Key infrastructure risk identification method and device in flood disaster chain scene

Abstract

The invention discloses a key infrastructure risk identification method and device in a flood disaster chain scene. The device comprises an acquisition module used for acquiring multi-source data and performing time-space alignment and fusion processing; the extraction module is used for extracting spatio-temporal evolution characteristics of a flood disaster chain based on the multi-source data; the construction module is used for constructing a key infrastructure coefficient according to the prior infrastructure information; the coupling module is used for coupling the spatio-temporal evolution characteristics and the key infrastructure coefficient, and calculating a regional risk index through a dynamic coupling coefficient and a space sensitivity factor; and the early warning module is used for generating a risk space distribution map and triggering dynamic early warning. According to the technical scheme, through multi-source data collection and space-time fusion, the space-time evolution characteristics of the flood disaster chain are extracted, key infrastructure coefficients based on information of buildings, roads, population and the like are constructed, the regional risk index is calculated through dynamic coupling, risk distribution is visually reflected, accurate early warning and emergency response are achieved, and the reliability of the system is improved. The urban disaster prevention and reduction level is effectively improved, and the public safety guarantee efficiency is remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster prevention and mitigation, and particularly to a method and device for risk identification of critical infrastructure in the scenario of a flood disaster chain. Background Art

[0002] In recent years, the continuous intensification of global climate change and urbanization has led to frequent extreme meteorological events, and flood disasters have shown the characteristics of a flood disaster chain. Heavy rain triggers floods, which in turn cause waterlogging and may induce chain effects such as traffic jams, posing a major threat to urban critical infrastructure (such as buildings, roads, and public service facilities). Existing technologies mainly focus on the static risk assessment of single disaster events, and most methods rely on online maps, remote sensing images, or local sensor data, making it difficult to comprehensively reflect the spatio-temporal evolution characteristics among various links in the flood disaster chain and their cumulative impact on the vulnerability of infrastructure. At the same time, the real-time data update and dynamic early warning response capabilities are insufficient, making it difficult to meet the actual needs of disaster prevention, mitigation, and emergency management. How to properly solve the above problems has become an urgent issue in the industry. Summary of the Invention

[0003] The present invention provides a method and device for risk identification of critical infrastructure in the scenario of a flood disaster chain. This method integrates multi-source data, extracts the characteristics of the flood disaster chain, dynamically calculates the risk index, and promotes precise early warning and disaster prevention improvement.

[0004] According to the first aspect of the present invention, there is provided a system for risk identification of critical infrastructure in the scenario of a flood disaster chain, which includes:

[0005] A collection module, configured to collect multi-source data and perform spatio-temporal alignment and fusion processing, where the multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data;

[0006] An extraction module, configured to extract the spatio-temporal evolution characteristics of the flood disaster chain based on the multi-source data, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain;

[0007] A construction module, configured to construct a critical infrastructure coefficient according to prior infrastructure information, where the critical infrastructure coefficient is based on any one or more of building density, road network density, population density, function type factor, and spatial connectivity;

[0008] A coupling module, configured to couple the spatio-temporal evolution characteristics and the critical infrastructure coefficient, and calculate the regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor;

[0009] An early warning module for generating a risk space distribution map and triggering dynamic early warnings.

[0010] In one embodiment, the acquisition module is further configured to: automatically identify and classify buildings, roads, and infrastructure through drone aerial photography and image recognition technologies; perform spatio-temporal alignment and resolution unification on multi-source data using geocoding and spatial interpolation technologies; establish a dynamic data update mechanism based on a cloud platform, and real-time correct meteorological, traffic, and infrastructure status data through edge computing.

[0011] In one embodiment, the extraction module is further configured to: model the causal relationships of rainstorms, floods, waterlogging, and traffic jams through a Bayesian network, and quantify the triggering threshold and delay parameters; analyze the propagation paths of flood disaster chain events using spatio-temporal clustering algorithms and diffusion models, and construct a spatio-temporal evolution matrix; dynamically correct flood disaster chain parameters through chain triggering coefficients, cumulative effect indices, and delay impact factors.

[0012] In one embodiment, the construction module is further configured to: identify the structural types and functional attributes of buildings, bridges, and drainage facilities based on deep learning algorithms; calculate the spatial distribution indicators of building density, road density, and population density; calculate the key infrastructure coefficient through the following formula:

[0013] KIC = α×BD + β×RD + γ×PD + δ×FT + ε×SI

[0014] Where the symbol KIC is the key infrastructure coefficient, the symbol BD is the building density, the symbol RD is the road network density, the symbol PD is the population density, the symbol FT is the functional type factor, the symbol SI is the spatial connectivity, and the weights α, β, γ, δ, ε are automatically calibrated through a machine learning model and historical disaster data.

[0015] In one embodiment, the coupling module is further configured to:

[0016] Dynamically adjust the weight w of flood disaster chain parameters by combining fuzzy comprehensive evaluation and entropy weight method j ; calculate the regional risk index based on the following formula:

[0017]

[0018] Where the symbol R is the regional risk index, the symbol C t is the cumulative impact coefficient of the flood disaster chain, the symbol P j is the parameter of each link in the flood disaster chain, the symbol S i is the spatial sensitivity factor; perform sensitivity analysis on the weights and parameters through Monte Carlo simulation to generate a risk confidence interval.

[0019] In one embodiment, the early warning module is further configured to: map the risk index to a heat map through a GIS platform, and overlay the infrastructure distribution and the evolution path of the flood disaster chain; set a risk threshold, and trigger a warning message in combination with the real-time data of the Internet of Things sensors; send a classified warning instruction to the emergency management department through SMS, email, and APP push methods.

[0020] In one embodiment, it further includes:

[0021] An emergency module, configured to receive the risk space distribution map and the classified warning instruction generated by the early warning module, determine the key protection area according to the regional risk index; generate a risk conduction topology map based on the dynamic coupling coefficient, deploy a precise on-site duty plan for transportation hubs, power facilities, and communication nodes, and configure emergency supplies and mobile drainage equipment; monitor the water level, structural deformation, and equipment operation data of the on-site area in real time through the Internet of Things sensors, and use the mobile command terminal to feedback the on-site situation information.

[0022] In one embodiment, it further includes:

[0023] A dispatching module, configured to adjust the deployment density of the stationed personnel and the material dispatching path through a multi-department collaborative platform according to the dynamic correction result of the spatio-temporal evolution matrix of the flood disaster chain; when the risk confidence interval exceeds the preset threshold, trigger a chain emergency response of traffic control - medical rescue - power supply, and synchronously update the spatial connectivity parameter in the infrastructure coefficient; perform local model training on the monitoring data of the stationed points through the edge computing nodes, and dynamically optimize the weight allocation strategy of the key infrastructure coefficient.

[0024] According to the second aspect of the present invention, an electronic device is provided, and the electronic device includes: a communication interface, a processor, and a memory;

[0025] Wherein, the memory is used to store program instructions, and when the program instructions are executed by the processor communicatively connected to the memory through the communication interface, the above-mentioned key infrastructure risk identification system in any flood disaster chain scenario is implemented.

[0026] According to the third aspect of the present invention, a computer-readable storage medium is provided, and computer program instructions are stored on the computer-readable storage medium. When the computer program instructions are executed by a computer (for example, the processor in the computer), the above-mentioned key infrastructure risk identification system in any flood disaster chain scenario is implemented.

[0027] In summary, the present invention provides a key infrastructure risk identification system and device in the scenario of a flood disaster chain. The method includes: collecting multi-source data and performing spatio-temporal alignment and fusion processing, where the multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data; based on the multi-source data, extracting the spatio-temporal evolution characteristics of the flood disaster chain, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain; according to prior infrastructure information, constructing a key infrastructure coefficient, which is based on any one or more of building density, road network density, population density, functional type factor, and spatial connectivity; coupling the spatio-temporal evolution characteristics and the key infrastructure coefficient, calculating the regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor, generating a risk spatial distribution map, and triggering a dynamic warning. The technical solution of this application extracts the spatio-temporal evolution characteristics of the flood disaster chain through multi-source data collection and spatio-temporal fusion, constructs a key infrastructure coefficient based on information such as buildings, roads, and population, calculates the regional risk index through dynamic coupling, intuitively reflects the risk distribution, realizes accurate warning and emergency response, effectively improves the urban disaster prevention and mitigation level, and significantly enhances the public safety guarantee efficiency.

