Fire extinguishing bomb positioning device and method
By optimizing the launch of fire extinguishing projectiles through multi-sensor and ballistic calculation modules, the problems of dynamic fire source identification and response delay in existing technologies have been solved, achieving precise positioning and rapid fire extinguishing.
Patent Information
- Application Number
- CN202511110783.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-08
- Publication Date
- 2025-10-28
AI Technical Summary
Existing fire extinguishing devices cannot effectively identify dynamic fire sources, have high response delays, and cannot optimize the launch trajectory of fire extinguishing projectiles, resulting in unsatisfactory fire extinguishing effects.
The system employs a multi-sensor module to acquire multi-source data, and combines it with a positioning module and a ballistic calculation module. By using an ideal ballistic model and obstacle detection, the system optimizes the launch trajectory and uses an extended Kalman filter to predict the movement of the fire source, achieving precise positioning and dynamic compensation.
It achieves centimeter-level fire source location accuracy, dynamic fire source tracking capability, improved response speed, reduced resource waste, adaptability to complex scenarios, and improved fire extinguishing efficiency and safety.
Smart Images
Figure CN120837875A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of fire extinguishing technology, and in particular to a fire extinguishing bullet positioning device and method. Background Technology
[0002] Traditional firefighting methods primarily rely on passive response technologies, which suffer from the following drawbacks: 1. Delayed response and inaccurate positioning: Sprinkler systems rely on temperature / smoke thresholds for activation (e.g., glass bulb nozzles need to melt at 68°C or higher), causing the fire to spread to the middle or late stages before activation, missing the golden rescue period; fire extinguishers and fire hydrants rely on manual operation, which is inefficient and can easily cause casualties in high-rise buildings, chemical plants and other scenarios.
[0003] 2. Insufficient adaptability to different scenarios: Electrical fires require non-water-based extinguishing agents (such as gas extinguishing agents), and oil fires require foam coverage, but traditional systems can only use a single water-based extinguishing agent; there is a lack of effective detection methods for dynamic fire sources (moving vehicles) or concealed fire sources (smoldering on walls), resulting in blind spots in fire suppression.
[0004] 3. Waste of resources and inefficient management: High false alarm rate (such as kitchen smoke triggering the whole area sprinkler), causing equipment to be falsely triggered and water damage, with an average annual indirect loss of more than 4 billion yuan. The status of fire protection facilities cannot be monitored in real time, and 30% of the equipment cannot be used when malfunctioning due to lack of maintenance.
[0005] To address the aforementioned technical shortcomings, significant flaws still exist in the upgraded technologies developed by the industry: 1. Use infrared sensor-based fixed-point fire extinguishing devices (such as kitchen equipment) and mechanically triggered fire extinguishing bombs. However, these are only suitable for small, fixed areas and cannot dynamically locate the fire source. Infrared sensors are easily interfered with by non-flame heat sources such as heaters, resulting in a false trigger rate of up to 15%.
[0006] 2. Tracked / wheeled robots equipped with water cannons (such as Japan's KOHZU) and drones that drop fire extinguishing balls (such as DroneFire) rely on manual remote control or preset paths. Their autonomous navigation failure rate in complex environments exceeds 40%. Their vision systems are only used for obstacle avoidance and are not deeply integrated with the fire extinguishing execution mechanism, resulting in a response delay of >5 seconds.
[0007] 3. Fire alarms are achieved by analyzing monitoring footage using CNN models (such as ResNet). However, this only completes the alarm function and does not link the fire extinguishing execution module, thus failing to form a closed-loop control. In smoke-covered or nighttime scenarios, the recognition accuracy drops sharply to below 60%.
[0008] In summary, the existing technical problems are as follows: fixed-point fire extinguishing devices cannot identify dynamic fire sources, and due to the interference of on-site temperature and fire intensity at the fire scene, existing fire extinguishing devices rely on manual operation for fire extinguishing. Furthermore, the calculated trajectory of the fire extinguishing projectiles cannot be optimized as the fire source moves, resulting in a high response delay and unsatisfactory fixed-point extinguishing effect for dynamic fire points. Summary of the Invention
[0009] Therefore, it is necessary to provide a fire extinguishing bomb positioning device and method to address the aforementioned technical problems.
