Building fire internal rescue path planning method, device, equipment and medium

Abstract

The invention relates to a building fire internal rescue path planning method and device, equipment and a medium. The method comprises the following steps: acquiring an indoor vector map, fire-fighting resource position information and multi-dimensional fire scene environment data of a target building; the multi-dimensional fire scene environment data comprises at least one of fire source thermal imaging distribution data, multi-floor smoke concentration gradient data and building structure monitoring data; generating an initial static traffic network diagram based on the indoor vector map and the fire-fighting resource position information; according to the multi-dimensional fire scene environment data, carrying out dynamic risk quantitative evaluation on the traffic cost of each connection edge in the initial static traffic network diagram to obtain a dynamic risk weight value of each connection edge; generating a dynamic risk passing network diagram according to the dynamic risk weight value; and obtaining an internal rescue path of the building fire based on the dynamic risk traffic network diagram. By adopting the method, a safe and efficient building fire internal rescue path can be provided for firefighters.

Description

Technical Field

[0001] This invention belongs to the field of fire rescue route planning technology, and in particular relates to a method, device, equipment and medium for planning internal rescue routes in building fires. Background Technology

[0002] With the development of urban construction and building industrialization technology, the number of high-rise, super high-rise and large-scale complex buildings continues to grow. These buildings are characterized by complex structures, dense functions and large flow of people, which leads to the emergence of fire response methods that rely on fixed fire protection facilities and conventional emergency plans.

[0003] Traditional fire rescue operations are often carried out through pre-set evacuation route signs, static escape route planning, and manual command and dispatch. Some emergency systems can output fixed evacuation routes based on building floor plans to assist trapped personnel in evacuating from the fire scene.

[0004] However, current fire response methods and emergency systems have the following core flaws: First, the path planning logic focuses on staying away from the fire source, which is only applicable to the evacuation of trapped personnel and cannot meet the needs of firefighters who need to actively enter the fire scene and advance towards the fire. Second, the path planning data relies on static building drawings and does not take into account dynamic parameters such as the real-time spread of fire, smoke diffusion, and structural collapse risk at the fire scene. Third, there is a lack of path adaptation adjustment mechanisms for extreme environments such as dense smoke, high temperature, and low visibility, which makes the planned paths prone to failure due to sudden changes in the fire scene environment, not only delaying the rescue opportunity but also significantly increasing the operational risks for firefighters. Summary of the Invention

[0005] Therefore, it is necessary to provide a method for planning internal rescue routes in building fires to address the aforementioned technical problems.

[0006] Firstly, this application provides a method for planning internal rescue routes in building fires, including:

[0007] Acquire indoor vector maps, fire resource location information, and multi-dimensional fire environment data of the target building; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data;

[0008] An initial static access network diagram is generated based on the indoor vector map and fire resource location information.

[0009] Based on multidimensional fire scene environment data, the passage cost of each connecting edge in the initial static passage network diagram is dynamically risk-quantified and evaluated to obtain the dynamic risk weight value of each connecting edge.

[0010] A dynamic risk access network diagram is generated based on the dynamic risk weight values;

[0011] Based on the dynamic risk access network diagram, the internal rescue path for building fires is obtained.

[0012] In one embodiment, a dynamic risk access network graph is generated based on dynamic risk weight values, including:

[0013] Based on the dynamic risk weight values ​​and the preset dynamic risk level mapping rules, an initial network graph with risk level labels is obtained.

[0014] Based on the thermal imaging distribution data of the fire source and the multi-story smoke concentration gradient data, the predicted risk level sequence of each connecting edge in the initial network graph with risk level labels is obtained in multiple consecutive time slices in the future.

[0015] Based on the predicted risk level sequence, a multi-layered dynamic risk network diagram with a time dimension is obtained;

[0016] By spatiotemporally fusing the multi-layered dynamic risk network diagram with the initial network diagram, a dynamic risk access network diagram is obtained.

[0017] In one embodiment, a multi-layer dynamic risk network graph is spatiotemporally fused with an initial network graph to obtain a dynamic risk access network graph, including:

[0018] Based on the network topology of each time slice layer of the multi-layer dynamic risk network graph and the topology of the initial network graph, the node space topology mapping and alignment of the network topology of the time slice layer and the initial network topology are performed to obtain the spatiotemporally aligned network sequence.

[0019] Based on the spatiotemporally aligned network sequence and the preset spatiotemporal fusion weight function, the time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence is obtained;

[0020] Based on time-varying risk feature vectors, a three-dimensional spatiotemporal correlation matrix is ​​obtained; the first and second dimensions of the three-dimensional spatiotemporal correlation matrix index identify the connecting edges in the space, the third dimension index identifies the time series, and the matrix element values ​​represent the risk weights of the corresponding edges after fusion at the corresponding time.

[0021] Based on the three-dimensional spatiotemporal correlation matrix, a dynamic risk access network diagram is obtained.

[0022] In one embodiment, based on the spatiotemporally aligned network sequence and a preset spatiotemporal fusion weight function, the time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence is obtained, including:

[0023] Calculate the time-varying risk feature vector using the following formula:

[0024]

[0025] in, Represents the time-varying risk feature vector. Representing an edge At the present moment Real-time risk observations Representing an edge In the future The predicted risk value, This represents the fusion weighting coefficient between real-time observations and short-term forecasts. Represents the time-decay weighting coefficient and satisfies , This indicates the final moment of the prediction time window.

[0026] In one embodiment, based on a dynamic risk access network diagram, the internal rescue routes for a building fire are obtained, including:

[0027] Obtain the firefighters' interior attack and rescue mission instructions, and extract the coordinates of the rescue start point and the target endpoint from the firefighters' interior attack and rescue mission instructions;

[0028] Based on the preset protection parameters and safety criteria of firefighters' personal protective equipment, multi-dimensional risk constraints are obtained for the planning of internal rescue routes in building fires. The multi-dimensional risk constraints include instantaneous environmental risk constraints, cumulative risk threshold constraints, equipment tolerance and adaptation constraints, and structural safety redundancy constraints.

[0029] Using the coordinates of the rescue starting point and the coordinates of the target ending point as the starting and ending points, and based on the multidimensional risk constraints, an iterative search is performed in the dynamic risk access network graph to obtain at least one internal rescue path for building fires.

[0030] In one embodiment, based on multi-dimensional fire scene environment data, the passage cost of each connecting edge in the initial static passage network graph is dynamically risk-quantified and assessed to obtain the dynamic risk weight value of each connecting edge, including:

[0031] Based on the thermal imaging distribution data of the fire source and the spatial coverage area of ​​each connecting edge in the initial static traffic network diagram, the real-time thermal radiation flux intensity value of each connecting edge in the corresponding channel area is obtained.

[0032] Based on multi-story smoke concentration gradient data, the predicted values ​​of key toxic gas concentrations in the corresponding channel areas of each connecting edge are obtained within a preset time period in the future.

[0033] Based on building structure monitoring data, the instantaneous instability probability value of related building components is obtained;

[0034] Based on real-time thermal radiation flux intensity, predicted concentrations of key toxic gases, and instantaneous instability probability, a comprehensive dynamic risk quantification value for each connected edge is obtained.