[0028] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in the written specification and the drawings.

[0029] The following will further describe the technical solution of the present invention in detail through the drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0031] Figure 1 Schematic diagram of a key infrastructure risk identification system in the scenario of a flood disaster chain provided for an embodiment of the present invention;

[0032] Figure 2 Structural diagram of an electronic device provided for an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The features and exemplary embodiments of various aspects of the present application will be described in detail below. To make the objectives, technical solutions, and advantages of the present application clearer and more understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by showing examples of the present application.

[0034] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, elements defined by the statement "comprising..." do not exclude the presence of additional identical elements in the process, method, article or device comprising the said elements.

[0035] As Figure 1 shown, the present invention provides a risk identification system for critical infrastructure in the scenario of a flood disaster chain. The risk identification system for critical infrastructure in the scenario of a flood disaster chain includes:

[0036] A collection module, configured to collect multi-source data and perform spatio-temporal alignment and fusion processing. The multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data;

[0037] An extraction module, configured to extract spatio-temporal evolution features of the flood disaster chain based on the multi-source data, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain;

[0038] A construction module, configured to construct a critical infrastructure coefficient according to prior infrastructure information. The critical infrastructure coefficient is based on any one or more of building density, road network density, population density, functional type factor, and spatial connectivity;

[0039] A coupling module, configured to couple the spatio-temporal evolution features and the critical infrastructure coefficient, and calculate a regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor;

[0040] An early warning module, configured to generate a risk spatial distribution map and trigger a dynamic early warning.

[0041] In one embodiment, a risk identification system based on the coupling of the spatio-temporal evolution characteristics of flood disaster chains and the spatial attributes of critical infrastructure is proposed. By collecting and integrating online maps, remote sensing images, Internet of Things sensors, meteorological monitoring, and historical disaster case data, the system constructs a causal relationship model and a spatio-temporal evolution matrix of flood disaster chains to achieve dynamic modeling of disaster links such as rainstorms, floods, waterlogging, and traffic jams. At the same time, based on prior information such as building density, road network density, population density, functional type factors, and spatial connectivity, a critical infrastructure coefficient is constructed, and the regional risk index is calculated through dynamic coupling coefficients and spatial sensitivity factors. Finally, a risk spatial distribution map is generated through the GIS platform and combined with real-time data to trigger dynamic early warnings, effectively solving the problems of single data, inaccurate risk assessment, and lagging early warning response in the existing technology, and improving the scientificity and effectiveness of disaster prevention, mitigation, and emergency response.

[0042] The function of the acquisition module is to integrate multi-source heterogeneous data, including online maps (such as road networks and building distributions), remote sensing images (satellite / drone aerial photography), Internet of Things sensors (water level and structural deformation monitoring), meteorological monitoring (real-time rainfall and wind speed), and historical disaster case data (spatio-temporal records of flood events). Further, drone aerial photography and image recognition technology are used to automatically label building, road, and infrastructure types; the coordinate system is unified through geocoding technology, and spatial interpolation algorithms are used to align data with different resolutions (such as fusing low-resolution meteorological data with high-precision remote sensing images); data dynamic updates are realized based on the cloud platform, for example, the traffic congestion or drainage facility status collected by sensors is corrected in real time through edge computing nodes.

[0043] The function of the extraction module is to analyze the dynamic propagation law of flood disaster chains and construct a causal chain model of rainstorm → flood → waterlogging → traffic jam. Further, Bayesian networks are used to quantify the disaster trigger threshold (such as the relationship between rainstorm intensity and the probability of flood occurrence) and the time delay parameter (such as the lag time from flood to waterlogging); the spatio-temporal clustering algorithm (such as DBSCAN) is used to identify the hot spots of disaster diffusion, and combined with the diffusion model to generate a spatio-temporal evolution matrix, which characterizes the spread path of the flood disaster chain in space and the intensity change in time; the chain trigger coefficient (such as the conduction weight of waterlogging on traffic jams) and the cumulative effect index (such as the pressure accumulation of continuous rainfall on the drainage system) are dynamically corrected.

[0044] The function of the construction module is to quantify the vulnerability of critical infrastructure and generate a spatialized critical infrastructure coefficient (KIC). Further, based on the GIS platform, the spatial distributions of building density, road network density, and population density are extracted; the functional types of infrastructure are identified through deep learning models (such as CNN) (such as hospitals and bridges are given higher weights); the spatial connectivity index (such as the road connectivity between a substation and surrounding hospitals) is calculated.

[0045] The function of the coupling module is to integrate the evolution characteristics of the flood disaster chain and the vulnerability of infrastructure to generate a dynamic risk index. Further, spatial superposition of the disaster parameters (such as the depth of waterlogging) in the spatio-temporal evolution matrix and the KIC value is carried out, and the local risk value is calculated through the dynamic coupling coefficient (reflecting the correlation strength between disasters and facilities); a spatial sensitivity factor (such as the sensitivity coefficient of low-lying areas to floods) is introduced, and a risk confidence interval is generated in combination with Monte Carlo simulation; based on the edge computing node, local model iteration is carried out on real-time sensor data (such as a sudden rise in water level), and the weight allocation strategy is dynamically adjusted.

[0046] The function of the early warning module is to visualize the risk distribution and trigger a hierarchical response. Further, a heat map is generated through the GIS platform, and the location of infrastructure and the evolution path of the flood disaster chain are superimposed (such as highlighting the hospital areas that may be flooded by floods in red); risk thresholds are set (such as KIC>0.8 and rainfall>50mm / h), and text messages and APP push are triggered to the emergency management department; it is linked with the traffic signal system and the medical rescue platform, and a chain of emergency instructions for traffic control - power supply guarantee - personnel evacuation is automatically activated when the risk confidence interval exceeds the limit.

[0047] Taking the rainstorm and flood scenario of a certain city as an example, the distribution of urban buildings is updated by drone aerial photography, the sensors deployed on the bridge transmit the water level data in real time, and the meteorological radar predicts the rainfall distribution in the next 3 hours; the Bayesian network determines that the rainstorm intensity has reached the flood trigger threshold, and the spatio-temporal evolution matrix shows that the flood will spread along the river to the eastern low-lying area; if the KIC value in the eastern region is relatively high (due to dense residential areas + old drainage systems), the coupling module calculates that the risk index in this region exceeds the threshold; the GIS platform marks the eastern region as a red high-risk area, sends instructions to the emergency center, dispatches mobile drainage vehicles to key bridges, and closes the flooded sections through the traffic signal system.

[0048] Facing the problem of risk identification of critical infrastructure in the context of the flood disaster chain, through the collection and integration of multi-source data, the spatio-temporal evolution characteristics of the flood disaster chain are extracted, and then combined with prior infrastructure information to construct a critical infrastructure coefficient. Finally, the regional risk index is calculated through the coupling of the two, so as to generate a risk spatial distribution map and trigger a dynamic early warning. Collect multi-source data and perform spatio-temporal alignment and fusion processing; extract the spatio-temporal evolution characteristics of the flood disaster chain based on multi-source data; construct a critical infrastructure coefficient according to prior infrastructure information; couple the spatio-temporal evolution characteristics and the critical infrastructure coefficient, calculate the regional risk index and generate a risk early warning. It not only considers the impact of a single disaster event, but comprehensively and dynamically reflects the interaction between various links in the flood disaster chain and its impact on critical infrastructure by integrating different data sources and multiple influencing factors, with high real-time and forward-looking, thus providing a scientific basis for disaster prevention, mitigation and emergency response.

[0049] The required data collection sources are diverse, including not only online map data, remote sensing images, but also Internet of Things sensor data, meteorological monitoring data, and historical disaster case data. Multi-source data collection can obtain high-resolution urban images through drone aerial photography, environmental information of large-scale areas through satellite remote sensing, and real-time meteorological, traffic, and infrastructure status data through Internet of Things sensors. In addition, online map data provides basic information such as urban roads and building distributions, and historical disaster case data provides statistical support for the flood disaster chain model. In order to make these data comparable and fusible, technologies such as geocoding and spatial interpolation are used to align the data in time and space to ensure the consistency of data from different sources in time and space. For example, remote sensing images may have a high spatial resolution but a long time update cycle, while sensor data has strong real-time performance but a low resolution. By fusing the two, complementary advantages can be achieved. For example, before a heavy rainstorm in a certain city, sensors deployed on main roads and bridges are used to collect rainfall and water level data in real time, while drones are used to conduct aerial photography of the urban area to obtain the latest building and road information. Combining with satellite image data of historical flood events, unified data processing and spatio-temporal alignment are realized through the cloud platform, providing a data basis for subsequent flood disaster chain analysis. In addition, non-traditional data sources such as social media data and meteorological radar data can be considered. By using natural language processing and image recognition technologies to mine the information of people reporting disasters, the data dimension can be further enriched, and the comprehensiveness and real-time performance of risk assessment can be improved.

