrain, human flow, and energy consumption: C(n) = ∑ i (k1d i +k2ρ i +k3|Δh i |+k4T i ), where d i is the road segment length, ρ i is the human flow density, |Δh i | is the elevation change, T i is the temperature influence factor, and k1, k2, k3, and k4 are 0.6, 0.3, 0.08, and 0.02, respectively.

[0094] The rescue vehicle path needs to meet the height and weight limits, and the evacuation path needs to include at least two safe exits. The R-tree is used to quickly filter out infeasible road segments.

[0095] The planning results are displayed as paths of different colors, with green being the optimal path (with the smallest cost), yellow being the alternative path (with moderate cost), and red being the prohibited road segment (such as a bridge with insufficient load capacity). Real-time data of key points on the path (such as the current human flow density, temperature, etc.) are also marked.

[0096] S5 dynamic feedback and optimization

[0097] Through real-time monitoring of the execution process and iteration of model parameters, the emergency command system is continuously optimized as the scene changes.

[0098] S51 real-time monitoring

[0099] Real-time calculation of the positional deviation between the actual trajectory and the planned path: where (x i (t),y i (t)) represents the coordinates of the i-th point on the actual trajectory at time t, where (x i (t),y i (t)) represents the coordinates of the i-th point on the planned path at time t. When δ(t) > 30 m and lasts for more than 3 time steps, the re-planning process is triggered. The re-planning algorithm preferentially uses the calculated neighborhood node data, and the response time is controlled within 2 s.

[0100] In addition, 12 types of abnormal states (such as vehicle failure, sensor drop, and personnel gathering overload) are defined, and each abnormality is associated with a 3-level response mechanism:

[0101] Level 1 (yellow warning): Specifically, the remaining amount of materials < 30%, triggering a nearby warehouse dispatch reminder;

[0102] Level 2 (orange warning): Specifically, the personnel density > 8 people / m 2 , automatically adjusting the evacuation path to the standby channel;

[0103] Level 3 (red warning): Specifically, the building collapse risk > 0.8, immediately starting the regional blockade instruction.

[0104] S52 model update

[0105] The random gradient descent method is used to optimize the risk assessment model, and the objective function is the root mean square error (RMSE): The weight update formula is: where η = 0.001, and after completing 10 rescue missions, the historical data (sample size ≥ 500) are used for batch update to ensure the stability of weight adjustment.

[0106] Weekly regular analysis of path planning efficiency indicators:

[0107] Planning accuracy = number of paths successfully reaching the target point / total planning times;

[0108]

[0109] The particle swarm optimization (PSO) algorithm is used to adjust the cost function coefficients. Specifically, in the winter snow scene, the elevation change weight k3 is increased to 0.15 to reduce the risk of ice and snow road travel.

[0110] Example 2:

[0111] This embodiment introduces an emergency command system based on visualization, which includes a data acquisition and preprocessing module, a visualization module, a decision deduction module, a command and dispatch module, and an evaluation module.

[0112] 1 Data acquisition and preprocessing module

[0113] The data acquisition and preprocessing module includes a perception unit and an edge computing unit.

[0114] 1.1 Perception unit

[0115] The perception unit is the data entry of the entire emergency command system, responsible for collecting a wide range of emergency-related data from different channels, including Internet of Things sensors, video surveillance, and manual terminal data. The perception unit uses protocol conversion middleware to access and format data from various devices, solving the problem of inconsistent data formats. Using a dynamically configurable sampling frequency mechanism, the system can adjust data collection frequency based on the urgency of the emergency scene and the frequency of data changes. Specifically, during the initial stage of a fire, the sampling frequency of temperature sensors is increased to multiple times per second to more accurately monitor the development of the fire.

[0116] 1.2 Edge computing unit

[0117] The edge computing unit is mainly responsible for preliminary processing of data at the source of data collection, reducing the pressure on subsequent data transmission and processing, and timely detecting anomalies in the data to ensure data quality. This unit runs a lightweight filtering algorithm to process sensor data in real time and remove noise interference. The system uses a real-time anomaly detection algorithm based on Z-score, with the formula: where X is the current data value, μ is the sliding window mean, and σ is the standard deviation. When |Z| > 3, the data is marked as abnormal. When monitoring heart rate data, if the Z value of the heart rate data at a certain time is greater than 3, the system will immediately mark the data as abnormal and perform further analysis and processing.

[0118] 2 Visualization module

[0119] The visualization module includes a scene construction unit and a physical simulation unit.

[0120] 2.1 Scene construction unit

[0121] The scene building unit is responsible for converting various geographic information and emergency-related data into intuitive three-dimensional visual scenes, providing clear and comprehensive emergency scene pictures for command personnel, enabling them to quickly understand the on-site situation. This unit uses WebGL and the Three.js framework for three-dimensional scene rendering and display. Using the LOD (Level of Detail) technique, the model details are loaded based on the user's perspective and the distance of the scene objects. When the user's perspective is far from the building, a low-detail model is loaded to improve rendering efficiency; when the perspective is close to the building, a high-detail model is automatically loaded to display more detailed information about the building. By importing digital elevation model DEM and building BIM data, a realistic three-dimensional terrain and building model is constructed.

