STDE-based emergency evacuation path generation method and system

Abstract

The invention relates to an emergency evacuation path generation method and system based on STDE, and the method comprises the following steps: extracting a topological network structure of a target building, and collecting smoke detection data and crowd density; building a building topology layer; obtaining a room grid layer according to the building topology layer; constructing a smoke risk value; constructing a congestion risk value; performing dynamic risk fusion on the smoke risk value and the congestion risk value to obtain a total risk; constructing a cost function from each node to an edge of an adjacent node in the building topology layer, and obtaining an optimal path of the building topology layer through a space-time joint reverse search algorithm; performing path planning on each room in the optimal path of the building topology layer in the corresponding room grid layer to obtain the optimal path of the room grid layer; and integrating the optimal paths of the building topology layer and the room grid layer to obtain a final emergency evacuation path. Compared with the prior art, the accuracy of evacuation risk assessment is improved, and different evacuation strategy requirements can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of building fire escape navigation, and in particular to a method and system for generating an emergency evacuation path based on STDE. Background Art

[0002] In the field of building fire escape navigation, existing technologies mainly focus on three aspects: static path planning, real-time monitoring navigation, and disaster propagation prediction models. These technologies have improved the efficiency of escape guidance to a certain extent, but they still have significant limitations.

[0003] Static path planning methods pre-generate fixed escape routes based on building floor plans and use fire signs to guide people. Their advantages lie in their simplicity and low cost. However, these methods rely entirely on pre-set information and are unable to respond to dynamic threats such as fire spread and structural collapse, resulting in poor reliability in actual fire scenarios.

[0004] Real-time monitoring navigation uses deployed temperature and smoke sensors to detect hazardous areas in real time and calculates the shortest path using algorithms such as Dijkstra. Compared to static path planning, this approach offers improved dynamism. However, these methods often overlook the nonlinear effect of crowd density on movement speed, resulting in the calculated "shortest path" not being the "fastest path."

[0005] Disaster propagation prediction models apply principles such as fluid dynamics to simulate the spread of fires, predicting future hazardous areas. While these models are theoretically forward-looking, they are computationally complex (a single simulation often takes minutes), making them difficult to meet the timeliness requirements of real-time path updates.

[0006] The above three technical solutions all have the following major defects in practical applications:

[0007] (1) Spatiotemporal decoupling problem: Existing solutions generally fail to effectively couple the spatial structural characteristics of buildings (such as fire compartments) with temporal evolution parameters such as fire spread and crowd congestion. This results in inaccurate path risk assessment, and the guided escape path may not be optimal or even pose safety hazards.

[0008] (2) Lack of multi-source information fusion: Existing methods usually rely on a single type of sensor data (such as temperature or smoke) and fail to effectively integrate multi-source heterogeneous information from video surveillance, personnel positioning, building structure, etc. This limits the comprehensive perception of the fire situation and affects the reliability of path planning.

[0009] (3) Rough risk quantification: Most existing methods use a binary approach to risk determination (i.e., simply labeling areas as “safe” or “dangerous”), failing to establish a fine-grained risk field model. This coarse risk quantification approach limits the accuracy of path planning and makes it impossible to achieve optimal path search based on risk gradients.

[0010] For example, patent application CN119150573A discloses a 3D-based building safety visualization emergency response method and system. This method simulates the 3D dynamic airflow diffusion distribution of smoke; integrates multiple risk factors to assess the harmfulness of smoke, combining the assessment results with the dynamic airflow diffusion distribution simulation results to generate a real-time 3D risk map; based on the real-time 3D risk map's customized views and risk warning level coloring, multi-dimensional warning annotation is performed on the risk areas along the smoke diffusion path; and based on the multi-dimensionally labeled real-time 3D risk map, evacuation routes are dynamically planned. This method relies solely on smoke data and only uses risk warning level coloring for warning annotation, resulting in low path planning accuracy. Summary of the Invention

[0011] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide an emergency evacuation path generation method and system based on STDE, which improves the accuracy of evacuation risk assessment and can meet the needs of different evacuation strategies.

