Intelligent monitoring method and system for fire control room

Abstract

The application discloses an intelligent monitoring method and system for a fire control room, and particularly relates to the technical field of fire safety monitoring. Firstly, the device static operation point coordinate set and the personnel position coordinates are acquired through positioning collection technology, and the space-time graph containing the personnel node and the device operation node is constructed based on the device static operation point coordinate set and the personnel position coordinates. The real-time obstacle information is acquired through the target temporary obstacle detection model, the compliance road network path set is updated, and the global shortest reachable time from the personnel node to each device is calculated. Finally, the corresponding color classification early warning signal is generated based on the comprehensive delay risk index and output to the alarm device of the fire control. The space-time graph containing the personnel node and the device operation node is constructed, the data accuracy is improved, the reachable edge and the passing time are updated in real time, the emergency passing scene is truly restored, and the emergency response time calculation is ensured to be in line with the actual situation.

Description

Technical Field

[0001] This invention relates to the field of fire safety monitoring technology, specifically to an intelligent monitoring method and system for fire control rooms. Background Technology

[0002] The fire control room is the core command center of the automatic fire alarm system and the fire linkage control system. It undertakes key functions such as receiving fire alarms, controlling fire equipment linkage, and commanding emergency response. Its emergency response efficiency and handling accuracy directly determine the effectiveness of fire prevention and control and the safety of people and property.

[0003] Currently, fire alarm system monitoring technology in fire control rooms mainly focuses on passive monitoring of fire alarm signals and equipment operating status, and can only achieve basic status acquisition of "whether there is an alarm signal". The emergency path planning of fire emergency response personnel to manually operated key fire equipment in traditional fire alarm systems relies on static layout drawings of the control room and does not consider the impact of temporary obstacles (such as stacked items) on passage time. Existing fire alarm monitoring systems are mostly one-way processes of "data acquisition - status display" and have not achieved closed-loop intelligent monitoring of "perception - modeling - calculation - evaluation - early warning".

[0004] However, the emergency response efficiency of the fire control room directly determines the effectiveness of fire suppression. Existing monitoring technologies still have shortcomings. For example, basic status data collection does not integrate key information such as real-time personnel location, obstacles, and travel time, making it difficult to assess the actual emergency response capabilities of personnel reaching critical operating equipment. Static layout drawings result in a serious disconnect between preset paths and actual personnel travel scenarios, failing to reflect the actual reach time of personnel to equipment, which can easily lead to delays in fire alarm response and weaken the emergency response effectiveness of the fire alarm system. Existing technologies lack a quantitative emergency response effectiveness assessment mechanism and have not established a differentiated risk classification mechanism, making it impossible to proactively push graded early warning signals, causing emergency response in the fire control room to still rely on the experience of on-duty personnel. Therefore, there is an urgent need for an intelligent monitoring method that can integrate multi-source sensing, modeling, quantitative assessment, and proactive early warning to make up for the shortcomings of existing technologies. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an intelligent monitoring method and system for fire control rooms to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring method for a fire control room, comprising: S1: Acquire multi-source monitoring data from the fire control room through positioning and acquisition technology, including the coordinate set of static operating points of equipment and the coordinates of personnel positions; S2: Construct the road network topology using the BIM evacuation route centerline, match road network nodes based on the three-dimensional positions of personnel and equipment, obtain the final matched nodes, and construct a static passage distance matrix; S3: Obtain real-time obstacle information through the target temporary obstacle detection model, update the static travel distance matrix, and generate a dynamic travel distance matrix representing the travel distance of the road network path; S4: Based on the dynamic travel distance matrix and the baseline travel speed of personnel in emergency scenarios, combined with the obstacle travel correction coefficient, calculate the global shortest reach time from personnel nodes to each device; S5: Compare the global shortest reach time with the preset standard response time of the corresponding device, generate a comprehensive delay risk index based on the comparison result, generate a graded warning signal with corresponding color based on the comprehensive delay risk index, and output it to the alarm device of the fire control.

[0007] Preferably, an intelligent monitoring system for a fire control room includes: Fire control room positioning and sensing module: It acquires multi-source monitoring data of the fire control room through positioning and acquisition technology, including the coordinate set of static operating points of equipment and personnel position coordinates; Fire control room road network construction module: Constructs road network topology through BIM evacuation route centerlines, matches road network nodes based on the three-dimensional positions of personnel and equipment, obtains the final matched nodes, and constructs a static passage distance matrix; Fire control room road network update module: By using the target temporary obstacle detection model, real-time obstacle information is obtained, the static travel distance matrix is ​​updated, and a dynamic travel distance matrix representing the travel distance of the road network path is generated; Fire control room spatiotemporal accessibility monitoring module: Based on the dynamic access distance matrix and the baseline access speed of personnel in emergency scenarios, combined with the obstacle access correction coefficient, calculates the global shortest access time from personnel nodes to each device; Fire control room alarm output module: It compares the shortest global reach time with the preset standard response time of the corresponding equipment, generates a comprehensive delay risk index based on the comparison result, generates a graded warning signal with corresponding color based on the comprehensive delay risk index, and outputs it to the alarm device of the fire control.