[0035] Based on the comprehensive dynamic risk quantification value, the dynamic risk weight value of each connecting edge is obtained through a preset dynamic weight mapping function.

[0036] In one embodiment, an initial static access network diagram is generated based on an indoor vector map and fire resource location information, including:

[0037] Based on the building structure parameters of the indoor vector map, the channel nodes and connectivity relationships are extracted to construct a basic topology network;

[0038] Based on the location information of fire-fighting resources, resource attribute tags are added to the corresponding channel nodes in the basic topology network;

[0039] Based on the channel nodes and combined with the preset building access priority rules, static access cost initial values ​​are assigned to the edges corresponding to each connectivity relationship in the basic topology network.

[0040] By removing invalid connectivity relationships with structural obstacles from the basic topology network, an initial static access network graph is obtained.

[0041] Secondly, this application also provides an internal rescue route planning device for building fires, comprising:

[0042] The multi-source data acquisition module is used to acquire indoor vector maps of the target building, fire resource location information, and multi-dimensional fire environment data; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data;

[0043] The initial network construction module is used to generate an initial static access network diagram based on the indoor vector map and fire resource location information;

[0044] The dynamic risk quantification module is used to dynamically quantify and evaluate the passage cost of each connecting edge in the initial static passage network diagram based on multi-dimensional fire scene environment data, and obtain the dynamic risk weight value of each connecting edge.

[0045] The dynamic network generation module is used to generate a dynamic risk access network diagram based on dynamic risk weight values.

[0046] The path planning module is used to obtain internal rescue paths for building fires based on a dynamic risk access network diagram.

[0047] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.

[0048] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0049] The aforementioned method, apparatus, equipment, and medium for planning internal rescue routes in building fires involves acquiring an indoor vector map of the target building, fire resource location information, and multi-dimensional fire environment data. The multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data. Based on the indoor vector map and fire resource location information, an initial static access network diagram is generated. According to the multi-dimensional fire environment data, the access cost of each connecting edge in the initial static access network diagram is dynamically quantified and assessed to obtain a dynamic risk weight value for each connecting edge. Based on the dynamic risk weight values, a dynamic risk access network diagram is generated. Based on the dynamic risk access network diagram, the internal rescue route in the building fire is obtained. This method can provide firefighters with safe and efficient internal rescue routes in building fires. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 A flowchart illustrating an internal rescue route planning method for a building fire, provided as an exemplary embodiment of this application;

[0052] Figure 2 This is a schematic diagram of the structure of an internal rescue route planning device for building fires, provided as an exemplary embodiment of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0054] The building fire internal rescue route planning method provided in this application embodiment can be used in fire command center operation planning, on-site fire rescue route dynamic adjustment, building fire safety management system integration, personal protective equipment intelligent adaptation, structural safety intelligent monitoring, and multi-building linkage rescue route planning.

[0055] In one embodiment, such as Figure 1As shown, a method for planning internal rescue routes in building fires is provided. This embodiment illustrates the application of this method to a planning terminal. It is understood that this method can also be applied to a server, or to a system including both a terminal and a server, and is implemented through the interaction between the planning terminal and the server. In this embodiment, the method includes the following steps:

[0056] Step S101: Obtain the indoor vector map of the target building, fire resource location information, and multi-dimensional fire environment data; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data.

[0057] Among them, the indoor vector map can be a digital map of the interior space of the target building stored in a vector data format (such as SVG, SHP format), which includes geometric attribute information such as the layout of passageways, room boundaries, stairwell locations, door and window coordinates, floor elevations, and wall thickness, as well as semantic attribute information such as functional labels of each area (such as meeting rooms, evacuation routes, and equipment rooms).

[0058] Firefighting resource location information can be the specific spatial coordinates and attribute parameters of various fire-related facilities and equipment deployed inside and around the target building, including but not limited to: fire hydrants, fire water supply interfaces, air refueling equipment supply points, emergency lighting devices, fire extinguisher storage points, fire elevators, fire doors, evacuation signs, fire control rooms, mini fire stations, etc.

[0059] Multidimensional fire environment data can be a collection of multidimensional data reflecting the real-time status and evolution trend of a fire, collected through sensors, thermal imaging equipment, structural monitoring systems, etc. Specifically, it includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data. Fire source thermal imaging distribution data can be collected by infrared thermal imagers or temperature sensor arrays, presented as a heat map or coordinate-temperature mapping, characterizing the real-time temperature distribution in different areas inside the building, the location of the fire source center point, the range of thermal radiation, and other regional thermal radiation distributions. Multi-story smoke concentration gradient data can be collected through a smoke sensor network, including smoke concentration, toxic gas composition and concentration (such as carbon monoxide, hydrogen cyanide), visibility, and concentration change gradients for a future preset time period predicted based on a fluid dynamics model. Building structure monitoring data can be data collected by strain sensors and vibration sensors deployed on key load-bearing components such as beams, columns, and floor slabs, including real-time strain values, vibration frequencies, and displacements of the components, used to calculate the instantaneous instability probability of the components and reflect the safety status of the building structure.

[0060] Specifically, the planning terminal can read the pre-stored 3D model database of the target building, parse and load its contained indoor vector map layers, and obtain accurate geometric and semantic attribute information. Simultaneously, the planning terminal can extract or query the spatial coordinates and attribute parameters of registered fire resources in real time, generating structured fire resource location information. Subsequently, the planning terminal can poll and receive raw monitoring data of the multi-dimensional fire environment data at preset intervals (such as every second or every 5 seconds), perform format parsing and coordinate registration, and finally integrate them into a multi-dimensional fire environment data set with a unified spatiotemporal benchmark.

[0061] Step S102: Based on the indoor vector map and fire resource location information, generate an initial static access network diagram.

[0062] The initial static access network diagram can be a topological network model used to characterize the connectivity of accessible areas within a building and the static access cost.

[0063] Specifically, the planning terminal can parse the building structure parameters in the indoor vector map, extract key locations such as passage intersections and floor entrances as passage nodes, and construct a basic topology network based on spatial connectivity. Then, it matches the location information of fire protection resources with the coordinates of the basic topology network and adds resource attribute tags such as fire hydrants and emergency lighting devices to the corresponding passage nodes. Finally, it combines the preset building access priority rules to obtain an initial static access network map that only contains valid passable paths and fire protection resource association information.

[0064] Step S103: Based on the multi-dimensional fire scene environment data, perform dynamic risk quantification assessment on the passage cost of each connecting edge in the initial static passage network diagram to obtain the dynamic risk weight value of each connecting edge.

[0065] The dynamic risk weight value of each connecting edge can be a normalized weight parameter obtained by quantifying the passage risk of each connecting edge in the initial static passage network graph. It is used to characterize the passage danger level of the corresponding connecting edge in the current and future preset time periods.