[0012] The purpose of the present invention can be achieved by the following technical solutions:

[0013] A method for generating an emergency evacuation path based on STDE includes the following steps:

[0014] Extracting the topological network structure of the target building from the building information model of the target building, and collecting smoke detection data and crowd density in the target building in real time;

[0015] Constructing a building topology layer according to the topological network structure of the target building;

[0016] Finely dividing the internal space of each node in the building topology layer to form a room grid layer;

[0017] constructing a smoke risk value based on the smoke detection data;

[0018] Constructing a congestion risk value based on the crowd density;

[0019] Dynamically merging the smoke risk value and the congestion risk value to obtain the total risk of each edge in the building topology layer;

[0020] Obtaining a cost function of an edge from each node in the building topology layer to its adjacent node according to the total risk of each edge in the building topology layer;

[0021] Obtaining the optimal path of the building topology layer through a spatiotemporal joint reverse search algorithm according to the cost function;

[0022] Performing path planning on each room in the optimal path of the building topology layer at its corresponding room grid layer using the A* algorithm to obtain the optimal path of the room grid layer;

[0023] The optimal path of the building topology layer and the optimal path of the room grid layer are integrated to obtain the final emergency evacuation path.

[0024] Furthermore, the topological network structure of the target building includes geometric information and connection relationships of various elements such as rooms, doors, corridors and stairs.

[0025] Furthermore, the size of each grid in the room grid layer is:

[0026]

[0027] Where GridSize(n) is the size of the room grid, A(n) is the area of ​​the room corresponding to node n, and N evacuate (n) is the number of valid evacuation signs in room n, and n is a node in the building topology layer.

[0028] Furthermore, the total risk of each edge in the building topology layer is:

[0029] R total (e) = μ·R smoke (e)+(1-μ)·R crowd (e)

[0030]

[0031] Where R total (e) is the total risk of each edge in the building topology layer, e is the edge of the building topology layer, μ is the risk weight, R smoke (e) is the smoke risk of each edge in the building topology layer, R crowd (e) is the congestion risk of each edge in the building topology layer, w s (s) is the weight of smoke sensor s, w c (c) is the weight of camera c.

[0032] Furthermore, the smoke risk value includes the smoke risk value of a single grid in the room grid layer and the smoke risk value of a single node in the building topology layer;

[0033] The smoke risk value of a single grid in the room grid layer is:

[0034]

[0035] Where R smoke (g,s) is the smoke risk value of grid g affected by smoke detector s, R max is the maximum risk value when the detector is triggered, λ is the attenuation coefficient, d(g,s) is the distance to the center of the smoke detector, D smoke (s) is the digital signal of the smoke detector, r is the effective radius of the detector, g is the grid unit, and s is the smoke detector;

[0036] The smoke risk of a single node in the building topology layer is:

[0037] R smoke (s)=D smoke (s)·R max

[0038] Where R smoke (s) is the smoke risk value of the building topology node affected by the smoke detector s, D smoke (s) is the digital signal of the smoke detector, R max is the maximum risk value when the detector is triggered, and s is the smoke detector.

[0039] Furthermore, the congestion risk value of a single node on each edge in the building topology layer is:

[0040]

[0041] Where R crowd (c) is the congestion risk value within the video surveillance range c, ρ(c) is the crowd density, R max is the maximum congestion risk value, ρ low is the low density threshold, ρ medium is the medium density threshold, ρ high is the high density threshold, N detect is the number of people identified by the target monitoring algorithm, A ROI is the actual area of ​​the monitoring area, k is the slope parameter of the curve, F texture is the image texture complexity feature value, and F0 is the empirical threshold.

[0042] Furthermore, the specific steps of generating the optimal path of the building topology layer by using the spatiotemporal joint reverse search algorithm are:

[0043] Taking the exit node of the target building as the starting point, construct a breadth-first reverse graph;

[0044] During the reverse search process, time constraint pruning is performed. If the minimum remaining safe time is less than 0, the expansion is stopped;

[0045] Attempt to bypass nodes that cannot be escaped within the safe time, giving priority to traversing nodes adjacent to the reverse graph nodes;

[0046] The optimal predecessor node and arrival time of each node are recorded through the backtracking mechanism, and then the optimal path of the building topology layer is generated.

[0047] Furthermore, the cost function of the edge from each node to its adjacent node in the building topology layer is:

[0048]

[0049] Where C e is the cost function, α and β are weight coefficients, L(e) is the length of edge e, v max is the maximum speed, R total (e) is the total risk value of edge e at the current moment.