[0008] The technical effects and advantages of this invention are as follows: 1. This invention constructs a set of road network topology segments based on the centerline of fire evacuation routes in a BIM model, and obtains compliant road network paths from personnel matching nodes to equipment operation matching nodes through a path planning algorithm. Based on real-time updates of the traffic topology and traffic time, it realistically restores emergency traffic scenarios, gets rid of the limitations of static layout drawings, ensures that the emergency response time calculation is consistent with the actual situation on site, and avoids delays in handling due to path misjudgment. 2. This invention constructs a comprehensive delay risk index based on equipment weight and delay time, transforming abstract emergency response capabilities into quantifiable values. It also achieves three-level risk classification (green, yellow, and red) through preset thresholds, proactively pushing graded early warning signals to the alarm devices of the fire control system. This accurately identifies potential and severe delay risks, improving the standardization and proactivity of emergency response in the fire control room. 3. This invention forms a closed-loop process of multi-source perception, modeling, response risk quantification assessment, and proactive early warning. It simultaneously displays information on delayed equipment and personnel locations, providing intuitive data support for emergency dispatch. It transforms passive duty into proactive prevention and control, improves the timeliness and accuracy of emergency response, and comprehensively enhances the fire control room's ability to handle sudden fire incidents and its overall safety level. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the overall process of the present invention.

[0010] Figure 2 This is a schematic diagram of the method flow of the present invention.

[0011] Figure 3 This is a schematic diagram of the dynamic travel distance matrix generation process of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Please see Figure 1 As shown, the present invention provides an intelligent monitoring system for a fire control room, including a fire control room positioning and sensing module, a fire control room road network construction module, a fire control room road network updating module, a fire control room spatiotemporal accessibility monitoring module, and a fire control room alarm output module.

[0014] The fire control room positioning and sensing module is connected to the fire control room road network construction module, the fire control room road network update module is connected to both the fire control room road network construction module and the fire control room spatiotemporal accessibility monitoring module, and the fire control room alarm output module is connected to the fire control room spatiotemporal accessibility monitoring module.

[0015] Fire control room positioning and sensing module: It acquires multi-source monitoring data of the fire control room through positioning and acquisition technology, including the coordinate set of static operating points of equipment and personnel position coordinates; Fire control room road network construction module: Constructs road network topology through BIM evacuation route centerlines, matches road network nodes based on the three-dimensional positions of personnel and equipment, obtains the final matched nodes, and constructs a static passage distance matrix; Fire control room road network update module: By using the target temporary obstacle detection model, real-time obstacle information is obtained, the static travel distance matrix is ​​updated, and a dynamic travel distance matrix representing the travel distance of the road network path is generated; Fire control room spatiotemporal accessibility monitoring module: Based on the dynamic access distance matrix and the baseline access speed of personnel in emergency scenarios, combined with the obstacle access correction coefficient, calculates the global shortest access time from personnel nodes to each device; Fire control room alarm output module: compares the shortest global reach time with the preset standard response time of the corresponding equipment, generates a comprehensive delay risk index based on the comparison result, generates a graded warning signal with corresponding color based on the comprehensive delay risk index, and outputs it to the alarm device of the fire control. Please see Figure 2 As shown, an intelligent monitoring method for a fire control room includes: S1: acquiring multi-source monitoring data of the fire control room through positioning acquisition technology, including the coordinate set of static operating points of equipment and the coordinates of personnel positions; S2: constructing a road network topology through the center line of BIM evacuation channels, matching road network nodes based on the three-dimensional positions of personnel and equipment, obtaining the final matched nodes, and constructing a static travel distance matrix; S3: acquiring real-time obstacle information through a target temporary obstacle detection model, updating the static travel distance matrix, and generating a dynamic travel distance matrix representing the travel distance of road network paths; S4: calculating the global shortest reachable time from personnel nodes to each piece of equipment based on the dynamic travel distance matrix and the baseline travel speed of personnel in emergency scenarios, combined with obstacle travel correction coefficients; S5: comparing the global shortest reachable time with the preset standard response time of the corresponding equipment, generating a comprehensive delay risk index based on the comparison result, generating a graded warning signal with corresponding color based on the comprehensive delay risk index, and outputting it to the alarm device of the fire control system.

[0016] S1: Acquire multi-source monitoring data from the fire control room using location acquisition technology, including the coordinate set of static operating points of equipment and personnel location coordinates, including: S1.1: Extract the 3D coordinates (X, Y, F, Z) of the operation points of all critical fire-fighting equipment from the BIM model of the fire control room. k ,Y k Z k For situations where the same equipment has multiple operating points (such as manual buttons on both sides of a fireproof roller shutter), the coordinates of each operating point are extracted. Let there be K key operating points for the equipment; then the set of static operating point coordinates is D, where D = {D...} k =(X k ,Y kZ k The key fire-fighting equipment refers to equipment that meets any of the following conditions during a fire emergency response: equipment requiring on-site confirmation of remote start status (equipment that can be remotely started from the fire control room, but requires on-site confirmation or manual forced start due to reasons such as the on-site control cabinet being in manual mode or line faults, including but not limited to fire pump control cabinets, smoke-proof fan control cabinets, and smoke exhaust fan control cabinets); equipment that must be physically operated on-site (including but not limited to gas extinguishing emergency stop buttons, gas extinguishing emergency start buttons, fireproof roller shutter manual control buttons, and fire damper manual reset devices); and equipment used to confirm the authenticity of a fire alarm on-site (including but not limited to manual fire alarm buttons and fire telephone jacks). S1.2: Pre-deploy n UWB positioning base stations (e.g., in evacuation routes, stairwells, fire control rooms, etc. on each floor of the building, at intervals of 5-8 meters, typically 4-6 base stations per floor), and integrate an IMU (three-axis accelerometer and three-axis gyroscope) into the UWB tag, with base station location coordinates as B1, B2, ..., B n Each base station is located at B. b =(x b ,y b ,z b x, y, and z are the coordinates in the x, y, and z directions, respectively. All personnel involved in the fire emergency response wear UWB positioning tags (which can be integrated into work badges or safety helmets). The positioning tags periodically (e.g., 10Hz) transmit pulse signals to the base station, and each base station records the arrival time t of the pulse signal. b (b=1,2,...,n-1), the arrival time difference between the b-th base station and the reference base station (e.g., B1) is ∆t. b1 ∆t b1 =t b -t1, the corresponding distance difference is ∆d b1 ,∆d b1 =c×∆t b1 c is the speed of light; the TDOA positioning algorithm is used, based on ∆t b1 ∆d b1 and B b Establish the TDOA positioning equation system and calculate the initial estimate P of the personnel's position. t 0 =(x t 0 ,y t 0 ,z t 0 ); In this embodiment, it is necessary to specifically explain that the TDOA positioning algorithm is a well-known technology. It is a positioning method that uses time difference. By comparing the absolute time difference of the signal arriving at each monitoring station, a hyperbola with the monitoring station as the focus and the distance difference as the major axis can be drawn. The intersection of the hyperbola is the position of the signal. For the TDOA positioning equation system, the Chan algorithm can be used to linearize the nonlinear equation system, and the initial estimate of the personnel's position can be obtained by solving it using the least squares method.