[0050] Extracting the spatio-temporal evolution characteristics of flood disaster chains based on multi-source data focuses on extracting the dynamic characteristics of each link of the flood disaster chain from the collected data. By constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain, the interactions among links such as heavy rain, flood, waterlogging, and the resulting traffic jams are quantified. The causal relationship model of the flood disaster chain can adopt statistical methods such as Bayesian networks to describe the triggering thresholds and delay parameters among links. For example, when heavy rain reaches a certain intensity, it may trigger a flood, and the flood may cause waterlogging after a certain period of time. The spatio-temporal evolution matrix is used to describe the diffusion paths and intensity distributions of disaster events in different times and spaces, and dynamic analysis is realized through spatio-temporal clustering and diffusion models. The chain triggering coefficient, cumulative effect index, and delay impact factor are quantitative indicators of the interactions among nodes in the flood disaster chain, which can reflect the chain and cumulative nature of disaster propagation. For example, when a certain area encounters heavy rain, the causal relationship among heavy rain, flood, waterlogging, and traffic jams is modeled through a Bayesian network to quantify the critical value of heavy rain triggering a flood and the delay time of the flood triggering waterlogging, and then the evolution matrix of the flood disaster in the urban area is constructed through a spatio-temporal clustering algorithm to determine the main paths and high-risk areas of disaster propagation. In addition to traditional statistical methods, machine learning algorithms can be introduced to train historical data, update model parameters in real time, and improve the accuracy of flood disaster chain prediction. At the same time, a virtual simulation platform is constructed through digital twin technology to conduct multiple simulation drills for different disaster scenarios, providing decision-making support for disaster prevention and control strategies.

[0051] Based on the prior infrastructure information in the city, by calculating indicators such as building density, road network density, population density, functional type factors, and spatial connectivity, the Key Infrastructure Coefficient (KIC) is constructed. For example, in a certain urban area, the distribution data of buildings and roads are obtained through GIS technology, the building density and road network density of the area are counted using remote sensing image recognition technology, and at the same time, the population density is obtained according to the census data. Combining the functional classification of buildings (such as hospitals, schools, bridges, etc.) and their spatial connectivity, the KIC value of the area is calculated. If the KIC value of an area is high, it indicates that its key infrastructure is relatively vulnerable and may cause more serious losses during a disaster. In addition to the existing indicators, information such as the maintenance status and construction age of the infrastructure can be considered to further expand and refine the KIC formula to form a more accurate infrastructure vulnerability assessment system.

[0052] Coupling the spatio-temporal evolution characteristics with the critical infrastructure coefficient, calculating the regional risk index and generating risk warnings is to couple the spatio-temporal evolution characteristics of the flood disaster chain extracted above with the constructed critical infrastructure coefficient, calculate the regional risk index through the dynamic coupling coefficient and the spatial sensitivity factor, and generate a risk spatial distribution map using the GIS platform. At the same time, combined with real-time data, dynamic warnings are triggered. For example, in an urban area, through the first two steps, various parameters and spatio-temporal evolution matrices of the flood disaster chain are obtained, and at the same time, the KIC value of this area is calculated using infrastructure data. If the cumulative impact of floods caused by heavy rains in a certain area is large, and the infrastructure in this area is relatively vulnerable (high KIC), then the regional risk index R calculated according to the formula will be high. At this time, the GIS platform can mark this area as a high-risk area and push text messages, emails and APP notifications to the emergency management department through the warning system. Methods such as Monte Carlo simulation can also be used to conduct sensitivity analysis on the parameters and weights in the risk index, generate a risk confidence interval, and further improve the robustness and reliability of risk assessment. In addition, combined with real-time updated Internet of Things data and edge computing technology, the dynamic adjustment of the risk index can be realized, forming a closed-loop feedback system to ensure the timeliness and accuracy of warning information.

[0053] The technical solution in this embodiment involves the collection and spatio-temporal alignment of multi-source data, which provides a solid data foundation for the extraction of flood disaster chain characteristics; using algorithms such as Bayesian networks and spatio-temporal clustering to construct a flood disaster chain model, revealing the causal relationships and spatio-temporal evolution laws among various links of the disaster; combining prior information of urban infrastructure to construct critical infrastructure coefficients, effectively evaluating the vulnerability of infrastructure within the region; generating a regional risk index through coupling calculation methods, and using the GIS platform for intuitive display and warning response, thus realizing the intelligence and refinement of disaster prevention, mitigation and emergency management. Integrating various advanced technologies such as big data, intelligent algorithms, GIS technology and the Internet of Things, it can not only realize the dynamic analysis of complex flood disaster chains, but also quantify the risk level according to the actual situation of regional infrastructure, providing a scientific basis for urban safety management and emergency response. At the same time, by introducing dynamic updates, Monte Carlo sensitivity analysis and closed-loop feedback mechanisms, this method has significant advantages in terms of real-time performance, robustness and accuracy. The technical solution in this embodiment extracts the spatio-temporal evolution characteristics of the flood disaster chain through multi-source data collection and spatio-temporal fusion, constructs critical infrastructure coefficients based on information such as buildings, roads, and population, calculates the regional risk index through dynamic coupling, intuitively reflects the risk distribution, realizes accurate warning and emergency response, effectively improves the urban disaster prevention and mitigation level, and significantly enhances the public safety guarantee efficiency.

[0054] In one embodiment, the acquisition module is further configured to: automatically identify and classify buildings, roads, and infrastructure through drone aerial photography and image recognition technologies; perform spatio-temporal alignment and resolution unification on multi-source data using geocoding and spatial interpolation technologies; establish a dynamic data update mechanism based on a cloud platform, and real-time correct meteorological, traffic, and infrastructure status data through edge computing.

[0055] In one embodiment, by obtaining comprehensive, accurate, and real-time raw data, it provides data support for subsequent extraction of flood disaster chain characteristics, construction of critical infrastructure coefficients, and risk assessment.

[0056] Automatically identify and classify buildings, roads, and infrastructure through drone aerial photography and image recognition technologies. Combining an advanced drone platform and image processing algorithms, it realizes the automated detection and classification of infrastructure such as buildings, roads, bridges, and drainage facilities within the target area. Using drone aerial photography can obtain high-resolution image data, with a wide coverage area and the ability to complete the acquisition of a large-scale area in a short time. After being trained, image recognition algorithms (such as convolutional neural networks) can automatically identify different objects in the image and classify them according to predefined categories. By using drones to obtain real-time images, the position information and structural information of targets such as buildings and roads can be automatically extracted through image recognition, which not only improves the efficiency of data acquisition but also ensures the timeliness and accuracy of the data. For example, in a certain urban disaster prevention system, by regularly conducting aerial photography of the urban area using drones and using deep learning algorithms to identify various buildings and transportation hubs, a dynamically updated geographic information database is formed, providing detailed infrastructure distribution data for subsequent analysis. In addition to traditional aerial photography, data from fixed surveillance cameras, satellite remote sensing images, and lidar data can be introduced to further enrich the data sources and improve the accuracy and robustness of target recognition through multi-source data fusion.

[0057] Spatio-temporal alignment and resolution unification of multi-source data are carried out using geocoding and spatial interpolation techniques. After data collection, due to differences in collection time, spatial resolution, and format among different data sources, spatio-temporal alignment and unified processing must be performed. Geocoding technology can convert unstructured data (such as objects in images) into information with specific geographic coordinates, and spatial interpolation technology is used to fill in the missing parts in data collection and achieve spatial consistency among data sources. The purpose of spatio-temporal alignment is to ensure that data from different sources can be compared and fused under the same geographic coordinate system, which is crucial for subsequent construction of the causal model of the flood disaster chain and risk assessment. By unifying the resolution, errors caused by different scales of data can be eliminated, making the analysis results more accurate. For example, using online map data and drone-captured images, after geocoding, all data are converted into standard longitude and latitude coordinates; then, through spatial interpolation technology, high-resolution images are matched with low-resolution sensor data to form a unified spatio-temporal dataset, providing a consistent data basis for the subsequent construction of dynamic models. In addition, in the future, high-precision GPS positioning, differential positioning technology, and real-time satellite navigation data can be combined to further improve the data alignment accuracy. Even the introduction of deep learning models for intelligent prediction of data missing areas and optimization of data completion effects can be considered.