[0122] 2.2 Physical simulation unit

[0123] The physical simulation unit is used to simulate the disaster diffusion process and the response of building structures in disasters, helping command personnel predict disaster trends, develop countermeasures in advance, and protect personnel safety and reduce property losses.

[0124] For disaster diffusion simulation, the finite element method in Example One is used to solve the related partial differential equations. By solving the equation and combining the actual scene parameters, such as ventilation conditions in buildings, fire location and intensity, the diffusion process of smoke in space is simulated.

[0125] In terms of building structure response simulation, Hooke's Law is used: F = kΔx, where F is the force, k is the structural stiffness coefficient, and Δx is the displacement vector. According to the structural design parameters and material properties of the building, the value of k is determined. In earthquake simulation, the stability of the building structure is judged by calculating the displacement vector of different parts, and the possible damage location and degree are predicted.

[0126] 3 Decision deduction module

[0127] The decision deduction module includes a risk assessment unit and a path planning unit.

[0128] 3.1 Risk assessment unit

[0129] The risk assessment unit considers multiple factors to quantitatively assess the risk of the current emergency scene, providing a scientific basis for decision-making for command personnel, helping them judge the severity of the disaster and the possible impact. This unit builds a three-element risk assessment model containing time, cost, and safety, with the objective function being the model function as in Example One. Considering the uncertainty of risk factors, the comprehensive risk value is more accurately evaluated. In assessing earthquake disaster risk, by simulating different earthquake magnitudes and focal depths multiple times, the probability distribution of personnel casualties and property losses under different scenarios is calculated, resulting in more accurate risk assessment data and results.

[0130] 3.2 Path planning unit

[0131] The path planning unit generates optimal evacuation and rescue paths according to real-time conditions and target requirements of emergency scenarios, ensuring that personnel can quickly and safely evacuate and rescue forces can efficiently reach designated locations. The improved Dijkstra algorithm is used, introducing a cost function that includes road segment length, pedestrian density, elevation change, and risk factors. The function is shown in Example One, and different adjustment coefficients are set according to different emergency scenarios and requirements. Specifically, in a fire scenario, the weights of pedestrian density and temperature risk can be appropriately increased to evacuate personnel more quickly. In a mountain rescue scenario, the weight of elevation change is increased. At the same time, considering the traffic restrictions of roads, such as height limit, weight limit, and real-time road conditions, the path is optimized and screened to ensure that the generated path is feasible and optimal.

[0132] 4 Command and dispatch module

[0133] The command and dispatch module includes a resource allocation unit and a communication management unit.

[0134] 4.1 Resource allocation unit

[0135] The resource allocation unit reasonably allocates resources and personnel according to the requirements of emergency scenarios and existing resource conditions, improves resource utilization efficiency, and ensures the smooth progress of rescue and emergency disposal work. This unit uses the Hungarian algorithm in the transportation problem to solve the optimal matching problem. The supply points and demand points of materials are matched, considering factors such as the type, quantity, transportation cost of materials, and the urgency of demand. Assuming there are multiple material storage warehouses and multiple disaster areas, each warehouse has different types and quantities of materials, and each disaster area has different material demands. Through the Hungarian algorithm, the optimal material allocation scheme can be found to minimize transportation costs and meet the needs of disaster areas. In terms of personnel scheduling, according to the nature of rescue tasks and factors such as personnel skills and experience, personnel are reasonably arranged to different posts and tasks to achieve optimal allocation of human resources.

[0136] 4.2 Communication management unit

[0137] The communication management unit is responsible for ensuring the reliability of command transmission in the emergency command system, ensuring that the decisions of the command personnel can be timely and accurately conveyed to the execution personnel, while ensuring that the on-site data can be transmitted back to the command center in real time and stably. This unit uses a combination of multiple communication technologies to improve the coverage and stability of communication. Through data redundancy transmission and error checking mechanism, the integrity and accuracy of data in the transmission process are ensured. When sending instructions, multiple copies of the same data are sent at the same time, and CRC cyclic redundancy check algorithm is used to check the data. The receiving end judges whether the data is correctly received according to the check result, and if an error is found, it requires retransmission.

[0138] 5evaluation module

[0139] The evaluation module quantitatively evaluates the operation effect of the entire emergency command system, provides basis for the optimization and improvement of the system, and helps to continuously improve the efficiency and quality of emergency command. A variety of evaluation indexes are used to comprehensively evaluate the system. The response efficiency evaluation uses the formula Where, t ideal is the theoretical optimal time, t actual is the actual time spent. When evaluating the response efficiency of the rescue operation, the shortest time required to complete the rescue task in theory is compared with the actual time spent, and the response efficiency is calculated. The resource utilization rate evaluation measures the utilization of resources by calculating the proportion of the actual amount of resources used to the total amount of resources. Specifically, in a fire rescue, the actual amount of extinguishing agent consumed, the fuel consumption of rescue vehicles, and the number of personnel participating in the rescue are counted and compared with the total amount of resources reserved to evaluate the resource utilization rate. At the same time, the accuracy of the decision is evaluated by comparing the actual rescue effect with the expected target, analyzing whether the decision is reasonable and effective, and providing reference and improvement direction for subsequent decision-making.