[0017] S1.3: An adaptive extended Kalman filter (EKF) is used to fuse the initial position estimate from the UWB with the IMU inertial measurement data to obtain a smooth and continuous real-time position of the personnel. A state transition model X is defined. t|t-1 =F×X t-1 +w(t) and observation model Z t|t-1 =H×X t-1 +W(t), F and H are the state transition matrix and observation matrix, respectively, and w(t) and W(t) are the process noise and observation noise, respectively. The state transition matrix is ​​designed as a 6×6 matrix, X t Let X be the state vector. t =[x t ,y t ,z t ,v x,t ,v y,t ,v z,t ] T , (x t ,y t ,z t ) and (v x,t ,v y,t ,v z,t The three-dimensional position and three-dimensional velocity of the person at time t are respectively given by P. The velocity component v in each direction is calculated by central difference, v = (P t+1 -P t-1 ) / 2Δt, where Δt is the sampling interval (e.g., 0.1s), the observation matrix is ​​a 3×6 matrix, and the initial estimate of the personnel position is used as the observation vector [x t 0 ,y t 0 ,z t 0 ] T The noise covariance matrix Q is dynamically adjusted based on the real-time measurement variance of the IMU accelerometer. The IMU acceleration variance characterizes the intensity of the person's movement; a larger variance (more intense movement) increases the Q matrix, improving the filter's ability to track changes in movement; a smaller variance (smoother movement) decreases the Q matrix, suppressing noise interference. An adaptive extended Kalman filter prediction step and update step are executed to obtain the three-dimensional coordinates P of the person in the UWB coordinate system at time t. tUWB P t UWB =(x t UWB ,y t UWB ,z t UWB Using a pre-calibrated coordinate transformation matrix, the coordinates in the UWB coordinate system are transformed to the unified global coordinate system of the BIM model, and the final personnel position coordinates P are output. t =(x t ,y t ,z t ); This embodiment requires specific explanation of the state transition matrix: A block diagonal matrix is ​​constructed based on the sampling interval Δt to obtain the state transition matrix F. I3 is a 3×3 identity matrix, and 03 is a 3×3 zero matrix; observation matrix The process noise w(t) and the observation noise W(t) are both zero-mean Gaussian white noise, with covariance matrices Q and R, respectively, i.e., w(t) ~ N(0,Q) and V(t) ~ N(0,R), where Q is constructed based on personnel acceleration. The system acquires three-dimensional instantaneous measurements from the IMU accelerometer. To eliminate gravity components and random noise, the instantaneous measurements are first subtracted from the gravitational acceleration. Then, the data is passed through a low-pass filter to calculate the preprocessed three-dimensional acceleration (α). x,t ,a y,t ,a z,t Amplitude a t , Within a sliding window of length W1 (W takes values ​​of 10-20, corresponding to a 1-2 second time window), the sample variance of the acceleration is calculated as the initial value σ. a0,t 2 The initial value is corrected by an adaptive adjustment formula, and the final output is σ after clamp constraint. a,t 2 One possible implementation is based on the innovation vector V t =Z t -H×X t|t-1 g X t|t-1 g To estimate the state at time t based on the state vector up to time t-1, the acceleration variance σ in the process noise covariance matrix is ​​adaptively adjusted. a,t 2 Calculate the total deviation energy e t =||V t || 2 ,||V t || 2Calculate the trace s of the theoretical innovation covariance matrix, given the square of the L2 norm of the innovation vector. t s t =trace(H×P t|t-1 ×H T +R), P t|t-1 For the prediction error covariance, σ a,t 2 =clamp(σ a0,t 2 ×e t / s t ,σ min 2 ,σ max 2 ), σ min 2 and σ max 2 These are preset minimum values ​​(obtained by collecting IMU acceleration data of the corresponding person when they are completely stationary or moving at a constant speed, and calculating the variance of the acceleration amplitude, such as 0.05m). 2 / s 4 ~0.1m 2 / s 4 (In the fire control room, simulate an emergency response, quickly run from your workstation to the furthest fire-fighting equipment and stop abruptly, repeatedly collect acceleration data, calculate the variance of acceleration for each movement, and take the maximum value, such as 3m) 2 / s 4 ~4m 2 / s 4 ), clamp() is the amplitude limiting function, σ a0,t 2 The initial value of the acceleration sample variance is given; R is a 3×3 diagonal matrix, pre-calibrated according to the UWB noise characteristics. UWB tags are placed in fixed positions, and static data of three-dimensional coordinates are collected over a period of time. The diagonal elements represent the variances of the three-dimensional directions. The pre-calibrated coordinate transformation matrix is ​​pre-selected in key areas such as the fire control room and evacuation routes and stairwells on each floor, selecting n1 fixed physical feature points (n1≥3, such as corners, column centers, equipment reference points, etc.) to ensure that feature points within each calibration area are not collinear. Their coordinates P in the UWB coordinate system are measured respectively. i UWB And extracting the coordinates P of the BIM model in the global coordinate system i BIM The optimal transformation parameters are solved using the least squares method: first, the center point P is calculated using the average of n1 coordinates. avg UWB and P avg BIM Subtracting the center point from each of the n1 coordinates yields the decentralized result Q.i UWB and Q i BIM Construct the covariance matrix HH. Perform singular value decomposition on HH to obtain the rotation matrix RR, and calculate the translation vector TT, where TT = P. avg BIM -RR×P avg UWB Unified coordinate system P t =RR×P t UWB +TT.