[0058] A dynamic data update mechanism is established based on the cloud platform, and meteorological, traffic, and infrastructure status data are corrected in real time through edge computing to achieve dynamic update and real-time processing of data. The cloud platform is used as the core for centralized management and processing of data, aggregating, storing, and uniformly processing the data uploaded by each collection terminal. At the same time, edge computing technology is used for preliminary data processing and correction at locations closer to the data collection source, thereby improving data transmission efficiency and response speed. The cloud platform can support the storage and calculation of large-scale data and achieve efficient integration of data through a distributed architecture. Edge computing distributes computing tasks near the data collection terminal, reducing data transmission latency and being able to quickly respond in emergencies to achieve real-time correction and update of meteorological, traffic, and infrastructure status data. For example, in the disaster warning system of a certain city, a cloud platform center and edge computing nodes distributed in different regions are set up to collect meteorological station and traffic sensor data in real time. The edge nodes can preliminarily screen and correct abnormal data and upload it to the cloud platform after unified processing to ensure the timeliness and accuracy of the data, thus providing a decision-making basis for emergency response in a timely manner. Artificial intelligence algorithms can also be introduced to perform real-time prediction and anomaly detection on edge data to improve the intelligent level of data preprocessing. In addition, a cross-platform data exchange standard can be constructed to achieve data interconnection and interoperability among different departments and institutions.

[0059] The data collection and spatio-temporal alignment and fusion processing link provides a solid data foundation for the entire flood disaster chain risk identification system. Through the comprehensive collection, precise alignment, and real-time update of multi-source data, the accuracy and timeliness of subsequent flood disaster chain model construction and risk assessment are ensured. It not only covers traditional data collection technologies but also comprehensively applies advanced technologies such as UAV aerial photography, image recognition, geocoding, spatial interpolation, cloud platforms, and edge computing, fully demonstrating the application advantages of modern information technology in the field of disaster prevention and mitigation. By continuously expanding data sources and optimizing data processing algorithms, it is expected to achieve a more intelligent, automated, and refined urban risk monitoring and emergency management system in the future.

[0060] In one embodiment, the extraction module is further configured to: model the causal relationships among heavy rain, flood, waterlogging, and traffic congestion through a Bayesian network, and quantify the triggering threshold and delay parameters; analyze the propagation paths of flood disaster chain events using spatio-temporal clustering algorithms and diffusion models to construct a spatio-temporal evolution matrix; dynamically correct the flood disaster chain parameters through chain triggering coefficients, cumulative effect indices, and delay impact factors.

[0061] In one embodiment, a Bayesian network is used to model the causal relationships among heavy rain, flood, waterlogging, and traffic congestion. A Bayesian network is a probabilistic graphical model that expresses the conditional probability dependence relationships among variables through a directed acyclic graph (DAG). Key disaster events (such as heavy rain, flood, waterlogging, and traffic congestion) serve as network nodes, and the edges between the nodes reflect the causal associations between the events. For example, when heavy rain reaches a certain intensity (triggering threshold), it will trigger a flood with a relatively high probability; after a flood occurs, waterlogging may occur within a certain time delay, and waterlogging may in turn cause traffic congestion. Through statistical analysis of historical data, a Bayesian network can be used to quantify the triggering thresholds and delay parameters between nodes, thereby clarifying the causal transmission mechanism between events. For example, assume that in a certain city, historical data shows that when the rainfall in a certain area reaches 50 mm, the probability of a flood occurring increases significantly. The Bayesian network can use "rainfall ≥ 50 mm" as the critical condition for heavy rain to trigger a flood, and at the same time set the average delay from heavy rain to flood to 15 minutes; after a flood occurs, the triggering probability and delay of waterlogging can also be quantified accordingly through data, such as the probability of waterlogging occurring is 70%, and the delay is about 20 minutes. In this way, through the Bayesian network, a hierarchical flood disaster chain model can be constructed to achieve a quantitative description of the causal relationships. In practical applications, the parameters of the Bayesian network model can be continuously updated through machine learning methods, and the model can be corrected online by combining real-time monitoring data to ensure that the model parameters can reflect the latest disaster characteristics. In addition, other influencing factors (such as terrain, soil quality, and drainage system conditions) can be introduced into the model to improve the accuracy and adaptability of causal relationship modeling.

[0062] After constructing the causal model of the flood disaster chain, the spatio-temporal clustering algorithm and diffusion model are further used to analyze the propagation path of disaster events and construct a spatio-temporal evolution matrix. The spatio-temporal clustering algorithm can classify disaster events that are adjacent or similar in time and space, revealing the spatial distribution and evolution characteristics of event groups, while the diffusion model simulates the diffusion process of disaster effects from the center to surrounding areas. By combining the two, the propagation law and evolution trend of the flood disaster chain at different time nodes and different spatial regions can be intuitively displayed. Spatio-temporal clustering mainly relies on the geographical coordinates and timestamps of event occurrences to gather events with high spatial proximity and time correlation together, forming high-risk areas or disaster hotspots; the diffusion model is based on physical or mathematical models to describe how disaster factors spread outward from the initial point, such as the rising flood level gradually affecting surrounding low-lying areas. After constructing the spatio-temporal evolution matrix, the distribution, diffusion direction, and speed of disaster events in each time period can be quantitatively analyzed, providing parameter support in the spatial and temporal dimensions for the subsequent calculation of risk indices. For example, after a heavy rain in a certain city, the flood and waterlogging events that occurred within several consecutive hours are clustered and analyzed through the spatio-temporal clustering algorithm, and it is determined that the city center area and some low-lying areas are the main high-risk areas. The diffusion model is used to simulate the change of the flood level, determine the speed and scope of the flood spreading from the city center to surrounding areas, and then construct a spatio-temporal evolution matrix reflecting the disaster propagation path and evolution process. This matrix not only records the occurrence order of disaster events but also includes the change trend of spatial distribution, providing detailed spatio-temporal dynamic information for risk assessment. Based on the existing methods, deep learning methods can be introduced to extract features and recognize patterns from spatio-temporal data, further improving the accuracy of spatio-temporal clustering and diffusion models. At the same time, combined with digital twin technology, the virtual simulation is fused with real-time monitoring data to update the spatio-temporal evolution matrix in real time, enhancing the dynamics and accuracy of disaster propagation path prediction.

[0063] After obtaining the causal relationship and spatio-temporal evolution characteristics of the flood disaster chain, the scheme dynamically modifies the flood disaster chain model by introducing the chain trigger coefficient, cumulative effect index, and delay impact factor. The chain trigger coefficient is used to quantify the intensity of the trigger effect between each link in the flood disaster chain. The cumulative effect index reflects the cumulative effect of each link, while the delay impact factor describes the impact brought by the time delay between each link. Through these three parameters, the initially constructed flood disaster chain model can be corrected in real time to adapt to the dynamically changing disaster scenario and ensure that the risk assessment results can truly reflect the current actual situation. Specifically, the chain trigger coefficient reflects the degree of influence of a certain disaster link on the subsequent link; the cumulative effect index is to superimpose the effects of multiple consecutive links to form an overall risk increase effect; the delay impact factor takes into account the impact of the time delay between each link on risk transmission. Using these parameters, the risk contribution of each node in the flood disaster chain can be dynamically adjusted, thus realizing the adaptive optimization of the model. For example, if the real-time monitoring data shows that the trigger effect of a certain link has increased, the chain trigger coefficient can be correspondingly increased to more quickly reflect the risk upgrade. For example, during a rainstorm process, the real-time data shows that the flood propagation speed is faster than expected by historical data. At this time, the chain trigger coefficient can be appropriately increased so that the model pays more attention to the impact between the rainstorm and the flood when calculating the risk. At the same time, if it is found that the flood duration is extended, the cumulative effect index and the delay impact factor also need to be adjusted accordingly, so that the finally calculated regional risk index is closer to the actual situation. Through the above dynamic correction mechanism, the system can continuously self-correct during the disaster occurrence process, improving the accuracy of risk prediction and early warning. To further improve the intelligence level of the dynamic correction mechanism, online learning algorithms and feedback control mechanisms can be introduced to optimize the chain trigger coefficient, cumulative effect index, and delay impact factor in real time. Using Monte Carlo simulation or reinforcement learning techniques, comprehensive analysis of historical data and real-time data is carried out to automatically optimize the values of each parameter, enabling the entire flood disaster chain model to adapt to the risk transmission laws under different scenarios. In addition, the corrected parameters can be fed back to the digital twin platform to realize the closed-loop update of virtual simulation and real-time monitoring data, further enhancing the system's response ability and early warning efficiency to the dynamic evolution of the flood disaster chain.