[0066] Specifically, the planning terminal can analyze multi-dimensional fire environment data and match each data point with the spatial coordinates in the initial static traffic network diagram, associating corresponding fire environment information with each connecting edge. For fire source thermal imaging distribution data, the planning terminal calculates the shortest Euclidean distance from the fire source to the center point of each connecting edge and the intensity of thermal radiation influence, thereby quantifying the fire source threat factor. For multi-story smoke concentration gradient data, the planning terminal obtains the current measured concentration value of the area where each connecting edge is located, and combines it with the predicted future gradient data to assess the risk level of smoke coverage or toxic gas intrusion along the path within a preset time period, quantifying the smoke hazard factor. For building structure monitoring data, the planning terminal analyzes the monitoring values ​​of the structural area where each connecting edge is located. If the values ​​exceed a preset safety threshold, the terminal assesses the risk of structural collapse or instability, quantifying the structural risk factor. Subsequently, the planning terminal integrates multiple risk factors to generate dynamic risk weight values.

[0067] Step S104: Generate a dynamic risk access network diagram based on the dynamic risk weight values.

[0068] The dynamic risk access network graph can be a topology network model generated by fusing the initial static access cost of each connecting edge in the initial static access network graph with the dynamic risk weight value corresponding to that connecting edge.

[0069] Specifically, the planning terminal can load an initial static access network graph and obtain the dynamic risk weight value corresponding to each calculated connection edge. Then, based on a preset cost fusion algorithm, the planning terminal performs calculations on the initial static access cost of each edge and its corresponding dynamic risk weight value. For example, this fusion algorithm can be an additive or multiplicative model, with the specific model and coefficients determined by pre-configuration. Finally, the planning terminal replaces the original weights of each edge in the initial static access network graph with the updated comprehensive access cost weights to generate a dynamic risk access network graph.

[0070] Step S105: Based on the dynamic risk access network diagram, obtain the internal rescue path for a building fire.

[0071] Among them, the internal rescue route for building fires can be the optimal passable route from the rescue starting point to the target endpoint, planned based on a dynamic risk access network diagram and under the condition of satisfying multi-dimensional risk constraints.

[0072] Specifically, the planning terminal can obtain the pre-set coordinates of the rescue starting point and the target endpoint, and map them to the corresponding vertices of the dynamic risk access network graph. Subsequently, the planning terminal searches on the dynamic risk access network graph. The goal of the search is to find a path connecting the starting point and the endpoint, consisting of a series of continuous edges in the network graph, such that the sum of the comprehensive access cost weights of all edges of the path is minimized or meets a preset cost threshold. Finally, one or more candidate path sequences that meet the conditions are obtained. These paths are the calculated internal rescue paths for building fires.

[0073] In the aforementioned method for planning internal rescue routes in building fires, the planning terminal acquires an indoor vector map of the target building, fire resource location information, and multi-dimensional fire environment data. The multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data. Based on the indoor vector map and fire resource location information, an initial static access network diagram is generated. According to the multi-dimensional fire environment data, the access cost of each connecting edge in the initial static access network diagram is dynamically quantified and assessed to obtain a dynamic risk weight value for each connecting edge. Based on the dynamic risk weight values, a dynamic risk access network diagram is generated. Based on the dynamic risk access network diagram, the internal rescue route in the building fire is obtained. This method can provide firefighters with safe and efficient internal rescue routes in building fires.

[0074] In one embodiment, generating a dynamic risk access network graph based on dynamic risk weight values ​​may include the following steps:

[0075] Step S201: Based on the dynamic risk weight value and the preset dynamic risk level mapping rule, an initial network graph with risk level labels is obtained.

[0076] The preset dynamic risk level mapping rule can be a pre-configured quantitative judgment rule used to map the dynamic risk weight value of each connection edge to the corresponding risk level. This rule can be dynamically adjusted according to the building type and rescue scenario requirements.

[0077] An initial network graph with risk level labels can be a topology network model generated by adding corresponding risk level labels to each connecting edge based on an initial static passable network graph and according to a preset dynamic risk level mapping rule.

[0078] Specifically, the planning terminal can load the initial static access network graph and the calculated dynamic risk weight values ​​for each connecting edge. Then, the planning terminal retrieves a pre-set dynamic risk level mapping rule, which defines the corresponding range between dynamic risk weight values ​​and discrete risk levels. The planning terminal then compares the dynamic risk weight value of each connecting edge with this rule to determine its risk level. Finally, the planning terminal appends this risk level as a label attribute to the connecting edge data structure corresponding to the initial static access network graph, generating a network model that retains the original topology and static access cost, while each edge has a clearly defined risk level label—that is, an initial network graph with risk level labels.

[0079] Step S202: Based on the fire source thermal imaging distribution data and the multi-story smoke concentration gradient data, obtain the predicted risk level sequence of each connecting edge in the initial network graph with risk level labels on multiple consecutive time slices in the future.

[0080] The predicted risk level sequence can be a sequence of data formed by arranging the risk levels of each connection edge in the initial network graph with risk level labels in chronological order over multiple consecutive time slices in the future.

[0081] Specifically, the planning terminal can quantitatively predict the potential thermal radiation intensity and smoke concentration of each connected edge in the initial network diagram with risk level labels, based on the fire spread trend information contained in the fire source thermal imaging distribution data and the smoke diffusion pattern information contained in the multi-story smoke concentration gradient data, across multiple consecutive discrete time slices in the future. Subsequently, the planning terminal compares the predicted thermal radiation intensity and smoke concentration values ​​with preset thermal radiation risk thresholds and smoke concentration risk thresholds, respectively, to determine their compliance status in each time slice.

[0082] Furthermore, the planning terminal determines a risk level for each future time slice of each connection edge based on a preset risk level determination logic (which defines the risk level corresponding to different combinations of compliance conditions). Finally, for each connection edge, the planning terminal generates a list of risk levels arranged in the order of time slices, i.e., a predicted risk level sequence.

[0083] Step S203: Based on the predicted risk level sequence, a multi-layered dynamic risk network diagram with a time dimension is obtained.

[0084] Among them, the multi-layer dynamic risk network graph can be a multi-level topological network model that includes spatial connectivity and temporal risk evolution dimensions, which is constructed by layering the initial network graph with risk level labels according to multiple consecutive time slices in the future based on the predicted risk level sequence.

[0085] Specifically, the planning terminal can extract the time slice parameters and the risk level of each connecting edge at the corresponding time from the predicted risk level sequence; then, based on the initial network graph with risk level labels, an independent subnet is constructed for each time slice, retaining the original network's nodes, connecting edges, and fire resource association information, while binding the corresponding time risk level labels to the connecting edges in each subnet; finally, all subnets are superimposed in chronological order to generate a complete multi-layer dynamic risk network graph.

[0086] Step S204: The multi-layer dynamic risk network diagram is spatiotemporally fused with the initial network diagram to obtain a dynamic risk access network diagram.

[0087] Specifically, the planning terminal can extract risk level label information from each time slice layer of the multi-layer dynamic risk network graph and extract the static access cost weights from the initial network graph. Subsequently, the planning terminal uses a preset spatiotemporal fusion weight function to weight and fuse the quantified risk level value of each time slice with the initial static access cost, generating a time-varying comprehensive weight matrix that comprehensively represents the changes in the comprehensive access cost of each connection edge at different times. This fusion process can be represented by the following formula:

[0088]

[0089] in, Representing an edge At any moment The weighted average of the combined passage costs after fusion Representing an edge Initial static passage cost, Representing an edge At any moment Risk level quantification value, Represents the static cost base coefficient. This represents the risk impact coefficient, used to adjust the amplification effect of risk level on the overall weight.