[0010] The following technical solution is adopted in this specification: This manual provides a fire extinguishing bomb positioning device, including: A multi-sensor module is used to acquire multi-source data; wherein, the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data; The positioning module is used to convert image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and to calculate the three-dimensional fire source center point. The ballistic calculation module is used to obtain the launch position of the fire extinguishing projectile positioning device; it is used to input the coordinates of the launch position and the center point of the three-dimensional fire source into the ideal ballistic model, and calculate the launch trajectory through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; This is used for safety verification of obstacle collision detection based on point cloud data. When the verification passes, the launched trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set; after calculating the trajectory point set... N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Launching mechanism, used to launch fire extinguishing projectiles according to the final launch trajectory.
[0011] Optionally, the device also includes a fire detection module, which is used to perform fire detection using temperature data, and when the temperature data is higher than the temperature threshold, uses a trained YOLOv8 model to perform flame detection on the RGB-D image data, and outputs bounding boxes and confidence scores; and uses the DeepSORT model to identify pixel-level temperature gradient data of infrared images and track dynamic fire sources. The positioning module is also used to activate the dynamic compensation mode when the moving speed of the fire source exceeds a preset threshold.
[0012] Optionally, the dynamic compensation mode includes updating the coordinates of the three-dimensional fire source center point at a frequency of 10Hz and predicting the coordinates of the three-dimensional fire source center point 0.5s later using an extended Kalman filter. The formula for predicting the coordinates of the three-dimensional fire source center point after 0.5 seconds using extended Kalman filtering is as follows: in, This represents the predicted fire source state vector after 0.5 seconds. This represents the current state vector of the fire source. Represents the current velocity vector. Indicates the prediction time step. This represents the acceleration compensation term caused by wind speed.
[0013] Optionally, the device further includes: The launch strategy adjustment module is used to adjust and analyze the fire distribution characteristics based on reinforcement learning algorithms, identify key fire points, and control the launch mechanism to attack the key fire points.
[0014] Optionally, the device further includes: The human-computer interaction module is used to switch the launch mode of the launch mechanism. When the launch mode is manual mode, after locking the fire source position, the launch mechanism will launch the fire extinguishing projectile after fine-tuning the launch position with manual intervention. When the firing mode is automatic firing, the firing mechanism fires the fire extinguishing projectile according to the final firing trajectory after locking the fire source location.
[0015] Optionally, the multi-sensor module includes: Multispectral and depth cameras are used to acquire RGB-D image data; LiDAR sensors are used to collect point cloud data; Infrared sensors are used to acquire infrared images; Wind speed sensor, used to collect real-time wind speed vectors; Temperature sensor, used to collect temperature data.
[0016] Optionally, the device further includes: The energy storage module is used to power the positioning device for fire extinguishing bombs; The control module is used to control the movement of the fire extinguishing bomb positioning device.
[0017] This manual provides a method for positioning fire extinguishing bombs, including: Multi-source data is acquired through a multi-sensor module; wherein, the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data; The positioning module converts image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and calculates the three-dimensional fire source center point. The launch position of the fire extinguishing projectile positioning device is obtained through the ballistic calculation module; the coordinates of the launch position and the center point of the three-dimensional fire source are input into the ideal ballistic model, and the launch trajectory is calculated through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; Safety verification for obstacle collision detection is performed based on point cloud data. When the verification passes, the launch trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set. The trajectory point set is then used for calculation. N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Fire extinguishing projectiles are launched from the launch mechanism according to the final launch trajectory.
[0018] The fire extinguishing bomb positioning device provided in this manual can detect fires based on multi-source data and calculate the three-dimensional fire source center point. Furthermore, it can optimize the launch trajectory of the fire extinguishing bomb based on the movement of the three-dimensional fire source center point, thereby achieving precise positioning and extinguishing of the fire source.
[0019] This invention can be operated in both manual and automatic modes. In manual mode, the nozzle automatically locks onto the fire source based on fire source identification, and fine adjustments can be made manually. In automatic fire extinguishing mode, simply place the device at the fire scene; it will automatically locate the fire source and calculate and predict the impact point of the extinguishing projectile. Attached Figure Description
[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0021] Figure 1 This is a schematic diagram of a fire extinguishing bomb positioning device provided in this specification; Figure 2 This is a flowchart illustrating the positioning and launching process of the fire extinguishing projectile of the present invention. Figure 3 This is a schematic diagram of the system of the present invention; Figure 4 This is a flowchart illustrating a method for positioning fire extinguishing bombs provided in this specification.