[0050] Furthermore, the minimum remaining safety time is:

[0051] T Margin =min(∑Δt-T danger (n))

[0052] Where, T Margin is the minimum remaining safety time, Δt is the moving time, T danger (n) is the danger time threshold of each node, which is obtained by accumulating the smoke propagation delay along the path from the smoke source to the target node. The smoke propagation delay is:

[0053]

[0054] Where, T prop (e) is the smoke propagation delay, L(e) is the length of edge e, v base is the basic diffusion speed, Δh(e) is the absolute value of the vertical height difference between the nodes connected by edge e, and k v is the vertical acceleration coefficient, L(e) is the length of side e, A eff (e) is the effective flow area of ​​side e.

[0055] According to another aspect of the present invention, there is provided a system for generating an emergency evacuation route based on STDE, comprising:

[0056] A data acquisition module is used to extract the topological network structure of the target building from the building information model of the target building, and to collect smoke detection data and crowd density in the target building in real time;

[0057] A building topology layer construction module, configured to construct a building topology layer according to the topological network structure of the target building;

[0058] A room grid layer construction module is used to finely divide the internal space of each node in the building topology layer to form a room grid layer;

[0059] a smoke risk value acquisition module, configured to construct a smoke risk value based on the smoke detection data and the room grid layer;

[0060] a congestion risk value acquisition module, configured to construct a congestion risk value according to the crowd density;

[0061] A dynamic risk fusion module, configured to dynamically fuse the smoke risk value and the congestion risk value to obtain a total risk of each edge in the building topology layer;

[0062] A cost function construction module is used to obtain a cost function of an edge from each node in the building topology layer to its adjacent node according to the total risk of each edge in the building topology layer;

[0063] A building topology layer optimal path acquisition module, configured to obtain the optimal path of the building topology layer through a spatiotemporal joint reverse search algorithm according to the cost function;

[0064] The room grid layer optimal path acquisition module is used to perform path planning on each room in the optimal path of the building topology layer at its corresponding room grid layer using the A* algorithm to obtain the optimal path of the room grid layer;

[0065] The optimal path integration module is used to integrate the optimal path of the building topology layer and the optimal path of the room grid layer to obtain the final emergency evacuation path.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. This invention improves the accuracy and robustness of environmental perception by integrating multi-source heterogeneous data from BIM models, smoke detectors, and video surveillance. It also introduces smoke risk and congestion risk values, performs spatiotemporal co-encoding of building space information and dynamic risk information (smoke, congestion), and constructs a path description model, enabling a more comprehensive and accurate assessment of evacuation risks.

[0068] 2. This paper designs smoke risk values, congestion risk values, and a dynamic risk fusion method that can reflect the dynamic changes of the environment in real time. By combining the building topology layer and the room grid layer, it takes into account the propagation delay of smoke and sets a dangerous time threshold for each node to ensure that the generated evacuation path is reachable within the safe time. It realizes the unification of macroscopic path planning and microscopic obstacle avoidance, and meets the requirements of different evacuation strategies through a risk-adjustable cost function. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1This is a flow chart of a method for generating an emergency evacuation route based on STDE proposed in the present invention;

[0070] Figure 2 This is a schematic diagram of the STDE-based emergency evacuation path generation method proposed in the present invention. DETAILED DESCRIPTION

[0071] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0072] Abbreviations involved:

[0073] Spatial-Temporal Dynamic Encoding (STDE)

[0074] Building Information Modeling (BIM)

[0075] Example 1

[0076] This embodiment provides a method for generating an emergency evacuation path based on STDE. Figure 1 As shown, the following steps are included:

[0077] S1. Extract the topological network structure of the target building from the building information model of the target building, and collect smoke detection data and crowd density in the target building in real time.

[0078] This embodiment takes a two-story building as an example, combined with Figure 2 The schematic diagram of the STDE-based emergency evacuation path generation method is shown to generate the emergency evacuation path.

[0079] Extract the target building's topological network structure from the Building Information Model (BIM), including the geometric information and connectivity of various elements such as rooms, doors, corridors, and staircases. For example, the first floor of a two-story building contains Room A, Corridor 1, and a stairwell, while the second floor contains Room B, Corridor 2, and a stairwell, with the stairwell connecting the first and second floors.