[0090] Finally, the planning terminal updates these updated comprehensive weights, which incorporate spatiotemporal risk information, onto the edges corresponding to the initial network graph, resulting in a dynamic risk access network graph.

[0091] In this embodiment, the planning terminal combines dynamic risk quantification, temporal risk prediction, and spatiotemporal fusion to generate a dynamic risk access network map that combines spatial connectivity and temporal dynamic adaptability, providing core network support for subsequent planning of safe and efficient internal rescue routes in building fires.

[0092] In one embodiment, spatiotemporal fusion of a multi-layer dynamic risk network graph with an initial network graph to obtain a dynamic risk access network graph may include the following steps:

[0093] Step S301: Based on the network topology of each time slice layer of the multi-layer dynamic risk network graph and the topology of the initial network graph, perform node spatial topology mapping and alignment between the network topology of the time slice layer and the initial network topology to obtain a spatiotemporally aligned network sequence.

[0094] Specifically, when each time slice layer in the multi-layer dynamic risk network graph has the same vertex set and edge connectivity as the initial network graph, the planning terminal uses the topology of the initial network graph as a spatiotemporal reference. It then sequentially matches and aligns each time slice layer in the multi-layer dynamic risk network graph with this reference. Ultimately, the planning terminal obtains a serialized data structure where each layer of the network, arranged in chronological order, corresponds topologically to the initial network graph, generating a spatiotemporally aligned network sequence.

[0095] Step S302: Based on the spatiotemporally aligned network sequence and the preset spatiotemporal fusion weight function, obtain the time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence.

[0096] The preset spatiotemporal fusion weight function can be a pre-configured calculation rule or function expression used to weight and fuse risk level data on different time slices.

[0097] A time-varying risk feature vector can be a set of characteristic data that represents the changing trend and intensity distribution of the risk level or comprehensive risk quantification value of a certain connection edge in a spatiotemporally aligned network sequence over multiple consecutive time slices in the future.

[0098] Specifically, the planning terminal can extract the risk level quantification value (or the value mapped to the risk level) of each connection edge in the spatiotemporally aligned network sequence at each time slice layer. Then, the planning terminal uses a preset spatiotemporal fusion weighting function, which assigns a weight coefficient to each time slice. Finally, the planning terminal performs a weighted calculation on the risk quantification value of each time slice and its corresponding weight coefficient, and arranges the weighted results (or new values ​​obtained after further processing) into an ordered array according to time order, which is the instantaneous risk feature vector.

[0099] Step S303: Based on the time-varying risk feature vector, a three-dimensional spatiotemporal correlation matrix is ​​obtained; the first and second dimensions of the three-dimensional spatiotemporal correlation matrix index the connecting edges in the space, the third dimension index identifies the time series, and the matrix element values ​​represent the risk weights of the corresponding edges after fusion at the corresponding time.

[0100] Among them, the three-dimensional spatiotemporal correlation matrix can be a three-dimensional matrix structure with both spatial and temporal dimensions, used to uniformly store and represent the risk weights of all connecting edges in the dynamic risk passage network graph after fusion at different times.

[0101] Specifically, when the planning terminal can generate a spatiotemporally aligned network sequence, it traverses all connections on each time slice in the sequence. For each connection and its corresponding time-varying risk feature vector, the planning terminal calculates its fused risk weight on each time slice according to a preset mapping rule from feature vector to weight value. Subsequently, the planning terminal fills the obtained fused risk weights into a preset three-dimensional matrix data structure according to the spatial index of its corresponding connection and the time index of the time slice, ultimately obtaining a complete three-dimensional spatiotemporal correlation matrix.

[0102] Step S304: Based on the three-dimensional spatiotemporal correlation matrix, a dynamic risk access network diagram is obtained.

[0103] Specifically, the planning terminal can extract the fused risk weights of each connecting edge in the three-dimensional spatiotemporal correlation matrix at different time slices. Based on the topology of the initial static access network graph, the dynamic risk weights of each connecting edge are updated to the fused risk weights of the corresponding spatiotemporal dimensions in the matrix. At the same time, the attributes of fire resource nodes and effective connectivity are preserved, generating a dynamic network model that can reflect the access risks of each passage at different times in real time, i.e., a dynamic risk access network graph.

[0104] In this embodiment, the planning terminal generates a dynamic risk access network diagram to characterize the spatiotemporal risk evolution of each channel through topology alignment, spatiotemporal fusion weighting, three-dimensional matrix modeling, and network weight updates, providing dynamic and comprehensive risk network support for rescue route planning.

[0105] In one embodiment, based on the spatiotemporally aligned network sequence and a preset spatiotemporal fusion weight function, the time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence is obtained, including:

[0106] Calculate the time-varying risk feature vector using the following formula:

[0107]

[0108] in, Represents the time-varying risk feature vector. Representing an edge At the present moment Real-time risk observations Representing an edge In the future The predicted risk value, This represents the fusion weighting coefficient between real-time observations and short-term forecasts. Represents the time-decay weighting coefficient and satisfies , This indicates the final moment of the prediction time window.

[0109] Specifically, Representing an edge At the present moment The real-time risk observation value is derived from the risk quantification result of the multi-dimensional fire environment data (such as fire source thermal imaging distribution data and multi-story smoke concentration gradient data) after real-time analysis and quantification, and then mapped to the spatial area where the edge is located. Representing an edge In the future The predicted risk value is derived from the predicted state of the fire environment within a preset time period, and is obtained by mapping the risk level to the spatial area where the edge is located. This represents the fusion weighting coefficient between real-time observation and short-term forecasting, used to adjust the contribution ratio of current measured risk and predicted risk at the next moment in short-term risk assessment; This represents the time-series decay weighting coefficient, used to weight and fuse risk values ​​at multiple future prediction times. Generally, the more recent the prediction value, the greater its weight.

[0110] In this embodiment, the planning terminal uses a formula to fuse real-time risk observations and future predicted risk values ​​to generate a time-varying risk feature vector that characterizes the spatiotemporal risk evolution of the connection edge.

[0111] In one embodiment, obtaining the internal rescue path for a building fire based on a dynamic risk access network diagram may include the following steps:

[0112] Step S401: Obtain the firefighter's interior attack rescue mission instruction, and extract the rescue starting point coordinates and target endpoint coordinates from the firefighter's interior attack rescue mission instruction.

[0113] Among them, the firefighter interior attack and rescue mission instructions can be issued by the fire command system or the on-site command center. These instructions are specific to the target building fire scenario and are used to clarify the core mission requirements and key parameters for firefighters to go deep into the fire scene to carry out rescue operations.

[0114] Specifically, the planning terminal can acquire structured firefighter interior attack rescue mission instructions in real time when communicating with the fire command system data interface or rescue personnel's handheld terminals. Subsequently, the planning terminal further parses the firefighter interior attack rescue mission instructions, identifies the text descriptions or logical identifiers corresponding to the rescue starting point and the target endpoint, and then performs semantic matching and coordinate transformation based on the pre-loaded indoor vector map to map them into precise spatial coordinates in the map coordinate system, namely the coordinates of the rescue starting point and the target endpoint.