[0064] In this embodiment, the technical solution models the causal relationships among rainstorms, floods, waterlogging, and traffic jams through a Bayesian network, achieving the quantification of the triggering thresholds and delay parameters of the flood disaster chain. By constructing a spatio-temporal evolution matrix using a spatio-temporal clustering and diffusion model, the propagation paths and dynamic distribution characteristics of disaster events are effectively revealed. Finally, the flood disaster chain parameters are dynamically corrected using the chain triggering coefficient, cumulative effect index, and delay impact factor, enabling the entire model to flexibly respond to different disaster scenarios and enhancing the real-time performance and accuracy of risk assessment. Through the organic combination of these technical means, this solution not only provides a scientific basis for urban flood disaster risk prediction but also offers technical support for disaster warning and emergency response, having high application value and promotion prospects.

[0065] In one embodiment, the construction module is further configured to: identify the structural types and functional attributes of buildings, bridges, and drainage facilities based on a deep learning algorithm; calculate the spatial distribution indicators of building density, road density, and population density; calculate the critical infrastructure coefficient through the following formula:

[0066] KIC = α×BD + β×RD + γ×PD + δ×FT + ε×SI

[0067] where the symbol KIC is the critical infrastructure coefficient, the symbol BD is the building density, the symbol RD is the road network density, the symbol PD is the population density, the symbol FT is the functional type factor, the symbol SI is the spatial connectivity, and the weights α, β, γ, δ, ε are automatically calibrated through a machine learning model and historical disaster data.

[0068] In one embodiment, deep learning algorithms are used to automatically identify infrastructure such as buildings, bridges, and drainage facilities in urban areas, and extract their structural types and functional attributes. Traditional manual identification or rule extraction methods are not only inefficient but also poorly adaptable to complex scenarios. Deep learning algorithms (such as convolutional neural networks, YOLO, Faster R-CNN, etc.) can automatically learn target features with the support of a large amount of labeled data to achieve high-precision identification. Using deep learning for image recognition mainly includes three steps: data preprocessing, model training, and target detection. First, high-resolution images are obtained through drone aerial photography or satellite remote sensing, and then the images are processed such as normalization and data augmentation. Next, a large number of labeled samples are used to train the network model so that the model can distinguish different categories such as buildings, bridges, and drainage facilities. Finally, the trained model is applied to actual data to achieve automatic detection and classification, and extract the structural type (e.g., reinforced concrete structure, brick-wood structure, etc.) and functional attribute (such as residential, commercial, public facilities, or transportation hub) of each facility. For example, in a certain urban disaster prevention system, after the urban area images obtained by drones are input into the deep learning model, the system can automatically identify multiple buildings, several main bridges, and the urban drainage network, and automatically label the uses of the buildings (such as residential buildings, hospitals, schools) and the structural types of the bridges. This process not only greatly improves the efficiency of data collection but also provides accurate prior information for subsequent infrastructure vulnerability assessment.

[0069] After completing the automatic identification of infrastructure, the next step is to calculate spatial distribution indicators such as building density, road network density, and population density in the area. These indicators can intuitively reflect the concentration of infrastructure and the aggregation of the population in the area, and are important parameters for evaluating the vulnerability of the area. The key infrastructure coefficient is calculated by the following formula:

[0070] KIC = α × BD + β × RD + γ × PD + δ × FT + ε × SI

[0071] Among them, the symbol KIC is the critical infrastructure coefficient, which reflects the comprehensive vulnerability of the infrastructure within a region. The symbol BD is the building density, the symbol RD is the road network density, the symbol PD is the population density, and the symbol FT is the functional type factor, which reflects the disaster resistance ability and criticality of different buildings or facilities in disasters (for example, hospitals and fire stations are usually given higher weights). The symbol SI is the spatial connectivity, which describes the degree of spatial connection and dependence among various facilities. The weights α, β, γ, δ, and ε are automatically calibrated through a machine learning model combined with historical disaster data to achieve the adaptive optimization of the index weights. The above spatial indicators are usually calculated with the help of a GIS system. By means of methods such as spatial overlay and kernel density analysis of infrastructure and population data, specific values within each region are obtained, providing basic data for the construction of the subsequent critical infrastructure coefficient (KIC). For example, in a certain urban area, through remote sensing images and urban planning data, the building density of a certain urban area can be calculated as 1,500 buildings per square kilometer, the main road density is 30 kilometers per square kilometer, and the population density is 8,000 people per square kilometer. When combined with the urban functional planning, these data can help determine the key protection objects and emergency resource allocation within the region. For another example, suppose the specific data of a certain urban area are: building density BD = 150, road density RD = 30, population density PD = 800, functional type factor FT = 0.7, and spatial connectivity SI = 0.5; at the same time, the weights determined through machine learning are α = 0.2, β = 1.5, γ = 0.25, δ = 3, and ε = 8 respectively. Then the calculated critical infrastructure coefficient is:

[0072] KIC = 0.2×150 + 1.5×30 + 0.25×800 + 3×0.7 + 8×0.5 = 281.1

[0073] Through the above method, a value of 281.1 can be calculated. The higher the value, the more vulnerable the infrastructure in the region and the greater the risk. This indicator can be used to divide high-risk areas and guide the allocation and priority setting of disaster prevention resources.

[0074] The technical solution in this embodiment automatically identifies infrastructure information through deep learning, combines GIS to calculate spatial indicators such as building, road, and population densities, and then constructs the critical infrastructure coefficient through a weighted formula, realizing the quantitative assessment of the vulnerability of the infrastructure within the region. It not only improves the data collection and processing efficiency but also ensures the accuracy and dynamics of risk assessment by automatically calibrating the weights through machine learning. This provides a solid data support and theoretical basis for subsequent disaster risk assessment, early warning, and emergency response.

[0075] In one embodiment, the coupling module is further configured to:

[0076] Dynamically adjust the parameter weights w of the flood disaster chain by combining fuzzy comprehensive evaluation and entropy weight method j ; Calculate the regional risk index based on the following formula:

[0077]

[0078] Among them, the symbol R represents the regional risk index, and the symbol C t is the cumulative impact coefficient of the flood disaster chain, and the symbol P j is the parameter of each link in the flood disaster chain, and the symbol S i is the spatial sensitivity factor; conduct sensitivity analysis on the weights and parameters through Monte Carlo simulation to generate a risk confidence interval.

[0079] In one embodiment, comprehensively evaluate the regional risk index by utilizing the parameters of the flood disaster chain, the dynamic coupling coefficient, and the spatial sensitivity factor. Dynamically adjust the parameter weights wj of the flood disaster chain by fuzzy comprehensive evaluation and entropy weight method, which can introduce expert experience or uncertainty factors into the evaluation process, and conduct a comprehensive evaluation combining qualitative and quantitative methods on the impact of each disaster parameter through the membership function. The entropy weight method determines the weights of each index based on the dispersion degree of the data itself, which can avoid the deviation caused by completely relying on subjective experience. By combining the two, the weight values wj of each parameter (such as rainfall intensity, flood level, waterlogging depth, etc.) in the flood disaster chain can be dynamically corrected to make the evaluation results more objective and accurate. Calculate the regional risk index based on the following formula:

[0080]