[0080] Real-time collection and fusion of multi-source sensor data, including digital signals from point-type smoke detectors and crowd density estimated from video surveillance data.

[0081] Collect smoke detector data in real time. Assume that detector s1 in room A is triggered at time t=0, then the digital signal provided by detector s1 is: D smoke (s1)=1.

[0082] Collect video surveillance data in real time and estimate the crowd density in each area through image analysis. The crowd density is:

[0083]

[0084] Where ρ(c) is the crowd density, N detect is the number of people identified by the target monitoring algorithm, A ROI is the actual area of ​​the monitoring area, k is the slope parameter of the curve, F texture is the image texture complexity feature value, and F0 is the empirical threshold.

[0085] S2. Construct a building topology layer based on the topological network structure of the target building, and finely divide the internal space of each node in the building topology layer to form a room grid layer.

[0086] The building topology layer describes the macroscopic spatial structure and connectivity of a building. Nodes represent key locations (rooms, exits, stairwells, etc.), and edges represent physical pathways. In this example, the nodes are defined as Room A, Corridor 1, Stairwell (First Floor), Stairwell (Second Floor), Corridor 2, Room B, and Exit. Edges are defined as the physical pathways connecting these nodes.

[0087] The internal space of each node in the building topology layer is finely divided to form a grid. The size of each grid in the room grid layer is:

[0088]

[0089] Where GridSize(n) is the size of the room grid, A(n) is the area of ​​the room corresponding to node n, and N evacuate (n) is the number of valid evacuation signs in room n, and n is a node in the building topology layer.

[0090] Divide Room A, Room B, and Corridors 1 and 2 into adaptive grids. For example, the area of ​​Room A is 50m 2 , the number of effective evacuation signs is 2, and the grid size is about 5m.

[0091] S3. Construct a smoke risk value based on the smoke detection data.

[0092] The smoke risk value is constructed based on smoke detection data. The smoke risk value includes the smoke risk value of a single grid in the room grid layer and the smoke risk value of a single node in the building topology layer.

[0093] The smoke risk value of a single grid in the room grid layer is:

[0094]

[0095] Where Rsmoke (g,s) is the smoke risk value of grid g affected by smoke detector s, R max is the maximum risk value when the detector is triggered, λ is the attenuation coefficient, d(g,s) is the distance to the center of the smoke detector, D smoke (s) is the digital signal of the smoke detector, r is the effective radius of the detector, g is the grid unit, and s is the smoke detector;

[0096] The smoke risk of a single node in the building topology layer is:

[0097] R smoke (s)=D smoke (s)·R max

[0098] Where R smoke (s) is the smoke risk value of the building topology node affected by the smoke detector s, D smoke (s) is the digital signal of the smoke detector, R max is the maximum risk value when the detector is triggered, and s is the smoke detector.

[0099] Since detector s1 is triggered in room A, the smoke risk value within the grid of room A is calculated according to the formula. The risk value decreases as the distance from the center of the detector increases. At the same time, the building topology node corresponding to room A also has a high smoke risk value.

[0100] S4. Construct a congestion risk value based on crowd density.

[0101] The congestion risk value is constructed based on the crowd density estimated through video surveillance data. The congestion risk value of a single node on each edge in the building topology layer is:

[0102]

[0103] Where R crowd (c) is the congestion risk value within the video surveillance range c, ρ(c) is the crowd density, R max is the maximum congestion risk value, ρ low is the low density threshold, ρ medium is the medium density threshold, ρ high is the high density threshold.

[0104] Assume R max =1,ρ low =0.5 people / m 2 , ρ medium =1.5 people / m 2 , ρ high =3 people / m 2 Since the crowd density in corridor 1 is ρ(c1) = 2 people / m 2, calculate its risk value according to the formula:

[0105]

[0106] Therefore, the building topology node corresponding to corridor 1 has a congestion risk value of approximately 0.667.

[0107] S5. Dynamically fuse the smoke risk value and the congestion risk value to obtain the total risk of each edge in the building topology layer.