[0115] Step S402: Based on the preset protection parameters and safety criteria of firefighters' personal protective equipment, obtain the multi-dimensional risk constraints for planning the internal rescue path in a building fire. The multi-dimensional risk constraints include instantaneous environmental risk constraints, cumulative risk threshold constraints, equipment tolerance and adaptation constraints, and structural safety redundancy constraints.

[0116] Among them, the preset protection parameters of firefighters' personal protective equipment can be the core technical indicators and tolerance limits of the personal protective equipment worn by firefighters (such as fire helmets, flame-retardant suits, air respirators, heat-resistant gloves, protective boots, etc.) when resisting the harmful environment of the fire scene.

[0117] Safety guidelines can be formulated based on fire and rescue industry standards, fire scene operation safety management regulations, and practical experience. They are fundamental and mandatory behavioral norms and judgment standards used to ensure the safety of firefighters in interior attack and rescue operations.

[0118] Multidimensional risk constraints can be a set of constraints that covers multiple dimensions such as fire scene environment, equipment tolerance, and structural safety, based on preset protection parameters and safety criteria for firefighters' personal protective equipment.

[0119] Specifically, the planning terminal can access a pre-built database of firefighter personal protective equipment (PPE) to obtain detailed protective parameters of the equipment worn by firefighters on the current mission (such as the rated usage time of the air respirator, the high-temperature resistance threshold of the thermal suit, and the smoke penetration resistance level of the mask). Simultaneously, based on safety guidelines built into the system or obtained through external interfaces, the planning terminal extracts the detailed protective parameters of the aforementioned equipment and various safety limit thresholds specified in the safety guidelines (such as upper limits for ambient temperature, limits for toxic gas concentration, thresholds for continuous operation time, and minimum structural stability requirements). Through a pre-set rule engine, it quantifies and correlates the equipment protective parameters with the safety limit thresholds, generating threshold constraints for instantaneous environmental risks, risk upper limit constraints based on cumulative exposure, equipment performance and environmental compatibility constraints, and redundancy constraints considering structural safety margins. Finally, it generates a multi-dimensional risk constraint set for path planning. The specific expression is as follows:

[0120]

[0121] in, This represents the instantaneous environmental risk constraint. This represents any node in the path (a specific location within the fire). This represents the set of rescue routes to be planned. Represents a node Instantaneous ambient temperature, Represents a node The concentration of toxic gases, Represents a node Flue gas infiltration pressure, Represents a node The safety factor of instantaneous ambient temperature. Represents a node The safety factor for the concentration of toxic gases, Represents a node The safety factor of flue gas infiltration pressure. This indicates the high-temperature resistance threshold in the equipment's protection parameters. This indicates the concentration limits for toxic gases in safety guidelines. This indicates the threshold pressure for preventing smoke infiltration.

[0122]

[0123] in, This represents the cumulative risk threshold constraint. Indicates the total time taken for the entire path. Indicates different times Ambient temperature, Indicates the concentration of toxic gases. Indicates the weighting coefficient. Indicates the weighting coefficient. Indicates the weighting coefficient. This represents the threshold for continuous operation time in safety guidelines.

[0124]

[0125] in, This indicates the equipment's tolerance and adaptation constraints. This represents the environmental intensity correction factor. The rated service life of an air-supplied breathing apparatus is derived from the equipment's protective parameters. This represents the average heat intensity at each node along the path.

[0126]

[0127] in, This represents the structural safety redundancy constraint condition. Represents a node The actual structural safety factor of the area (such as the load-bearing safety factor of beams and columns). Represents a node The critical safety factor of the structure in the area (the critical value at which the structure is on the verge of instability). This indicates the minimum structural safety factor requirement in the safety criteria. This represents the structural safety redundancy coefficient.

[0128] Step S403: Using the coordinates of the rescue starting point and the coordinates of the target ending point as the starting and ending points, perform an iterative search in the dynamic risk access network graph according to the multi-dimensional risk constraints to obtain at least one internal rescue path for building fires.

[0129] Specifically, the planning terminal can map the coordinates of the rescue starting point and the target endpoint to corresponding nodes in a dynamic risk access network graph. Subsequently, the planning terminal performs an iterative search on the dynamic risk access network graph. During the search, the planning terminal continuously visits network nodes and expands paths, while simultaneously calculating the comprehensive risk value of each edge on candidate paths in real time and comparing it with multi-dimensional risk constraints (such as instantaneous risk threshold, cumulative risk limit, equipment tolerance matching degree, and structural safety redundancy). If a candidate path violates any constraint at any time or as a whole, it is pruned and excluded from the search space. The planning terminal continues to iterate until it finds at least one connected path from the starting point to the endpoint that satisfies all multi-dimensional risk constraints throughout, and this path is then identified as the internal rescue path for the building fire.

[0130] In this embodiment, the planning terminal integrates task instruction parsing, refined risk constraint modeling based on equipment and criteria, and performs constrained path search in a dynamic risk network. This enables intelligent rescue path planning that comprehensively considers real-time fire threats, firefighters' personal protective capabilities, and structural safety and reliability, significantly improving the safety, adaptability, and feasibility of firefighters' interior attack rescue paths.

[0131] In one embodiment, based on multi-dimensional fire scene environment data, a dynamic risk quantification assessment is performed on the passage cost of each connecting edge in the initial static passage network graph to obtain the dynamic risk weight value of each connecting edge. This may include the following steps:

[0132] Step S501: Based on the fire source thermal imaging distribution data and the spatial coverage area of ​​each connecting edge in the initial static traffic network diagram, obtain the real-time thermal radiation flux intensity value of the corresponding channel area of ​​each connecting edge.

[0133] Among them, the real-time thermal radiation flux intensity value can be a quantitative value used to characterize the thermal radiation energy received per unit area per unit time from the fire source within the channel area covered by the connecting edge.

[0134] Specifically, the planning terminal can analyze the temperature field information and geometric parameters of the fire sources in the thermal imaging distribution data, and then extract the geometric center coordinates or coverage boundary range of each connecting edge in the initial static traffic network diagram. Further, the planning terminal calculates the effective thermal radiation flux from each fire source to the spatial region of each connecting edge, and may superimpose the radiation contributions from multiple fire sources. Finally, the planning terminal outputs a quantified real-time thermal radiation flux intensity value for each connecting edge, used to characterize the instantaneous thermal risk directly caused by the fire source along its traffic path.

[0135] Step S502: Based on the multi-story smoke concentration gradient data, obtain the predicted values ​​of key toxic gas concentrations in the corresponding channel areas of each connecting edge within a preset time period in the future.

[0136] Among them, the predicted value of key toxic gas concentration can be the quantitative prediction result of the concentration of core toxic gases (such as carbon monoxide, hydrogen cyanide, hydrogen sulfide, etc.) that pose a threat to the life safety of firefighters in the corresponding passage area of ​​each connecting edge within a future preset time period, calculated based on the multi-story smoke concentration gradient data, combined with the building structure characteristics, ventilation conditions and smoke diffusion laws.