[0140] Finally, it should be noted that: obviously, the above embodiments are only examples for clearly illustrating the present application, and are not limited to the implementation. For ordinary skilled in the art, on the basis of the above description, other different forms of changes or variations can also be made. Here, it is not necessary and impossible to exhaust all the implementation. The obvious changes or variations derived therefrom are still within the protection scope of the present application.

Claims

1. A visualization-based emergency command method, characterized in that: The following steps are involved: S1 Data Collection and Preprocessing: Deploy IoT sensors, drones, mobile terminals, and manual reporting platforms to collect multiple types of data. Set collection frequencies for different data types, use a sliding window filtering algorithm to clean sensor data, introduce confidence checks to clean manually reported data, and transmit the cleaned data to edge computing nodes. S2 Data Fusion and Feature Extraction: Heterogeneous data are uniformly mapped to the GIS geographic coordinate system, affine transformation is used to compensate for equipment positioning errors, and a multi-dimensional feature vector is constructed. Each component is weighted and normalized using the entropy weight method to integrate environmental, personnel, and resource characteristics. S3 3D visualization modeling: Import elevation model DEM data and building BIM data to generate a 3D terrain grid, overlay a real-time thermal layer, perform color mapping based on temperature values, solve partial differential equations based on the finite element method to simulate disaster spread, implement dynamic entity rendering, and design interactive functions; S4 Decision-making and Path Planning: Build a risk assessment model, determine weights through the Analytic Hierarchy Process, improve the Dijkstra algorithm, introduce a cost function for path planning, and display the planning results as color-coded paths with real-time data of key points. S5 Dynamic Feedback and Optimization: Calculates the position deviation between the actual trajectory and the planned path in real time. When the deviation exceeds a threshold and persists for a certain period of time, triggers the replanning process, defines abnormal states and sets response mechanisms, uses stochastic gradient descent to optimize the risk assessment model, and uses particle swarm optimization to adjust the path planning cost function coefficients. Continuously optimize the system based on historical data and actual scenario changes.

2. The visualization-based emergency command method according to claim 1, characterized in that: The sliding window filtering algorithm cleans the sensor data, and the formula is Where D(t) represents the filtered data value at time t, C L (i) represents the original sensor data corresponding to time i, the sliding window half-width k = 5, and the Gaussian kernel parameter σ = 1.

5.

3. The visualization-based emergency command method according to claim 1, characterized in that: The color mapping function is: Among them, T represents temperature, and the edges of the high-temperature area are processed by Gaussian blur to simulate the transition effect of fire spread.

4. The visualization-based emergency command method according to claim 1, characterized in that: The finite element method solves the partial differential equation as in, represents the rate of change of the variable φ with time t, ▽ is the Hamiltonian operator, D is the diffusion coefficient, u is the wind speed vector, and S represents the source term.

5. The visualization-based emergency command method according to claim 1, characterized in that: The cost function formula is: C(n)=∑ i (k1d i +k2ρ i +k3|Δh i |+k4T i ), where d i is the length of the road section, ρ i is the human flow density, |Δh i | is the elevation change, T i is the temperature influence factor, k1, k2, k3, and k4 are 0.6, 0.3, 0.08, and 0.02, respectively.

6. A visualization-based emergency command system that executes the visualization-based emergency command method according to claims 1 to 5, characterized in that: It includes data acquisition and preprocessing module, visualization module, decision-making and deduction module, command and dispatch module and evaluation module; The data acquisition and preprocessing module is used to acquire and preliminarily process multi-source heterogeneous data, including a perception unit and an edge computing unit; the perception unit unifies the data format through a protocol conversion middleware and collects emergency-related data; The edge computing unit performs preliminary processing on the collected data; The visualization module is used to realize the visualization presentation of emergency scenarios and physical process simulation, including a scenario construction unit and a physical simulation unit; The scene construction unit is used to load and render three-dimensional geographic information and dynamic entities, import digital elevation model (DEM) and building BIM data to build realistic scenes; the physical simulation unit is used to simulate disaster spread and building structure response, solve partial differential equations based on the finite element method to simulate disaster spread, and use Hooke's law to simulate building structure stress; The decision-making deduction module is used to perform risk assessment and path planning based on visual data, including a risk assessment unit and a path planning unit; The risk assessment unit is used to calculate the multi-objective comprehensive risk value and construct a risk assessment model; the path planning unit is used to generate the optimal evacuation and rescue path and set the cost function parameters according to different emergency scenarios and needs; The command and dispatch module is used to dispatch emergency resources and convey instructions, and includes a resource allocation unit and a communication management unit; the resource allocation unit is used to optimize the material and personnel dispatch plan; The communication management unit is used to ensure instruction transmission; The evaluation module is used to quantitatively evaluate the operational effect of the emergency command system, use the response efficiency evaluation formula to evaluate resource utilization, and compare multiple evaluation indicators.

Images