[0108] Calculate the total risk of each edge connecting each topological node in the building topology layer:

[0109] R total (e) = μ·R smoke (e)+(1-μ)·R crowd (e)

[0110]

[0111] Where R total (e) is the total risk of each edge in the building topology layer, e is the edge of the building topology layer, μ is the risk weight, R smoke (e) is the smoke risk of each edge in the building topology layer, R crowd (e) is the congestion risk of each edge in the building topology layer, w s (s) is the weight of smoke sensor s, w c (c) is the weight of camera c.

[0112] For example, when calculating the risk of edge e1 connecting room A and corridor 1, it is necessary to consider the information of sensors deployed in room A and corridor 1, build a reference path connecting the centers of the two nodes, calculate the sensor weights, and finally integrate the smoke risk and congestion risk to obtain R total (e1).

[0113] S6. Obtain the cost function of the edge from each node in the building topology layer to its adjacent nodes according to the total risk of each edge in the building topology layer.

[0114] Time dimension encoding: Calculate the smoke propagation delay of the edge connecting each topological node. For example, calculate the propagation delay T from room A to corridor 1 prop (e1), we need to consider their distance, height difference and effective flow area of ​​the channel. The smoke propagation delay is:

[0115]

[0116] Where, T prop (e) is the smoke propagation delay, L(e) is the length of edge e, v baseis the basic diffusion speed, Δh(e) is the absolute value of the vertical height difference between the nodes connected by edge e, and k v is the vertical acceleration coefficient, L(e) is the length of side e, A eff (e) is the effective flow area of ​​side e.

[0117] Starting from the smoke source room A, the propagation delay is accumulated along the building topology to calculate the danger time threshold of each node. For example, the danger time threshold T for corridor 1 is danger (Corridor 1) is equal to the propagation delay from room A to corridor 1.

[0118] The cost function of the edge from each node to its adjacent nodes in the building topology layer is:

[0119]

[0120] Where C e is the cost function, α and β are weight coefficients, L(e) is the length of edge e, v max is the maximum speed, R total (e) is the total risk value of edge e at the current moment.

[0121] In this embodiment, α=1 and β=0.5, which means that time efficiency is more emphasized, but risks must also be considered.

[0122] S7. Obtain the optimal path of the building topology layer through the spatiotemporal joint reverse search algorithm according to the cost function.

[0123] The specific steps of generating the optimal path of the building topology layer through the spatiotemporal joint reverse search algorithm are as follows:

[0124] Taking the exit node of the target building as the starting point, construct a breadth-first reverse graph;

[0125] During the reverse search process, time constraint pruning is performed. If the minimum remaining safe time is less than 0, the expansion is stopped;

[0126] Attempt to bypass nodes that cannot be escaped within the safe time, reduce the β weight, and give priority to traversing nodes adjacent to the reverse graph nodes;

[0127] The backtracking mechanism records the optimal predecessor node and arrival time of each node, thereby generating the optimal path at the building topology level. For example, suppose the optimal path from room B to the exit is: room B → corridor 2 → stairwell (second floor) → stairwell (first floor) → corridor 1 → exit.

[0128] The minimum remaining safety time is:

[0129] T Margin =min(∑Δt-T danger (n))

[0130] Where, T Margin is the minimum remaining safety time, Δt is the moving time, T danger (n) is the danger time threshold of each node, which is obtained by accumulating the smoke propagation delay along the path from the smoke source to the target node.

[0131] S8. Perform path planning on each room in the optimal path of the building topology layer at its corresponding room grid layer using the A* algorithm to obtain the optimal path of the room grid layer.

[0132] The A* algorithm's cost function considers the smoke risk of each cell in the room's grid layer and the heuristic distance from the current cell to the target cell. The grid-level path planning results guide evacuees' specific movement within the room, for example, avoiding areas with high smoke concentrations and directing them to the next topological node.

[0133] For each room in the path (Room B), the A* algorithm is used to perform path planning at the corresponding grid level. The A* algorithm considers the smoke risk distribution within Room B and guides evacuees away from high-risk areas, ultimately reaching Room B's exit (the door connecting to Corridor 2). Similarly, if the path passes through a large corridor, A* path planning can also be performed at the corridor's grid level to provide more precise evacuation guidance.

[0134] S9. Integrate the optimal path of the building topology layer and the optimal path of the room grid layer to obtain the final emergency evacuation path.