[0018] S2: Construct the road network topology using the BIM evacuation route centerline, match road network nodes based on the 3D positions of personnel and equipment, obtain the final matched nodes, and construct a static travel distance matrix, including: S2.1: Based on the centerline of the fire evacuation route in the BIM model, construct the set of road network topology segments L. ro L ro ={ l 1, l 1,..., l J}, each paragraph l j The road network centerline is a continuous line segment with endpoints at road network nodes, and J is the number of road network bus segments; then the personnel nodes vv p Equipment operation node vv k The global 3D coordinates are projected onto the 2D plane of the fire control room floor (z is the fixed height of the ground) to obtain the 2D projected coordinates. The Euclidean distance from the 2D projected coordinates to the 2D coordinates of all road network nodes in the BIM road network topology is calculated, and the road network node with the minimum distance is selected as the initial nearest node. Finally, a circular neighborhood is drawn with the 2D projection points of personnel and equipment as the center and half the design net width of the fire evacuation route as the radius. The overlap between this neighborhood and the ground area of ​​the evacuation route is verified. If there is an overlap, the initial node is confirmed as the final personnel and equipment operation matching node P. p_ro p and P d_ro k If there is no overlap, traverse the road network topology to find the first road network node that overlaps with the neighboring area as the final matching node, ensuring that the matching node is within the range of a passable evacuation route. S2.2: Based on the road network nodes in the set of road network topology segments, construct a static travel distance matrix A consisting of the number of road network nodes multiplied by the number of road network nodes. t The matrix elements represent the travel distance of the road network segments between corresponding two road network nodes; In this embodiment, it is important to specify that the boundary of a single road network centerline segment is based on the inherent geometric feature points of the fire evacuation route. The specific steps are as follows: 1. Basic boundary node extraction: From the BIM fire evacuation route layer, extract the intersection points (road junctions / forks / merges), turning points (centerline bends ≥ 15°), and endpoints (road start / end / closure points) of the evacuation route centerline as the core boundary nodes of the basic road network; 2. Single-segment basic line segment division: Connect the core boundary nodes sequentially according to the continuous direction of the evacuation route. A continuous centerline segment without intersections or bends between adjacent core boundary nodes is a single segment. 3. Scene node supplementation and adaptation: The ground two-dimensional projection points of key fire equipment operation points are matched to the nearest core boundary node of the basic road network through two-dimensional Euclidean distance + landing area verification; If the projection point is located in the middle of a single basic line segment, an auxiliary road network node is added on the line segment, the original basic line segment is divided into two continuous road network segments, the road network topology line segment set is updated, and the equipment node matching is completed; The ground two-dimensional projection points of personnel nodes are matched to the nearest node of the basic road network in real time, without adding any road network nodes, to ensure the static stability of the road network topology.

[0019] Please see Figure 3 As shown, S3: Using a target temporary obstacle detection model, real-time obstacle information is obtained, the static travel distance matrix is ​​updated, and a dynamic travel distance matrix representing the travel distance of road network paths is generated, including: S3.1: Real-time acquisition of video streams from a set of road network topology segments via deployed cameras, and decoding of these streams into continuous image frames. t For image frame I t Preprocessing is performed, including scaling the image frames to the input size adapted to the YOLOv8 model (e.g., 640×640), pixel value normalization (scaling pixel values ​​from [0,255] to the [0,1] range to eliminate the interference of brightness differences on model inference), and channel order correction (since the image frame channel order output by the video decoding library (e.g., OpenCV) is BGR, it needs to be converted to the RGB channel order required by the YOLOv8 model, which is achieved by reordering the channel dimension index to ensure that the color features are consistent with the input during model training); the preprocessed image frames are input into the fine-tuned YOLOv8 model, and forward inference is performed to obtain the detection result DD. t DD t ={(c m1 ,b m1 ,s m1 )} m1=1 M1 M1 is the number of detected obstacle instances, c m1For the m1th obstacle category (such as piled-up objects, temporary cables, fireproof roller shutters falling in evacuation routes, locked evacuation staircases, etc.), b m1 =(x min ,y min ,x max ,y max ) represents the bounding box of the obstacle (image coordinate system), s m1 To test the confidence level (e.g., setting the filter threshold to 0.6 to eliminate low-confidence false positives); then, combining the camera calibration parameters (intrinsic matrix + distortion coefficients + extrinsic matrix) and the three-dimensional coordinate constraints of the fixed plane on the ground of the fire control room (e.g., height z=0), the center of the bounding box in the image coordinates (x c ,y c Mapping to the global three-dimensional coordinate system of the control room, the three-dimensional position coordinates O of the temporary obstacle are obtained. m1 =(x m1 ,y m1 ,z m1 ), z m1 The obstacle height can be determined using scene priors (height ranges can be assigned based on obstacle type or estimated using the height ratio of the detection box in the image; scene priors could be 0.5–1.5m for stacked objects, 0m for temporary cables touching the ground, and z for fireproof roller shutters falling). m1 = Falling height (e.g., from 2.0m to 0.8m, the range is [0.8, 2.0]), evacuation staircase locked [0, 2.0]; height ratio estimation of the true height H of a known ground reference object (e.g., cabinet, baseboard). ref and height h in the image ref The height of the obstacle detection box in the image is h. obs Estimate the actual height of the obstacle proportionally = H ref ×(h obs / h ref Output the real-time obstacle location set O={O m1}; This embodiment specifically describes that the target detection model is built on the YOLOv8 architecture. It collects video frames of road network topology segments monitored in the fire control room under different lighting conditions, time periods, and obstacle types. Obstacle categories and bounding boxes are labeled, and a scene-specific dataset is constructed. The YOLOv8 model is then fine-tuned through transfer learning to obtain a fine-tuned temporary obstacle detection model based on the YOLOv8 architecture, improving the detection accuracy and scene adaptability for targets such as stacked objects and moving personnel. The fine-tuning process uses a cross-entropy loss function to optimize category prediction and a CIoU loss function to optimize bounding box regression, ensuring that the obstacle positions and categories output by the model meet the real-time perception requirements of this invention. Using pre-calibrated camera intrinsic matrix, distortion coefficients, and extrinsic matrix, and employing well-known camera calibration and coordinate mapping methods, the image coordinates are mapped to the global three-dimensional coordinate system of the fire control room, obtaining a set of three-dimensional spatial coordinates for the obstacles.