[0081] Among them, Ct is the cumulative impact coefficient of the flood disaster chain, representing the chain effect and superposition effect among multiple links such as rainstorm, flood, and waterlogging, and P j is the j-th disaster parameter, w j is the weight of this parameter, and S iLet \(S\) be the spatial sensitivity factor, which reflects environmental conditions such as terrain, soil quality, and the integrity of the drainage system. \(K_{IC}\) is the key infrastructure coefficient, which quantifies the vulnerability of infrastructure such as buildings, roads, and population density. This formula couples the characteristics of the flood disaster chain with the attributes of key infrastructure to obtain a comprehensive risk index \(R\). When there are uncertainties or fluctuations in parameter weights and disaster parameters, Monte Carlo simulation examines the distribution of the risk index through a large number of random samplings and can obtain the risk confidence interval. This process can quantify the reliability and stability of the evaluation results and provide a more flexible risk range judgment for decision-makers. For example, in a coastal city that is vulnerable to typhoon-induced heavy rains, which can lead to successive occurrences of flood disaster chain links such as floods, waterlogging, and traffic jams. Through fuzzy comprehensive evaluation, the expert group assigns a high membership degree to "rainstorm intensity" and "waterlogging depth", indicating that these two factors have a significant impact on the risk. At the same time, the entropy weight method is introduced to analyze the dispersion of historical data, and it is obtained that the weight of "flood water level" is slightly lower than that of "rainstorm intensity" because the drainage system in this area is relatively complete. Combining the above weights and parameters, the cumulative impact coefficient \(C_t\) of the flood disaster chain is calculated. Suppose the result shows that the triggering effect of rainstorm intensity and waterlogging depth on flood disasters is the most obvious, so the cumulative impact coefficient \(C_t\) of the flood disaster chain accounts for the main proportion. At the same time, the high population density and large road network density in the city center lead to a relatively high key infrastructure coefficient \(K_{IC}\); the flat terrain causes poor drainage, and the spatial sensitivity factor is also relatively high. A relatively high risk index \(R\) is obtained through formula calculation. The system can immediately issue a key warning for the city center area and generate a risk confidence interval in combination with Monte Carlo simulation, showing that in the most unfavorable scenario, the risk index may further increase, and it is necessary to reserve emergency supplies and evacuate residents in advance.

[0082] This technical solution combines fuzzy comprehensive evaluation, entropy weight method, Monte Carlo simulation, and dynamic coupling formula, which can not only accurately describe the relative importance of each parameter of the flood disaster chain, but also quantify the superimposed impact of infrastructure vulnerability and environmental sensitivity on risk, and improve the credibility of the evaluation results through sensitivity analysis, providing an efficient, flexible, and scalable technical means for urban disaster prevention, mitigation, and emergency management.

[0083] In one embodiment, the warning module is further configured to: map the risk index into a heat map through a GIS platform, and superimpose the infrastructure distribution and the evolution path of the flood disaster chain; set a risk threshold, and trigger a warning message in combination with the real-time data of the Internet of Things sensors; send a hierarchical warning instruction to the emergency management department through SMS, email, and APP push methods.

[0084] In one embodiment, the regional risk index is visualized as a heat map through a GIS platform. This heat map not only intuitively shows the spatial distribution of high-risk areas and low-risk areas, but also can overlay the spatial-temporal evolution paths of the locations of urban key infrastructure and flood disaster chains, thereby helping decision-makers quickly locate potential danger points. This process relies on the previous risk index calculation results, maps the risk values corresponding to each grid or administrative unit into a color gradient, and presents the disaster propagation trend and infrastructure distribution through GIS overlay analysis to form a complete urban disaster risk situation map. After obtaining the risk heat map, the system needs to set a risk threshold and combine the real-time data of Internet of Things sensors to trigger dynamic early warnings. When the risk index of any area exceeds the preset critical value, or the sensor detects abnormal situations (such as a sharp rise in water level or a sudden change in flow rate), the system will automatically determine that the area is in a high-risk state and immediately issue an alarm signal. In this way, the triggering of early warnings depends on both the existing risk assessment results and the real-time environmental changes, ensuring that the alarm information can quickly reflect the latest disaster situation. The hierarchical early warning instructions are sent to the emergency management department through text messages, emails and APP push methods to achieve multi-channel and fast transmission of early warning information. Hierarchical early warnings are usually divided into multiple levels such as blue, yellow, orange and red, corresponding to different disaster severity levels. After receiving early warnings of different levels, the emergency management department can quickly formulate response strategies, such as dispatching emergency rescue forces, organizing personnel evacuation, initiating traffic control or allocating emergency supplies, etc., to minimize disaster losses. For example, taking a coastal city as an example, when the city monitors that the water level along the coast exceeds the limit and the rainstorm intensity exceeds the set threshold, the system marks the key areas in the heat map as red high-risk and automatically pushes red early warning information to the city emergency management bureau and the transportation department, prompting to immediately close the low-lying sections and organize the evacuation of residents. This real-time linkage mechanism effectively improves the disaster prevention and mitigation efficiency of the city in extreme weather.

[0085] In one embodiment, it further includes an emergency module, which is used to receive the risk spatial distribution map and hierarchical early warning instructions generated by the early warning module, determine the key protection areas according to the regional risk index; generate a risk conduction topology map based on the dynamic coupling coefficient, deploy precise on-site duty plans for transportation hubs, power facilities, and communication nodes, and configure emergency supplies and mobile drainage equipment; real-time monitor the water level, structural deformation and equipment operation data of the on-site area through Internet of Things sensors, and use the mobile command terminal to feedback the on-site situation information.

[0086] In one embodiment, the risk transmission topology map is generated and key areas are identified based on a risk spatial distribution map (GIS heat map) generated by the early warning module, which includes a regional risk index (R value) and a risk confidence interval (e.g., the boundaries of high-risk areas with a 90% probability). Graded early warning instructions specify the response levels corresponding to different risk levels (e.g., a yellow warning requires increased inspections, a red warning requires traffic control). The dynamic coupling coefficient, output by the coupling module, is applied to reflect the transmission strength of each link in the flood disaster chain to infrastructure. For example, a flood coupling coefficient of 0.7 in a certain region indicates a high contribution of flood damage to infrastructure in that region. Using a graph theory model, infrastructure nodes (e.g., substation A, bridge B) are treated as vertices, and disaster transmission paths (e.g., flood spread along the river to bridge B) are treated as edges. The edge weights are determined by the dynamic coupling coefficient. A shortest path algorithm (e.g., Dijkstra) is used to identify key nodes that are likely to be impacted by disasters first, thus forming a risk transmission topology map. For example, the eastern part of a city was designated a red high-risk zone due to its high KIC value (outdated drainage and dense population). The risk transmission topology map showed that flooding could spread to three substations via the river, drainage network, and roads. Based on this, the system demarcated a key protection zone with a radius of 500 meters centered on the substations and marked the road intersections where drainage equipment should be deployed first.

[0087] In terms of precise on-site duty and emergency resource allocation, transportation hubs deploy patrol personnel and pre-position mobile pump trucks based on the vulnerability score of the road network in the topology map (for example, the KIC value of a tunnel is >0.9); power facilities are equipped with waterproof baffles and backup generators to address the risk of waterlogging around substations, and power repair teams are arranged to be on standby 24 hours a day; communication nodes set up temporary communication vehicles at the junction of base station blind spots and high-risk areas to ensure smooth transmission of emergency instructions. Material dispatch optimization is based on path planning algorithms (such as algorithm A) combined with real-time traffic data (from IoT road condition sensors) to calculate the optimal material transportation route. For example, the time to transport drainage equipment from the warehouse to the target station was shortened from 40 minutes to 25 minutes.

[0088] For example, the system detected that a subway station was listed as a first-level protection target due to a sudden increase in the KIC value in the surrounding area (excessive building density leading to increased drainage pressure), and automatically triggered the following responses, including: dispatching two mobile drainage vehicles to the subway entrance and pre-positioning 200 sandbags; arranging three groups of inspectors to monitor the structural deformation of the station in shifts (by deploying laser ranging sensors); and coordinating with the transportation department to adjust the timing of traffic lights around the subway to reserve a green wave belt for emergency vehicles to pass through.

[0089] In terms of the Internet of Things perception and dynamic feedback mechanism, pressure water level gauges (accuracy ±1 cm) are installed in the stagnant areas for water level monitoring, and data is uploaded to the edge computing node every 5 minutes; for structural safety monitoring, inclinometers (range ±15°) and crack gauges (resolution 0.01 mm) are used to detect the deformation of bridges and buildings in real time; for equipment status feedback, an Internet of Things module is installed on the drainage vehicle to transmit the working status (such as pump flow rate and remaining fuel). Patrol personnel take flood photos with a tablet computer, and AI automatically identifies the water depth and marks it on the GIS map; the terminal receives real-time risk alerts pushed by the system (such as "the rainfall in this area will reach 50 mm in the next 30 minutes"), and at the same time reports the consumption of materials (such as 80% of the sandbags have been used).