[0137] Specifically, the planning terminal can analyze multi-story smoke concentration gradient data to obtain the current smoke concentration distribution, composition, and diffusion trend of each floor and area. Subsequently, based on a pre-set smoke fluid dynamics model (such as a CFD model or its simplified version) and information such as ventilation and window / door opening / closing status within the building, the planning terminal simulates the smoke diffusion process within the target building. For each connecting edge corresponding to a passage area, the planning terminal calculates the concentration change curves of key toxic gases (such as carbon monoxide and hydrogen cyanide) at multiple discrete time points within a pre-set future time period, extracting or predicting the critical concentration peaks or cumulative exposure concentrations that pose a direct hazard to firefighters, ultimately generating predicted values ​​for key toxic gas concentrations.

[0138] Step S503: Based on the building structure monitoring data, obtain the instantaneous instability probability value of the associated building components.

[0139] The instantaneous instability probability value can be a quantitative value calculated based on building structure monitoring data (such as component stress and strain data, displacement monitoring data, crack width data, vibration frequency data, etc.) and directly related to the channel area where the connecting edge is located, representing the probability of sudden structural failure (such as collapse, fracture, instability) at the current moment.

[0140] Specifically, the planning terminal can acquire building structure monitoring data collected by a sensor network deployed on key building structural components (such as beams, columns, floor slabs, and walls), including parameters such as real-time strain, stress, vibration frequency, tilt angle, and crack width. Subsequently, based on a pre-established structural health assessment model or failure probability prediction algorithm (such as one based on limit state equations or machine learning models), the planning terminal uses monitoring data from multiple components associated with each connecting edge within the spatial channel area as input to analyze their current stress state, degree of damage accumulation, and mechanical performance degradation under high temperature. Furthermore, the planning terminal comprehensively calculates and assesses the probability of these components experiencing instantaneous instability (such as bending instability, shear failure, or localized crushing) at the current moment, ultimately generating an instantaneous instability probability value.

[0141] Step S504: Based on the real-time thermal radiation flux intensity value, the predicted value of the concentration of key toxic gases, and the instantaneous instability probability value, obtain the comprehensive dynamic risk quantification value of each connection edge.

[0142] Among them, the comprehensive dynamic risk quantification value can be a quantitative value used to comprehensively characterize the degree of passage risk of a single connecting edge in the initial static passage network graph.

[0143] Specifically, the planning terminal can establish a multi-dimensional risk data set for each connection edge, including real-time thermal radiation flux intensity, predicted concentrations of key toxic gases, and instantaneous instability probability values. Then, based on pre-set weighting coefficients, normalization rules, and threshold comparison functions, thermal risk, toxicity risk, and structural risk are standardized and quantitatively scored respectively. Finally, the planning terminal integrates the quantitative scores from different dimensions to generate a comprehensive dynamic risk quantification value.

[0144] The expression for the comprehensive dynamic risk quantification value is as follows:

[0145]

[0146] in, Indicates connecting edges The comprehensive dynamic risk quantification value, This represents the normalized thermal radiation risk score. This represents the normalized risk score for toxic gases. This represents the structural instability risk score after normalization. This represents the fusion weighting coefficient for the thermal radiation dimension. This represents the fusion weighting coefficient for key toxic gases. The fusion weighting coefficient represents the structural instability, and .

[0147] Step S505: Based on the comprehensive dynamic risk quantification value, obtain the dynamic risk weight value of each connecting edge through a preset dynamic weight mapping function.

[0148] The preset dynamic weight mapping function can be a predefined mathematical function or mapping rule stored in the planning terminal, used to convert the comprehensive dynamic risk quantification value into a normalized dynamic risk weight value suitable for subsequent network graph weight updates.

[0149] Specifically, the planning terminal can traverse all connected edges, read their comprehensive dynamic risk quantification values ​​one by one, and call the preset dynamic weight mapping function. Subsequently, the planning terminal calculates the dynamic risk weight value by executing the dynamic weight mapping function; finally, the planning terminal uses the calculated dynamic risk weight value as a key attribute and binds it to each connected edge.

[0150] In this embodiment, the planning terminal systematically quantifies real-time fire risks from multiple dimensions, such as thermal radiation, toxic gases, and structural instability, and integrates them into a comprehensive dynamic risk quantification value, which is then mapped into usable network weights, thereby realizing a dynamic assessment of the risks of the internal access network of a building.

[0151] In one embodiment, generating an initial static access network map based on an indoor vector map and fire resource location information may include the following steps:

[0152] Step S601: Based on the building structure parameters of the indoor vector map, extract the channel nodes and connectivity relationships to construct the basic topology network.

[0153] Among them, the building structure parameters of the indoor vector map can be various quantitative attributes parsed from the indoor vector map data to describe the spatial layout and connectivity of the building's interior, including but not limited to: the geometric position of building components (such as wall coordinates, door and window boundaries), floor plan outline, width and direction of passages (such as corridors and passageways), spatial location and connecting floor information of stairwells, vertical coordinates of elevator shafts, and boundary range of each functional area (such as rooms, halls, and equipment rooms).

[0154] The basic topology network can be an abstract representation of the internal passage structure of a building using a graph theory model. In this model, passage nodes represent key spatial locations within the building that are accessible (such as passage intersections, room entrances, stair landings, elevator lobbies, etc.), while connectivity relationships (i.e. edges) represent feasible path segments connecting two adjacent passage nodes (such as a corridor, a staircase, or a path through a door).

[0155] Specifically, the planning terminal can parse indoor vector map data, identify all geometric elements such as walls, doors, windows, and passageway outlines, and then automatically identify key locations such as passageway intersections, corridor endpoints, staircase connections, and elevator entrances and exits, and determine them as topological nodes. Subsequently, based on the spatial adjacency relationships of these nodes and the actual passable directions (such as one-way doors and two-way passageways), the planning terminal establishes connecting edges between adjacent nodes, thereby constructing a basic topological network that fully reflects the internal spatial passage framework of the building.

[0156] Step S602: Based on the fire resource location information, add resource attribute tags to the corresponding channel nodes in the basic topology network.

[0157] Among them, resource attribute tags can be structured data identifiers attached to basic topology network nodes, used to characterize the type of fire protection resources associated with the node, status parameters, and spatial relationships.

[0158] Specifically, the planning terminal can read fire resource location information data to obtain the precise coordinates of various fire resources (such as fire hydrants, fire extinguishers, emergency lighting, and air respirator supply points) within the building. Subsequently, the planning terminal matches these coordinates with the spatial locations of nodes in the basic topology network (e.g., based on the nearest neighbor distance algorithm). When the coordinates of a resource fall within a node's preset neighborhood (e.g., within 3 meters), the resource is determined to be associated with that node. The planning terminal then appends the resource's type, unique identifier, status, and other attribute information to the node's data structure in a preset format (e.g., a JSON object) as resource attribute tags.

[0159] Step S603: Based on the channel nodes and in conjunction with the preset building access priority rules, assign initial static access cost values ​​to the edges corresponding to each connectivity relationship in the basic topology network.