[0020] S3.2: Based on the static travel distance matrix, obtain the matching node P for the person using a path planning algorithm (such as Dijkstra). p_ro p To device operation matching node P d_ro k Composed of multiple road network segments l j,K1 The compliant road network path L pa p,k ={ l j,1 , l j,2 ,..., l j,K1 K1 is the total number of segments in the path, and 1 / 2 / K1 is the passage sequence number of the line segment in the current path. Traverse all device operation matching nodes K to obtain the personnel matching node P. p_ro p The set of compliant road network paths L to all devices pa p ={L pa p,k |k=1,2,...,K}, thus obtaining the set L of compliant road network paths from all personnel matching nodes to all devices. pa ={L pa p |p=1,2,...,P1}, where P1 is the number of personnel involved in the fire emergency response and K is the number of equipment operation nodes; This embodiment specifically explains that the well-known Dijkstra algorithm is used as the path planning algorithm. Based on the fire evacuation-specific layer of the BIM model, the set of nodes and reachable paths are constructed only within the scope of compliant fire evacuation passages, evacuation staircases, and vestibules, completely eliminating nodes and paths in non-evacuation areas.

[0021] S3.3: For the set of compliant road network paths L pa Each compliant road network path L pa p Each segment of the road network l j,K1 Perform a rapid rejection test, a crossover test, and a 3D height verification. If the rejection test determines that the road network segment does not intersect with the corresponding temporary obstacle, skip the crossover test and the 3D height verification. If it determines that they intersect, then perform the crossover test. If it determines that the road network segment does not intersect with the corresponding temporary obstacle, skip the 3D height verification. If it determines that they intersect, then perform the 3D height verification. First, perform the rapid rejection test: calculate the road network segment... l j,K1 Enclosing box and temporary obstacle O m1 The bounding boxes of the 3D bounding boxes are projected onto the 2D ground plane (height z=0). If the bounding boxes do not overlap, it is determined that the road network segment does not intersect with the corresponding temporary obstacle, and the straddle test and 3D height verification are skipped. Otherwise, if they intersect, the projected bounding boxes overlap, and the temporary obstacle O is extracted. m1 Let e1, e2, ..., e be all the closed boundary line segments of the projected polygon. M M is the number of sides of the polygon, representing the road network segments. l j,K1 With each boundary segment e of the polygon m Perform the cross-intersection test one by one. If any one of the four intersection conditions is met, then the road network segment is determined to be intersecting with the temporary obstacle O. m1 If two-dimensional projections intersect, and both non-intersection conditions are met simultaneously, then the road network segment is determined to be intersecting with the temporary obstacle O. m1 Non-intersecting; road network segments that are determined to intersect by two-dimensional projection. l j,K1 Three-dimensional height verification was conducted, using the fire control room floor (height z=0) as a reference, and temporary obstacles O were placed... m1 Height range (z) m1,min ,z m1,max ) and the effective passage height range of the road network segment (z ro,min ,z ro,max The overlap is severe, z m1,min z is the height of the bottom of the obstacle from the ground. m1,max z is the height of the top of the obstacle above the ground. ro,min Typically 0, z ro,max It is usually 2m, if max(z) m1,min ,z ro,min )≤min(z m1,max ,z ro,maxIf the height ranges overlap, the road network segment is ultimately determined to be blocked by the corresponding temporary obstacle; if the segment is blocked by any temporary obstacle, it is marked as a compliant road network path L. pa p If an obstacle is blocked, the remaining obstacles are no longer traversed. The static travel distance matrix between road network nodes is updated synchronously to obtain a dynamic travel distance matrix representing the travel distance of the road network path. The element values ​​of the two road network nodes corresponding to the blocked line segment in the matrix are set to ∞. If the two height intervals are completely non-overlapping, it is determined that the temporary obstacle does not have a blocking effect on the road network line segment. Then, the next temporary obstacle is traversed until all obstacles have been traversed, and the compliant road network path L is traversed. pa p If none of the segments intersect with any temporary obstacle in two dimensions or overlap in three dimensions, then the path from the personnel to the equipment operation node is deemed passable, and the elements of the corresponding line segments in the static passage distance matrix remain unchanged. This embodiment requires specific explanation of the fast repulsion test and the straddle test, which are common knowledge in computational geometry for determining line segment intersection. Fast repulsion test: Construct axis-aligned rectangles with the two line segments as diagonals. If the two rectangles do not intersect, then the two line segments definitely do not intersect. Straddle test: If the two line segments intersect, then the two endpoints of each line segment must be located on opposite sides of the line containing the other line segment. The straddle test has two intersection conditions: 1. The road network line segment straddles any boundary line segment of the polygon; 2. Both endpoints of the road network line segment fall inside the obstacle projection polygon; 3. Any endpoint of the road network line segment falls exactly on the boundary of the polygon; 4. The road network line segment is collinear with and overlaps with some boundary line segment of the polygon. The straddle test has two non-intersection conditions: 1. The road network line segment and all boundary line segments of the polygon... 1. The obstacles do not meet the following conditions: 1. They do not cross each other; 2. At least one endpoint of the road network segment falls outside the projected polygon of the obstacle; 3. Three typical temporary obstacles in the building were selected for height verification tests. The test results are as follows: 1. Cardboard boxes stacked on the ground (obstacle category: stacked objects), height range [0m, 1.2m], overlap with the effective passage range [0m, 2.0m] in the range of [0m, 1.2m], and it is determined that the road network segment is blocked; 2. Suspended cables above the passage (obstacle category: temporary cables), erection height range [2.1m, 2.2m], no overlap with the effective passage range, and it is determined that it is not blocked; 3. Half-height temporary fences (obstacle category: other stacked objects), height range [0m, 1.5m], overlap with the effective passage range in the range of [0m, 1.5m], and it is determined that the road network segment is blocked.