[0090] For example, the edge computing node conducts local analysis on the data of the stagnant area sensors (such as removing abnormal water level fluctuation noise); feeds the cleaned data back to the system coupling module to trigger the incremental update of the spatio-temporal evolution matrix (such as correcting the flood diffusion speed parameter); according to the latest risk index, automatically adjusts the allocation of emergency resources through the multi-department collaborative platform (such as temporarily redirecting the drainage vehicle originally planned to support Area A to Area B with higher risk).

[0091] The technical solution in this embodiment quantifies the disaster propagation path through the risk conduction topology map, avoiding the "one-size-fits-all" resource allocation; the Internet of Things perception network realizes minute-level data update, with an efficiency improvement of more than 3 times compared to traditional manual inspections; the on-site data feedback drives the dynamic correction of model parameters, forming a virtuous cycle of "monitoring - early warning - response - re-optimization". Deeply integrating the theoretical model with engineering practice provides a practical technical paradigm for infrastructure protection in the scenario of flood disaster chains.

[0092] In one embodiment, a scheduling module is further included, which is used to adjust the deployment density of the stationed personnel and the material scheduling path through the multi-department collaborative platform according to the dynamic correction result of the spatio-temporal evolution matrix of the flood disaster chain; when the risk confidence interval exceeds the preset threshold, trigger a chain emergency response of traffic control - medical rescue - power supply, and synchronously update the spatial connectivity parameter in the infrastructure coefficient; conduct local model training on the monitoring data of the stationed points through the edge computing node, and dynamically optimize the weight allocation strategy of the key infrastructure coefficient.

[0093] In one embodiment, the dynamic propagation characteristics of the flood disaster chain are combined with the static vulnerability of infrastructure to achieve precision and intelligence in emergency response through data-driven closed-loop control, forming a closed-loop control mechanism of "data-driven decision-making - decision feedback model". A flood disaster chain emergency response system is constructed around "dynamic correction - collaborative response - closed-loop optimization". Based on real-time monitoring data, the spatio-temporal evolution matrix of the flood disaster chain is corrected to drive the dynamic adjustment of multi-departmental resource scheduling strategies; the emergency response trigger conditions are quantified through risk confidence intervals to achieve the automated execution of cross-departmental chain commands; the local data of the stationed points are used to train the model to continuously optimize the critical infrastructure assessment system.

[0094] The dynamic correction and collaborative scheduling of the spatio-temporal evolution matrix of the flood disaster chain include a matrix correction mechanism and multi-departmental collaborative scheduling. The matrix correction mechanism means that the spatio-temporal evolution matrix is essentially a multi-dimensional tensor, and its dimensions include time slices (such as every hour), spatial grids (such as 100m×100m grids), and disaster parameters (flood depth, waterlogging range, etc.). When the Internet of Things sensors detect that the water level rising rate in a certain grid exceeds the model prediction value (such as the actual value of 20cm / h vs the predicted value of 15cm / h), the matrix correction process is triggered. The deviation degree between the actual value and the predicted value is calculated through the difference algorithm; the Kalman filter is used to re-estimate the future disaster parameters of this grid in the matrix; the diffusion coefficient of adjacent grids is updated (such as adjusting the weight of the flood spreading eastward from 0.6 to 0.8). Multi-departmental collaborative scheduling means that according to the distribution of high-risk grids in the corrected matrix, the responsibility areas are divided through the Voronoi diagram algorithm. For example: if the risk index of a certain grid is 0.9 and the area is 1km2, and a single person can cover 0.3km2 of the high-risk area, then 3 people need to be deployed. Optimization of the material path: Combining real-time road conditions (from traffic cameras and GPS data), the tabu search algorithm is used to generate the shortest transportation path avoiding flooded sections, and transportation vehicles are dynamically allocated (such as preferentially using high-chassis trucks to transport sandbags to deep water areas). For example, during a certain rainstorm, the bridge sensor detected that the inclination angle of the bridge pier reached 0.5° (exceeding the safety threshold of 0.3°), the spatio-temporal evolution matrix was corrected, and the disaster intensity coefficient of the grids within 200 meters around the bridge was increased by 30%; the collaborative platform automatically dispatched 2 nearby engineering teams to carry support equipment and re-planned the material transportation route to bypass the landslide section; the emergency plan library was updated, and this bridge was marked as a key object for future disaster drills.

[0095] Threshold triggering and parameter linkage of chain emergency response include confidence interval threshold design and spatial connectivity parameter update. Confidence interval threshold design means generating the probability distribution of risk index through Monte Carlo simulation and setting double thresholds. Yellow warning (60% probability of exceeding the benchmark risk): Trigger the enhanced inspection instruction; Red warning (90% probability of exceeding the benchmark risk): Initiate cross-departmental chain response. For example, when the upper limit of the risk confidence interval in a hospital area breaks through the red threshold, traffic control is automatically executed (set up an electronic fence around the surrounding roads and only allow rescue vehicles to pass), medical rescue (dispatch ambulances to standby within 3 kilometers of the hospital radius), and power supply guarantee (start the standby generator and cut off non-essential power supply lines). Spatial connectivity parameter update means that the spatial connectivity (SI) in the original infrastructure coefficient is mostly calculated based on static road network data and needs to be dynamically corrected during emergency response. For example, if the main road between a substation and a hospital is flooded, resulting in a 70% road interruption rate, the SI value drops to 30% of the original value, directly affecting subsequent risk calculations. Another example, when the confidence interval of the waterlogging risk in the central business district of a city reaches the red threshold due to a burst drainage pipe, the intelligent signal light system is linked, and 8 intersections are set to the "only out and no entry" mode; remotely turn off the charging piles in the underground garage of the business district to prevent electric leakage accidents; synchronously update the SI parameters in this area (the connectivity between the business district and the fire station decreases by 50% due to road waterlogging), so that the weight of this area in the subsequent risk model increases.

[0096] Edge computing-driven dynamic optimization of coefficient weights includes local model training and weight allocation strategy. Deploy edge computing nodes at the garrison points to store the monitoring data (water level, deformation, etc.) in this area for the past 3 months, and use the federated learning framework for local model training. Initialize the global model (the benchmark KIC weights α = 0.2, β = 0.3, etc. issued by the cloud); calculate the loss function based on local data (such as the mean square error between the predicted risk value and the actual loss value); update the weight parameters through the gradient descent method to generate a local model; encrypt and upload the model incremental parameters to the cloud for aggregation. For example, in an industrial park, due to frequent equipment failures, the actual loss is higher than the predicted value, and the edge node automatically increases the weight of the function type factor (δ) (from 0.15 → 0.22). The weight allocation strategy means introducing the idea of game theory and transforming the optimization of infrastructure coefficient weights into a multi-objective programming problem. Minimize the prediction error of historical disaster losses. Maximize the risk discrimination of different facility types, constraint condition: Σ(α + β + γ + δ + ε) = 1. Use the NSGA-II multi-objective optimization algorithm to solve the Pareto optimal solution set, and finally select the implementation plan by the expert system. For example, in a chemical industrial park, the edge node detects multiple small-scale leaks without triggering a system warning, and the original model underestimates the impact of spatial connectivity (ε) on chain accidents; after increasing the weight of ε from 0.1 to 0.18, the model successfully predicts the risk of a chain shutdown in the factory area that may be caused by the next pipeline leak; the updated weight parameters are synchronized to other chemical industrial park nodes after being reviewed by the cloud.

[0097] Through matrix correction and parameter linkage, the minute-level synchronization between the disaster model and the infrastructure status is achieved, and the response speed is increased by 80% compared with the traditional static model; the departmental data barrier is broken through, and an automated pipeline of "risk determination - instruction generation - execution feedback" is constructed to shorten the emergency decision-making chain; edge intelligence training enables the infrastructure assessment system to have the ability to adapt to regional differences, and solves the problem of resource misallocation caused by the "one-size-fits-all" assessment.

[0098] As Figure 2 shown, the present invention provides an electronic device 200, and the electronic device includes: a communication interface, a processor 201, and a memory 202;

[0099] Wherein, the memory 202 is used to store program instructions, and when the program instructions are executed by the processor 201 communicatively connected to the memory 202 through the communication interface, multi-source data is collected and subjected to spatio-temporal alignment and fusion processing, and the multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data; based on the multi-source data, spatio-temporal evolution characteristics of the flood disaster chain are extracted, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain; according to prior infrastructure information, a critical infrastructure coefficient is constructed, and the critical infrastructure coefficient is based on any one or more of building density, road network density, population density, functional type factor, and spatial connectivity; coupling the spatio-temporal evolution characteristics and the critical infrastructure coefficient, calculating a regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor, generating a risk spatial distribution map and triggering a dynamic warning.