[0135] The path planning results of the topology layer and the grid layer are integrated to generate the final emergency evacuation path. This path will indicate the evacuation sign in room B to guide people to move towards the exit of room B, then indicate the evacuation sign in corridor 2 to guide people to move towards the stairwell, and so on, until they reach the final exit.

[0136] Example 2

[0137] This embodiment provides an emergency evacuation route generation system based on STDE, including:

[0138] A data acquisition module is used to extract the topological network structure of the target building from the building information model of the target building and collect smoke detection data and crowd density in the target building in real time;

[0139] A building topology layer construction module is used to construct a building topology layer according to the topological network structure of the target building;

[0140] The room grid layer construction module is used to finely divide the internal space of each node in the building topology layer to form a room grid layer;

[0141] Smoke risk value acquisition module, used to construct smoke risk value based on smoke detection data and room grid layer;

[0142] Congestion risk value acquisition module, used to construct congestion risk value according to crowd density;

[0143] Dynamic risk fusion module, used to dynamically fuse smoke risk value and congestion risk value to obtain the total risk of each edge in the building topology layer;

[0144] A cost function construction module is used to obtain a cost function of an edge from each node in the building topology layer to its adjacent nodes according to the total risk of each edge in the building topology layer;

[0145] The optimal path acquisition module of the building topology layer is used to obtain the optimal path of the building topology layer through the time-space joint reverse search algorithm according to the cost function;

[0146] The room grid layer optimal path acquisition module is used to plan the path of each room in the optimal path of the building topology layer in its corresponding room grid layer through the A* algorithm to obtain the optimal path of the room grid layer;

[0147] The optimal path integration module is used to integrate the optimal path of the building topology layer and the optimal path of the room grid layer to obtain the final emergency evacuation path.

[0148] The rest is the same as Example 1.

[0149] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A method for generating emergency evacuation paths based on STDE, characterized in that: The following steps are involved: Extracting the topological network structure of the target building from the building information model of the target building, and collecting smoke detection data and crowd density in the target building in real time; Constructing a building topology layer according to the topological network structure of the target building; Finely dividing the internal space of each node in the building topology layer to form a room grid layer; constructing a smoke risk value based on the smoke detection data; Constructing a congestion risk value based on the crowd density; Dynamically merging the smoke risk value and the congestion risk value to obtain the total risk of each edge in the building topology layer; Obtaining a cost function of an edge from each node in the building topology layer to its adjacent node according to the total risk of each edge in the building topology layer; Obtaining the optimal path of the building topology layer through a spatiotemporal joint reverse search algorithm according to the cost function; Performing path planning on each room in the optimal path of the building topology layer at its corresponding room grid layer using the A* algorithm to obtain the optimal path of the room grid layer; The optimal path of the building topology layer and the optimal path of the room grid layer are integrated to obtain the final emergency evacuation path.

2. The method for generating an emergency evacuation path based on STDE according to claim 1, characterized in that: The topological network structure of the target building includes geometric information and connection relationships of various elements such as rooms, doors, corridors and stairs.

3. The STDE-based emergency evacuation path generation method according to claim 1, characterized in that: The size of each grid in the room grid layer is: Where GridSize(n) is the size of the room grid, A(n) is the area of ​​the room corresponding to node n, and N evacuate (n) is the number of valid evacuation signs in room n, and n is a node in the building topology layer.

4. The method for generating an emergency evacuation path based on STDE according to claim 1, characterized in that: The total risk of each edge in the building topology layer is: R total (e)6µ·R smoke (e)+(1-μ)·R crowd (e) Where R total (e) is the total risk of each edge in the building topology layer, e is the edge of the building topology layer, μ is the risk weight, R smoke (e) is the smoke risk of each edge in the building topology layer, R crowd (e) is the congestion risk of each edge in the building topology layer, w s (s) is the weight of smoke sensor s, w c (c) is the weight of camera c.