[0022] S4: Based on the dynamic travel distance matrix and the baseline travel speed of personnel in emergency scenarios, combined with obstacle travel correction coefficients, calculate the global shortest reach time from personnel nodes to each equipment operation node, including: S4.1: Based on the dynamic travel distance matrix, starting from the road network node currently matched by the person and ending at the road network node matched by the target device's operation point, the path planning algorithm is re-executed. The algorithm treats elements with a weight of ∞ in the matrix as "impassable," automatically skips blocked line segments, prioritizes the effective path with the shortest travel distance, and updates the person matching node P. p_ro p To device operation matching node P d_ro k Composed of multiple road network segments l j,K1 The compliant road network path L pa，g p,k ; S4.2: Based on the baseline passage speed v of personnel in emergency scenarios ba And the elements in the dynamic travel distance matrix, calculate the matching node P for the personnel. p_ro p To device operation matching node P d_ro k Compliant road network path L pa,g p,k The current travel time T t p,k T t p,k =d pa p,k / (v ba ×η t p,k ), d pa p,k For compliant road network path L pa,g p,k The travel distance of all road network segments, v ba The value is typically taken as 1.0-1.5 m / s, which conforms to fire evacuation regulations. η t p,k η is the passage correction factor. t p,k =exp(-λr), where λ>0 is the attenuation coefficient (controlling the attenuation curvature; the larger λ is, the greater the barrier density with respect to η). t p,k The more sensitive the impact), r is the proportion of the passable area covered by obstacles, and r is the ratio of the real-time obstacle coverage area S1 to the total passable ground area S2; wherein S2: based on the compliant road network path L pa,g p,k Using all segments as centerlines, and the design net width of the evacuation passage to which the centerline belongs as the bandwidth, a two-dimensional strip-shaped local passage area is delineated, resulting in S2; S1: based on L pa p,kThe set of real-time obstacle 3D positions and obstacle bounding boxes corresponding to all segments are used to project the 3D bounding box of each real-time obstacle onto the ground (height z). m1 =0), resulting in a two-dimensional polygon. Only the projected areas that fall entirely within the aforementioned strip-shaped passage area and do not overlap with fixed obstacles are accumulated to obtain S1. If the compliant road network path L pa p,k If unreachable, then T t p,k =∞; Finally, for each device operation node, extract the reachability time of all personnel nodes to it, and take the minimum value as the final global shortest reachability time T of that device. t,min k1 k1 is the k1th device, k1=1,2,...,K1, K1≤K1 is the critical operation point K of the device; S5: Compare the global shortest reach time with the preset standard response time of the corresponding device, generate a comprehensive delay risk index based on the comparison result, generate a graded warning signal with corresponding color based on the comprehensive delay risk index, and output it to the alarm device of the fire control system; The global shortest reach time T... t,min k1 The corresponding device's preset standard response time T std k1 Comparison based on the delay time ΔT of a single device k1 If the delay time ratio is greater than or equal to 50% but less than 100%, a yellow warning is triggered directly. The delay time ratio is the proportion by which the delay time of any device exceeds its standard response time. If it is greater than or equal to 100%, a red warning is triggered directly. If the reachability time of the k1th device to all personnel nodes is ∞, the device is determined to be globally unreachable, and a red warning signal is triggered directly. For all devices with a delay time < 50%, a pre-set device weight w is also considered. k1 The comprehensive delay risk index RI is obtained. t , K1 represents the quantity of critical fire-fighting equipment, w k1 Satisfying ∑w k1 =1; if RI t ≤First preset risk threshold RI th1 A green normal signal is generated, indicating that the response meets the standard. If RI th1 <RI t ≤Second preset risk threshold RI th2 A yellow warning signal is generated to indicate a potential risk of response delay. If RI t >RI th2 A red warning signal is generated, indicating a serious risk of response delay. The warning signal is then output to the fire control alarm device, simultaneously displaying the delayed equipment operation node, delay time ratio, and personnel location. In this embodiment, it is necessary to specifically explain the preset standard response time T. std k First, according to national fire safety standards, the response time must not exceed the permissible safe evacuation time specified in the building design. Then, based on engineering practice data, such as the standard response time of 30-60 seconds for signal confirmation + short-distance walking + operation confirmation of the manual fire alarm button / fire telephone jack, a linear adjustment is made according to the actual building height, personnel configuration, etc., in the project's specific scenario, such as a two-person shift mode (T). std k Discounted by 5% to 10%; preset equipment weight w k Based on the emergency importance allocation of fire protection equipment, fire engineers and maintenance personnel score the equipment according to its functions and emergency impact, and the scores are normalized to obtain weights. Preset risk thresholds are determined in conjunction with fire protection industry emergency response standards and scenario tolerance, such as the equipment's standard response threshold T. std k The delay times (T) for each device are calculated as follows: main unit 30s, sprinkler pump 40s, broadcast 45s, telephone 50s, lighting 60s. t (vv p ,vv d k )-T std k The main unit has a 10-second delay, the sprinkler pump has a 5-second delay, and there is no delay for the rest. The overall risk index is RI. t =0.5×10+0.25×5+0.1×0+0.1×0+0.05×0=6.25s. Based on the scene tolerance, the threshold RI is set. th1 =5s, RI th2 =15s,5s<RI t A delay of ≤15s indicates a potential delay, requiring the fire control room staff to pay attention. If the main unit delays by 20s, the delay rate is 66.7% ≥ 50%, triggering a yellow alert directly without calculating the overall risk index. If the main unit delays by 35s, the delay rate is 116.7% ≥ 100%, triggering a red alert directly.