[0100] The present invention provides a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, multi-source data is collected and subjected to spatio-temporal alignment and fusion processing, and the multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data; based on the multi-source data, spatio-temporal evolution characteristics of the flood disaster chain are extracted, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain; according to prior infrastructure information, a critical infrastructure coefficient is constructed, and the critical infrastructure coefficient is based on any one or more of building density, road network density, population density, functional type factor, and spatial connectivity; coupling the spatio-temporal evolution characteristics and the critical infrastructure coefficient, calculating a regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor, generating a risk spatial distribution map and triggering a dynamic warning.

[0101] It should be understood that the specific features, operations, and details described above regarding the method of the present invention can be similarly applied to the apparatus and system of the present invention, or vice versa. Additionally, each step of the method of the present invention described above can be executed by the corresponding components or units of the apparatus or system of the present invention.

[0102] It should be understood that the various modules / units of the apparatus of the present invention can be implemented in whole or in part by software, hardware, firmware, or a combination thereof. Each module / unit can be embedded in the processor of a computer device in the form of hardware or firmware or independent of the processor, or can be stored in the memory of the computer device in the form of software for the processor to call to execute the operations of each module / unit. Each module / unit can be implemented as an independent component or module, or two or more modules / units can be implemented as a single component or module.

[0103] In one embodiment, a computer device is provided, which includes a memory and a processor. A computer instruction executable by the processor is stored on the memory. When the computer instruction is executed by the processor, it instructs the processor to execute each step of the method of the embodiment of the present invention. The computer device can generally be a server, a terminal, or any other electronic device with necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, a memory, a network interface, a communication interface, etc. connected through a system bus. The processor of the computer device can be used to provide necessary computing, processing, and / or control capabilities. The memory of the computer device can include a non-volatile storage medium and an internal memory. An operating system, computer programs, etc. can be stored in or on the non-volatile storage medium. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the computer device can be used to connect and communicate with external devices through a network. When the computer program is executed by the processor, it executes the steps of the method of the present invention.

[0104] The present invention can be implemented as a computer-readable storage medium having stored thereon a computer program, which when executed by a processor causes the steps of the method of the embodiments of the present invention to be performed. In one embodiment, the computer program is distributed over a plurality of network-coupled computer devices or processors such that the computer program is stored, accessed, and executed by one or more computer devices or processors in a distributed manner. A single method step / operation, or two or more method steps / operations, can be performed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations can be performed by one or more computer devices or processors, and one or more other method steps / operations can be performed by one or more other computer devices or processors. One or more computer devices or processors can perform a single method step / operation, or perform two or more method steps / operations.

[0105] Those of ordinary skill in the art will appreciate that the method steps of the present invention can be completed by a computer program instructing relevant hardware such as a computer device or a processor. The computer program can be stored in a non-transitory computer-readable storage medium, and when the computer program is executed, the steps of the present invention are caused to be performed. Depending on the circumstances, any reference herein to a memory, storage, database, or other medium may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0106] The various technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features have been described, any combination of these technical features should be considered to be covered by this specification, provided that such a combination is not contradictory.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A risk identification system for critical infrastructure in the scenario of a flood disaster chain, characterized in that, Including: A collection module, configured to collect multi-source data and perform spatio-temporal alignment and fusion processing, where the multi-source data includes any one or more of online map data, remote sensing images, Internet of Things sensor data, meteorological monitoring data, and historical disaster case data; An extraction module, configured to extract spatio-temporal evolution characteristics of a flood disaster chain based on the multi-source data, including constructing a causal relationship model and a spatio-temporal evolution matrix of the flood disaster chain; A construction module, configured to construct a critical infrastructure coefficient according to prior infrastructure information, where the critical infrastructure coefficient is based on any one or more of building density, road network density, population density, functional type factor, and spatial connectivity; A coupling module, configured to couple the spatio-temporal evolution characteristics and the critical infrastructure coefficient, and calculate a regional risk index through a dynamic coupling coefficient and a spatial sensitivity factor; An early warning module, configured to generate a risk spatial distribution map and trigger a dynamic early warning.

2. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 1, wherein The collection module is further configured to: automatically identify and classify buildings, roads, and infrastructure through drone aerial photography and image recognition technology; perform spatio-temporal alignment and resolution unification on multi-source data by using geocoding and spatial interpolation technology; establish a dynamic data update mechanism based on a cloud platform, and real-time correct meteorological, traffic, and infrastructure status data through edge computing.

3. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 1, characterized in that The extraction module is further configured to: model the causal relationship between rainstorms, floods, waterlogging, and traffic jams through a Bayesian network, and quantify the triggering threshold and delay parameters; analyze the propagation path of flood disaster chain events by using a spatio-temporal clustering algorithm and a diffusion model, and construct a spatio-temporal evolution matrix; dynamically correct flood disaster chain parameters through a chain triggering coefficient, a cumulative effect index, and a delay impact factor.

4. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 1, wherein The construction module is further configured to: identify the structural type and functional attributes of buildings, bridges, and drainage facilities based on a deep learning algorithm; calculate the spatial distribution indexes of building density, road density, and population density; calculate the critical infrastructure coefficient through the following formula: KIC = α×BD + β×RD + γ×PD + δ×FT + ε×SI Where, the symbol KIC is the critical infrastructure coefficient, the symbol BD is the building density, the symbol RD is the road network density, the symbol PD is the population density, the symbol FT is the functional type factor, the symbol SI is the spatial connectivity, and the weights α, β, γ, δ, ε are automatically corrected through a machine learning model and historical disaster data.

5. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 4, wherein, The coupling module is further configured to: Dynamically adjust the parameter weights w of the flood disaster chain by combining fuzzy comprehensive evaluation and entropy weight method j ; Calculate the regional risk index based on the following formula: Among them, the symbol R represents the regional risk index, and the symbol C t represents the cumulative impact coefficient of the flood disaster chain, and the symbol P j represents the parameters of each link in the flood disaster chain, and the symbol S i represents the spatial sensitivity factor; sensitivity analysis of weights and parameters is carried out through Monte Carlo simulation to generate a risk confidence interval.

6. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 1, characterized in that, The early warning module is further configured to: map the risk index into a heat map through a GIS platform, and overlay the infrastructure distribution and the evolution path of the flood disaster chain; set a risk threshold, and trigger a warning message in combination with real-time data of Internet of Things sensors; send a hierarchical warning instruction to the emergency management department through SMS, email, and APP push methods.

7. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 6, wherein It further includes: An emergency module, configured to receive the risk spatial distribution map and the hierarchical warning instruction generated by the early warning module, and determine key protection areas according to the regional risk index; Generate a risk conduction topology map based on the dynamic coupling coefficient, deploy precise on-site duty plans for transportation hubs, power facilities, and communication nodes, and configure emergency supplies and mobile drainage equipment; real-time monitor the water level, structural deformation, and equipment operation data in the on-site area through Internet of Things sensors, and use intelligent terminals to feedback on-site situation information.

8. The critical infrastructure risk identification system in the flood disaster chain scenario according to claim 7, characterized in that, It further includes: A dispatching module, which is used to adjust the deployment density of stationed personnel and the material dispatching path through a multi-department collaborative platform according to the dynamic correction results of the spatio-temporal evolution matrix of the flood disaster chain; when the risk confidence interval exceeds the preset threshold, trigger a chain emergency response of traffic control - medical rescue - power supply, and synchronously update the spatial connectivity parameters in the infrastructure coefficient; perform local model training on the monitoring data of the stationed points through edge computing nodes, and dynamically optimize the weight allocation strategy of the key infrastructure coefficient.

9. An electronic device, characterized in that, It includes: A communication interface, a processor, and a memory; Among them, the memory is used to store program instructions, and when the program instructions are executed by the processor that is communicatively connected to the memory through the communication interface, the electronic device realizes the key infrastructure risk identification system in any one of claims 1 to 8 under the flood disaster chain scenario.

10. A computer-readable storage medium having program instructions stored thereon, characterized in that, When the program instructions are executed by a computer, the computer realizes the key infrastructure risk identification system in any one of claims 1 to 8 under the flood disaster chain scenario.

Images