5. The method for generating an emergency evacuation path based on STDE according to claim 1, characterized in that: The smoke risk value includes the smoke risk value of a single grid in the room grid layer and the smoke risk value of a single node in the building topology layer; The smoke risk value of a single grid in the room grid layer is: Where R smoke (g,s) is the smoke risk value of grid g affected by smoke detector s, R max is the maximum risk value when the detector is triggered, λ is the attenuation coefficient, d(g,s) is the distance to the center of the smoke detector, D smoke (s) is the digital signal of the smoke detector, r is the effective radius of the detector, g is the grid unit, and s is the smoke detector; The smoke risk of a single node in the building topology layer is: R smoke (s)=D smoke (s)·R max Where R smoke (s) is the smoke risk value of the building topology node affected by the smoke detector s, D smoke (s) is the digital signal of the smoke detector, R max is the maximum risk value when the detector is triggered, and s is the smoke detector.

6. The method for generating an emergency evacuation path based on STDE according to claim 1, characterized in that: The congestion risk value of a single node on each edge in the building topology layer is: Where R crowd (c) is the congestion risk value within the video surveillance range c, ρ(c) is the crowd density, R max is the maximum congestion risk value, ρ low is the low density threshold, ρ medium is the medium density threshold, ρ high is the high density threshold, N detect is the number of people identified by the target monitoring algorithm, A ROI is the actual area of ​​the monitoring area, k is the slope parameter of the curve, F texture is the image texture complexity feature value, and F0 is the empirical threshold.

7. The method for generating an emergency evacuation path based on STDE according to claim 1, characterized in that: The specific steps of generating the optimal path of the building topology layer by the spatiotemporal joint reverse search algorithm are as follows: Taking the exit node of the target building as the starting point, construct a breadth-first reverse graph; During the reverse search process, time constraint pruning is performed. If the minimum remaining safe time is less than 0, the expansion is stopped; Attempt to bypass nodes that cannot be escaped within the safe time, giving priority to traversing nodes adjacent to the reverse graph nodes; The optimal predecessor node and arrival time of each node are recorded through the backtracking mechanism, and then the optimal path of the building topology layer is generated.

8. The method for generating an emergency evacuation path based on STDE according to claim 7, characterized in that: The cost function of the edge from each node to its adjacent node in the building topology layer is: Where C e is the cost function, α and β are weight coefficients, L(e) is the length of edge e, v max is the maximum speed, R total (e) is the total risk value of edge e at the current moment.

9. The method for generating an emergency evacuation route based on STDE according to claim 7, characterized in that: The minimum remaining safety time is: T Margin =min(∑Δt-T danger (n)) Where, T Margin is the minimum remaining safety time, Δt is the moving time, T danger (n) is the danger time threshold of each node, which is obtained by accumulating the smoke propagation delay along the path from the smoke source to the target node. The smoke propagation delay is: Where, T prop (e) is the smoke propagation delay, L(e) is the length of edge e, v base is the basic diffusion speed, Δh(e) is the absolute value of the vertical height difference between the nodes connected by edge e, and k v is the vertical acceleration coefficient, L(e) is the length of side e, A eff (e) is the effective flow area of ​​side e.

10. An emergency evacuation route generation system based on STDE, characterized in that: include: A data acquisition module is used to extract the topological network structure of the target building from the building information model of the target building, and to collect smoke detection data and crowd density in the target building in real time; A building topology layer construction module, configured to construct a building topology layer according to the topological network structure of the target building; A room grid layer construction module is used to finely divide the internal space of each node in the building topology layer to form a room grid layer; a smoke risk value acquisition module, configured to construct a smoke risk value based on the smoke detection data and the room grid layer; a congestion risk value acquisition module, configured to construct a congestion risk value according to the crowd density; A dynamic risk fusion module, configured to dynamically fuse the smoke risk value and the congestion risk value to obtain a total risk of each edge in the building topology layer; A cost function construction module is used to obtain a cost function of an edge from each node in the building topology layer to its adjacent node according to the total risk of each edge in the building topology layer; A building topology layer optimal path acquisition module, configured to obtain the optimal path of the building topology layer through a spatiotemporal joint reverse search algorithm according to the cost function; The room grid layer optimal path acquisition module is used to perform path planning on each room in the optimal path of the building topology layer at its corresponding room grid layer using the A* algorithm to obtain the optimal path of the room grid layer; The optimal path integration module is used to integrate the optimal path of the building topology layer and the optimal path of the room grid layer to obtain the final emergency evacuation path.

Images