[0023] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent monitoring method for a fire control room, characterized in that: include: S1: Acquire multi-source monitoring data from the fire control room through positioning and acquisition technology, including the coordinate set of static operating points of equipment and the coordinates of personnel positions; S2: Construct the road network topology using the BIM evacuation route centerline, match road network nodes based on the three-dimensional positions of personnel and equipment, obtain the final matched nodes, and construct a static passage distance matrix; The final matching node obtained is: S2.1: Based on the centerline of the fire evacuation route in the BIM model, a set of road network topology segments L is constructed. ro L ro ={ l 1, l 1,..., l J }, each paragraph l j The road network centerline is a continuous line segment with endpoints at road network nodes, and J is the number of road network bus segments; then the personnel nodes vv p Equipment operation node vv k The global 3D coordinates are projected onto the 2D plane of the fire control room floor to obtain 2D projected coordinates. The Euclidean distances from the 2D projected coordinates to the 2D coordinates of all road network nodes in the BIM road network topology are calculated, and the road network node with the minimum distance is selected as the initial nearest node. Finally, a circular neighborhood is drawn with the 2D projection points of personnel and equipment as the center and half the design net width of the fire evacuation route as the radius. The overlap between this neighborhood and the ground area of ​​the evacuation route is verified. If there is an overlap, the initial node is confirmed as the final personnel and equipment operation matching node P. p_ro p and P d_ro k ; If there is no overlap, traverse the road network topology to find the first road network node that overlaps with its neighbors as the final matching node; S3: Obtain real-time obstacle information through the target temporary obstacle detection model, update the static travel distance matrix, and generate a dynamic travel distance matrix representing the travel distance of the road network path; S4: Based on the dynamic travel distance matrix and the baseline travel speed of personnel in emergency scenarios, combined with the obstacle travel correction coefficient, calculate the global shortest reach time from personnel nodes to each device; The implementation of S4 includes: S4.1: Based on the dynamic travel distance matrix, taking the road network node currently matched by the person as the starting point and the road network node matched by the target device operation point as the ending point, re-execute the path planning algorithm. The algorithm considers elements with a weight of ∞ in the matrix as "inaccessible", automatically skips blocked line segments, prioritizes the effective path with the shortest travel distance, and updates the person matching node P. p_ro p To device operation matching node P d_ro k Composed of multiple road network segments l j,K1 The compliant road network path L pa，g p,k ; S4.2: Based on the baseline passage speed v of personnel in emergency scenarios ba And the elements in the dynamic travel distance matrix, calculate the matching node P for the personnel. p_ro p To device operation matching node P d_ro k Compliant road network path L pa，g p,k The current travel time T t p,k T t p,k =d pa p,k / (v ba ×η t p,k ), d pa p,k For compliant road network path L pa，g p,k The travel distance of all road network segments, η t p,k η is the passage correction factor. t p,k =exp(-λr), where λ>0 is the attenuation coefficient and r is the proportion of the passable area covered by obstacles; if the compliant road network path L pa p,k If unreachable, then T t p,k =∞; Finally, for each device operation node, extract the reachability time of all personnel nodes to it, and take the minimum value as the final global shortest reachability time T of that device. t,min k1 k1 is the k1th device, k1=1,2,...,K1, K1≤K1 is the critical operation point K of the device; S5: Compare the global shortest reach time with the preset standard response time of the corresponding device, generate a comprehensive delay risk index based on the comparison result, generate a graded warning signal with corresponding color based on the comprehensive delay risk index, and output it to the alarm device of the fire control.

2. The intelligent monitoring method for a fire control room according to claim 1, characterized in that: The implementation of S2 includes: S2.2: Based on the road network nodes in the set of road network topology segments, construct a static travel distance matrix A consisting of the number of road network nodes multiplied by the number of road network nodes. t The matrix elements represent the travel distance of the road network segments between corresponding two road network nodes.