[0160] The preset building access priority rules can be a set of quantitative criteria and judgment logic predefined and stored in the planning terminal. These rules are used to assign different static access cost base values ​​to different types of access routes based on building structural characteristics, passageway functions, and fire rescue operation requirements. These rules are typically based on the following factors: passageway type (e.g., main corridor, secondary passageway, fire-fighting passageway, staircase), passageway direction (e.g., upstairs staircase, downstairs staircase, two-way passageway), passageway physical attributes (e.g., width, slope, whether fire doors are installed), and its strategic importance in evacuation or internal attack.

[0161] Specifically, the planning terminal can traverse every connected edge in the basic topology network, identifying the type of channel node it connects to and the actual physical channel attribute corresponding to that edge. Subsequently, the planning terminal assigns a cost value to each edge according to a preset building access priority rule: this rule defines the mapping relationship between various channel attribute combinations and static access costs, which can be calculated using the following cost function:

[0162]

[0163] in, Representing an edge The initial value of the static passage cost, The base cost representing the channel type. Representing an edge The Additional attributes (such as width, slope, fire door status, etc.) This represents the cost adjustment function. This indicates the adjustment weighting coefficient.

[0164] Step S604: Remove invalid connectivity relationships with structural obstacles in the basic topology network to obtain the initial static access network graph.

[0165] Specifically, the planning terminal can compare the actual physical passages corresponding to each connected edge in the basic topology network with the building structural elements in the indoor vector map to identify impassable situations caused by structural changes or obstacles. For the identified invalid connections, the planning terminal removes them from the basic topology network. Finally, the planning terminal performs a topology consistency check on the network and further generates an initial static access network diagram that accurately reflects the current accessible framework of the building.

[0166] In this embodiment, the planning terminal systematically analyzes the building structure, integrates fire resource information, assigns static costs according to traffic rules, and eliminates invalid paths, thus constructing an accurate, structured, and resource-enhanced initial static traffic network map. This lays a solid and reliable static data foundation for subsequent integration of dynamic fire risk and efficient and safe rescue route planning.

[0167] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0168] Based on the same inventive concept, this application also provides a device for planning internal rescue routes in building fires as described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the internal rescue route planning device for building fires provided below can be found in the limitations of the internal rescue route planning method for building fires described above, and will not be repeated here.

[0169] In one exemplary embodiment, such as Figure 2 As shown, a building fire interior rescue route planning device 700 is provided, comprising:

[0170] The multi-source data acquisition module 701 is used to acquire indoor vector maps of the target building, fire resource location information, and multi-dimensional fire environment data; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data.

[0171] The initial network construction module 702 is used to generate an initial static access network diagram based on the indoor vector map and fire resource location information;

[0172] The dynamic risk quantification module 703 is used to perform dynamic risk quantification assessment on the passage cost of each connecting edge in the initial static passage network diagram based on multi-dimensional fire scene environment data, and obtain the dynamic risk weight value of each connecting edge.

[0173] The dynamic network generation module 704 is used to generate a dynamic risk access network diagram based on the dynamic risk weight value.

[0174] The path planning module 705 is used to obtain internal rescue paths for building fires based on a dynamic risk access network diagram.

[0175] In one embodiment, the dynamic network generation module includes:

[0176] The risk level labeling unit is used to obtain an initial network graph with risk level labels based on the dynamic risk weight value and the preset dynamic risk level mapping rule.

[0177] The temporal risk prediction unit is used to obtain the predicted risk level sequence of each connecting edge in the initial network graph with risk level labels on multiple consecutive time slices in the future, based on the fire source thermal imaging distribution data and the multi-story smoke concentration gradient data.

[0178] Multi-layer network building units are used to obtain a multi-layer dynamic risk network diagram with a time dimension based on the predicted risk level sequence.

[0179] The spatiotemporal fusion unit is used to spatiotemporally fuse the multi-layer dynamic risk network diagram with the initial network diagram to obtain a dynamic risk access network diagram.

[0180] In one embodiment, the spatiotemporal fusion unit includes:

[0181] The spatiotemporal alignment subunit is used to perform node spatial topology mapping and alignment between the network topology of each time slice layer of the multi-layer dynamic risk network graph and the topology of the initial network graph, so as to obtain a spatiotemporally aligned network sequence.

[0182] The vector computation subunit is used to obtain the time-varying risk feature vectors corresponding to each connection edge in the spatiotemporally aligned network sequence based on the spatiotemporally aligned network sequence and the preset spatiotemporal fusion weight function;

[0183] The correlation matrix generation sub-unit is used to obtain a three-dimensional spatiotemporal correlation matrix based on time-varying risk feature vectors. The first and second dimensions of the three-dimensional spatiotemporal correlation matrix index identify the connecting edges in the space, the third dimension index identifies the time series, and the matrix element values ​​represent the risk weights of the corresponding edges after fusion at the corresponding time.

[0184] The dynamic network generation sub-unit is used to obtain a dynamic risk access network diagram based on a three-dimensional spatiotemporal correlation matrix.

[0185] In one embodiment, the vector computation subunit includes:

[0186] Calculate the time-varying risk feature vector using the following formula:

[0187]

[0188] in, Represents the time-varying risk feature vector. Representing an edge At the present moment Real-time risk observations Representing an edge In the future The predicted risk value, This represents the fusion weighting coefficient between real-time observations and short-term forecasts. Represents the time-decay weighting coefficient and satisfies , This indicates the final moment of the prediction time window.

[0189] In one embodiment, the path planning module includes:

[0190] The mission instruction parsing unit is used to obtain the firefighters' interior attack and rescue mission instructions and extract the coordinates of the rescue start point and the target end point from the firefighters' interior attack and rescue mission instructions.

[0191] The risk constraint generation unit is used to obtain multi-dimensional risk constraints for planning internal rescue routes in building fires based on preset protection parameters and safety criteria for firefighters' personal protective equipment. The multi-dimensional risk constraints include instantaneous environmental risk constraints, cumulative risk threshold constraints, equipment tolerance and adaptation constraints, and structural safety redundancy constraints.

[0192] The constrained path search unit is used to perform an iterative search in the dynamic risk access network graph based on the coordinates of the rescue starting point and the coordinates of the target ending point, according to multi-dimensional risk constraints, to obtain at least one internal rescue path for building fires.

[0193] In one embodiment, the dynamic risk quantification module includes:

[0194] The thermal radiation quantization unit is used to obtain the real-time thermal radiation flux intensity value of each connecting edge within the corresponding channel area based on the fire source thermal imaging distribution data and the spatial coverage area of ​​each connecting edge in the initial static traffic network diagram.

[0195] The toxic gas prediction unit is used to obtain the predicted value of the key toxic gas concentration in the channel area corresponding to each connecting edge within a preset time period based on the multi-story smoke concentration gradient data.

[0196] The structural instability assessment unit is used to obtain the instantaneous instability probability value of related building components based on building structure monitoring data;

[0197] The comprehensive risk calculation unit is used to obtain the comprehensive dynamic risk quantification value of each connected edge based on the real-time thermal radiation flux intensity value, the predicted value of the concentration of key toxic gases and the instantaneous instability probability value.