3. The intelligent monitoring method for a fire control room according to claim 1, characterized in that: The implementation of S3 includes: S3.1: Real-time acquisition of video streams of road network topology line segment sets through deployed cameras, and decoding into continuous image frames I. t For image frame I t Preprocessing is performed; the preprocessed image frames are input into the fine-tuned YOLOv8 model, and forward inference is used to obtain the detection results DD. t DD t ={(c m1 ,b m1 ,s m1 )} m1=1 M1 M1 is the number of detected obstacle instances, c m1 For the m1th obstacle category, b m1 =(x min ,y min ,x max ,y max ) represents the bounding box of the obstacle, s m1 To test the confidence level; then, combining the camera calibration parameters and the three-dimensional coordinate constraints of the fixed plane on the ground of the fire control room, the center of the bounding box in the image coordinates (x c ,y c Mapping to the global three-dimensional coordinate system of the control room, the three-dimensional position coordinates O of the temporary obstacle are obtained. m1 =(x m1 ,y m1 ,z m1 ), z m1 Given the obstacle height, output the real-time obstacle position set O1={O m1 } 4. The intelligent monitoring method for a fire control room according to claim 3, characterized in that: The implementation of S3 also includes: S3.2: Based on the static travel distance matrix, obtain the personnel matching node P through a path planning algorithm. p_ro p To device operation matching node P d_ro k Composed of multiple road network segments l j,K1 The compliant road network path L pa p,k ={ l j,1 , l j,2 ,..., l j,K1 K1 is the total number of segments in the path, and 1 / 2 / K1 is the passage sequence number of the line segment in the current path. Traverse all device operation matching nodes K to obtain the personnel matching node P. p_ro p The set of compliant road network paths L to all devices pa p ={L pa p,k |k=1,2,...,K}, thus obtaining the set L of compliant road network paths from all personnel matching nodes to all devices. pa ={L pa p |p=1,2,...,P1}, where P1 is the number of personnel involved in the fire emergency response, and K is the number of equipment operation nodes: S3.3: For the set of compliant road network paths L pa Each compliant road network path L pa p Each segment of the road network l j,K1 Perform a rapid rejection test, a straddle test, and a 3D height verification. If the rejection test determines that the road network segment does not intersect with the corresponding temporary obstacle, skip the straddle test and 3D height verification. If it determines that they intersect, perform the straddle test. If it determines that the road network segment does not intersect with the corresponding temporary obstacle, skip the 3D height verification. If it determines that they intersect, perform the 3D height verification. If the two height intervals overlap, the road network segment is ultimately determined to be blocked by the corresponding temporary obstacle. If the segment is determined to be blocked by any temporary obstacle, it is marked as a compliant road network path L. pa p If the obstacle is blocked, the remaining obstacles are no longer traversed. The static travel distance matrix between road network nodes is updated synchronously to obtain the dynamic travel distance matrix representing the travel distance of the road network path. The element values ​​of the two road network nodes corresponding to the blocked line segment in the matrix are set to ∞. If the two height intervals have no overlap, it is determined that the temporary obstacle does not have a blocking effect on the road network line segment. The compliant road network path L is traversed. pa p If none of the segments intersect with any temporary obstacle in two dimensions or overlap in three dimensions, then the path from the personnel to the equipment operation node is deemed passable, and the elements of the corresponding line segments in the static travel distance matrix remain unchanged.

5. The intelligent monitoring method for a fire control room according to claim 1, characterized in that: The S5 implementation includes: setting the global shortest reachability time T. t,min k1 The corresponding device's preset standard response time T std k1 Comparison based on the delay time ΔT of a single device k1 If the delay time ratio is greater than or equal to 50% but less than 100%, a yellow alert is triggered directly; if it is greater than or equal to 100%, a red alert is triggered directly; if the reachability time of the k1th device to all personnel nodes is ∞, the device is determined to be globally unreachable, and a red alert signal is triggered directly; for all devices with a delay time < 50%, combined with the preset device weight w k1 The comprehensive delay risk index RI is obtained. t ;If RI t ≤First preset risk threshold RI th1 A green normal signal is generated, indicating that the response meets the standard. If RI th1 <RI t ≤Second preset risk threshold RI th2 A yellow warning signal is generated to indicate a potential risk of response delay. If RI t >RI th2 A red warning signal is generated, indicating a serious risk of response delay. The warning signal is then output to the fire control alarm device, simultaneously displaying the operating nodes of the delayed equipment and the location of personnel.

6. An intelligent monitoring system for a fire control room, used with the intelligent monitoring method for a fire control room as described in any one of claims 1-5, comprising: Fire control room positioning and sensing module: It acquires multi-source monitoring data of the fire control room through positioning and acquisition technology, including the coordinate set of static operating points of equipment and personnel position coordinates; Fire control room road network construction module: Constructs road network topology through BIM evacuation route centerlines, matches road network nodes based on the three-dimensional positions of personnel and equipment, obtains the final matched nodes, and constructs a static passage distance matrix; Fire control room road network update module: By using the target temporary obstacle detection model, real-time obstacle information is obtained, the static travel distance matrix is ​​updated, and a dynamic travel distance matrix representing the travel distance of the road network path is generated; Fire control room spatiotemporal accessibility monitoring module: Based on the dynamic access distance matrix and the baseline access speed of personnel in emergency scenarios, combined with the obstacle access correction coefficient, calculates the global shortest access time from personnel nodes to each device; Fire control room alarm output module: It compares the shortest global reach time with the preset standard response time of the corresponding equipment, generates a comprehensive delay risk index based on the comparison result, generates a graded warning signal with corresponding color based on the comprehensive delay risk index, and outputs it to the alarm device of the fire control.

Images