[0198] The weight mapping unit is used to obtain the dynamic risk weight value of each connection edge based on the comprehensive dynamic risk quantification value and through a preset dynamic weight mapping function.

[0199] In one embodiment, the initial network building module includes:

[0200] The topology network building unit is used to extract channel nodes and connectivity relationships based on the building structure parameters of indoor vector maps, and to build a basic topology network.

[0201] The resource tag attachment unit is used to add resource attribute tags to the corresponding channel nodes in the basic topology network based on fire resource location information.

[0202] The static cost assignment unit is used to assign initial static passage cost values ​​to the edges corresponding to each connectivity relationship in the basic topology network based on the channel nodes and the preset building access priority rules.

[0203] The invalid connectivity removal unit is used to remove invalid connectivity relationships with structural obstacles in the basic topology network to obtain the initial static passable network graph.

[0204] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the building fire interior rescue route planning method as described above.

[0205] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0206] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0207] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for planning internal rescue routes in a building fire, characterized in that, The method includes: Acquire indoor vector maps, fire resource location information, and multi-dimensional fire environment data of the target building; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data; Based on the indoor vector map and the location information of the fire-fighting resources, an initial static access network diagram is generated; Based on the multidimensional fire scene environment data, the passage cost of each connecting edge in the initial static passage network graph is dynamically risk-quantified and evaluated to obtain the dynamic risk weight value of each connecting edge. Based on the dynamic risk weight values, a dynamic risk access network diagram is generated; Based on the dynamic risk access network diagram, the internal rescue path for a building fire is obtained.

2. The method according to claim 1, characterized in that, The step of generating a dynamic risk access network diagram based on the dynamic risk weight value includes: Based on the dynamic risk weight values ​​and the preset dynamic risk level mapping rules, an initial network graph with risk level labels is obtained. Based on the fire source thermal imaging distribution data and the multi-story smoke concentration gradient data, the predicted risk level sequence of each connection edge in the initial network graph with risk level labels is obtained in multiple consecutive time slices in the future. Based on the predicted risk level sequence, a multi-layered dynamic risk network diagram with a time dimension is obtained; The dynamic risk network graph is spatiotemporally fused with the initial network graph to obtain the dynamic risk access network graph.

3. The method according to claim 2, characterized in that, The step of spatiotemporally fusing the multi-layer dynamic risk network graph with the initial network graph to obtain the dynamic risk access network graph includes: Based on the network topology of each time slice layer of the multi-layer dynamic risk network graph and the topology of the initial network graph, the node space topology mapping and alignment of the network topology of the time slice layer and the initial network topology are performed to obtain a spatiotemporally aligned network sequence. Based on the spatiotemporally aligned network sequence and the preset spatiotemporal fusion weight function, the time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence is obtained; Based on the time-varying risk feature vector, a three-dimensional spatiotemporal correlation matrix is ​​obtained; the first and second dimensions of the three-dimensional spatiotemporal correlation matrix index identify the connecting edges in the space, the third dimension index identifies the time series, and the matrix element values ​​represent the risk weights of the corresponding edges after fusion at the corresponding time. Based on the three-dimensional spatiotemporal correlation matrix, the dynamic risk access network diagram is obtained.

4. The method according to claim 3, characterized in that, The time-varying risk feature vector corresponding to each connection edge in the spatiotemporally aligned network sequence is obtained by applying the spatiotemporally aligned network sequence to a preset spatiotemporal fusion weight function, including: The time-varying risk feature vector is calculated using the following formula: in, Represents the time-varying risk feature vector. Representing an edge At the present moment Real-time risk observations Representing an edge In the future The predicted risk value, This represents the fusion weighting coefficient between real-time observations and short-term forecasts. Represents the time-decay weighting coefficient and satisfies , This indicates the final moment of the prediction time window.

5. The method according to claim 1, characterized in that, The method for obtaining internal rescue routes in a building fire based on the dynamic risk access network diagram includes: Obtain the firefighters' interior attack and rescue mission instructions, and extract the rescue start point coordinates and target end point coordinates from the firefighters' interior attack and rescue mission instructions; Based on the preset protection parameters and safety criteria of firefighters' personal protective equipment, multidimensional risk constraints are obtained for the planning of internal rescue routes in building fires. The multidimensional risk constraints include instantaneous environmental risk constraints, cumulative risk threshold constraints, equipment tolerance and adaptation constraints, and structural safety redundancy constraints. Using the coordinates of the rescue starting point and the coordinates of the target ending point as the starting and ending points, and based on the multidimensional risk constraints, an iterative search is performed in the dynamic risk access network graph to obtain at least one internal rescue path for a building fire.

6. The method according to any one of claims 1 to 5, characterized in that, The step of dynamically quantifying the risk of each connecting edge in the initial static access network graph based on the multi-dimensional fire scene environment data to obtain the dynamic risk weight value of each connecting edge includes: Based on the fire source thermal imaging distribution data and the spatial coverage area of ​​each connecting edge in the initial static traffic network diagram, the real-time thermal radiation flux intensity value of each connecting edge in the corresponding channel area is obtained. Based on the multi-story smoke concentration gradient data, the predicted values ​​of key toxic gas concentrations in the channel areas corresponding to each connecting edge are obtained within a future preset time period. Based on the building structure monitoring data, the instantaneous instability probability value of the associated building components is obtained; Based on the real-time thermal radiation flux intensity value, the predicted concentration of the key toxic gas, and the instantaneous instability probability value, a comprehensive dynamic risk quantification value for each of the connected edges is obtained. Based on the comprehensive dynamic risk quantification value, the dynamic risk weight value of each connection edge is obtained through a preset dynamic weight mapping function.

7. The method according to claim 1, characterized in that, The process of generating an initial static access network diagram based on the indoor vector map and the fire resource location information includes: Based on the building structure parameters of the indoor vector map, channel nodes and connectivity relationships are extracted to construct a basic topology network; Based on the fire resource location information, resource attribute tags are added to the corresponding channel nodes in the basic topology network; Based on the channel nodes, and in conjunction with the preset building access priority rules, static access cost initial values ​​are assigned to the edges corresponding to each connectivity relationship in the basic topology network. By removing invalid connectivity relationships with structural obstacles from the basic topology network, an initial static access network graph is obtained.

8. A building fire internal rescue route planning device, characterized in that, The device includes: The multi-source data acquisition module is used to acquire indoor vector maps of the target building, fire resource location information, and multi-dimensional fire environment data; the multi-dimensional fire environment data includes at least one of the following: fire source thermal imaging distribution data, multi-story smoke concentration gradient data, and building structure monitoring data. An initial network construction module is used to generate an initial static access network diagram based on the indoor vector map and the fire resource location information; The dynamic risk quantification module is used to perform dynamic risk quantification assessment on the passage cost of each connecting edge in the initial static passage network graph based on the multi-dimensional fire scene environment data, and obtain the dynamic risk weight value of each connecting edge. The dynamic network generation module is used to generate a dynamic risk access network diagram based on the dynamic risk weight value. The path planning module is used to obtain the internal rescue path for a building fire based on the dynamic risk access network diagram.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images