[0022] Explanation of reference numerals in the attached figures: 1-Positioning module; 2-Infrared module; 3-Temperature sensor; 4-Mini PC; 5-Control module; 6-Charging and energy storage module; 7-Camera; 8-Transmitting mechanism. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without creative effort are within the scope of protection of this application.
[0024] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.
[0025] Figure 1 This specification provides a schematic diagram of a fire extinguishing bomb positioning device, which specifically includes: A multi-sensor module is used to acquire multi-source data; wherein, the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data; The positioning module is used to convert image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and to calculate the three-dimensional fire source center point. The ballistic calculation module is used to obtain the launch position of the fire extinguishing projectile positioning device; it is used to input the coordinates of the launch position and the center point of the three-dimensional fire source into the ideal ballistic model, and calculate the launch trajectory through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; This is used for safety verification of obstacle collision detection based on point cloud data. When the verification passes, the launched trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set; after calculating the trajectory point set... N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Launching mechanism, used to launch fire extinguishing projectiles according to the final launch trajectory.
[0026] For example, the device also includes a fire detection module for fire detection using temperature data, and when the temperature data is higher than a temperature threshold, using a trained YOLOv8 model to perform flame detection on RGB-D image data and output bounding boxes and confidence scores; and for using a DeepSORT model to identify pixel-level temperature gradient data of infrared images and track dynamic fire sources. The positioning module is also used to activate the dynamic compensation mode when the moving speed of the fire source exceeds a preset threshold.
[0027] The dynamic compensation mode includes updating the coordinates of the three-dimensional fire source center point at a frequency of 10Hz and predicting the coordinates of the three-dimensional fire source center point 0.5s later through extended Kalman filtering. The formula for predicting the coordinates of the three-dimensional fire source center point after 0.5 seconds using extended Kalman filtering is as follows: in, This represents the predicted fire source state vector after 0.5 seconds. This represents the current state vector of the fire source. Represents the current velocity vector. Indicates the prediction time step. This represents the acceleration compensation term caused by wind speed.
[0028] In this embodiment, the device further includes: The launch strategy adjustment module is used to adjust and analyze the fire distribution characteristics based on reinforcement learning algorithms, identify key fire points, and control the launch mechanism to attack the key fire points.
[0029] The human-computer interaction module is used to switch the launch mode of the launch mechanism. When the launch mode is manual mode, after locking the fire source position, the launch mechanism will launch the fire extinguishing projectile after fine-tuning the launch position with manual intervention. When the firing mode is automatic firing, the firing mechanism fires the fire extinguishing projectile according to the final firing trajectory after locking the fire source location.
[0030] The multi-sensor module includes: a multispectral camera and a depth camera for acquiring RGB-D image data; a lidar sensor for acquiring point cloud data; an infrared sensor for acquiring infrared images; a wind speed sensor for acquiring real-time wind speed vectors; and a temperature sensor for acquiring temperature data.
[0031] The energy storage module is used to power the fire extinguishing bomb positioning device; the control module is used to control the movement of the fire extinguishing bomb positioning device.
[0032] In one specific embodiment, the preset safety distance can be 0.3 meters or 0.35 meters, which can be set according to actual needs and is not specifically limited here. The preset ballistic correction method can be: to correct the launch trajectory according to the correction rule of 1% for the height and 1% for the initial velocity of each launch trajectory. Here, the height correction percentage and the initial velocity correction percentage can be set according to actual needs and are not specifically limited here.
[0033] For example, Figure 2 This is a flowchart illustrating the positioning and launching process of the fire extinguishing projectile of the present invention. The multi-sensor module can acquire real-time video streams of the fire scene using a multispectral camera, and simultaneously obtain temperature distribution data. Optional equipment includes a LiDAR or depth camera (such as Intel RealSense) to acquire 3D point cloud data of the scene.
[0034] The fire detection module employs a lightweight YOLOv8 or an improved Faster R-CNN model, optimized for flame / smoke features. Combined with infrared thermal imaging data, feature-level fusion (such as Early Fusion) enhances the detection accuracy of small-target fires. The fire identification threshold can be adaptively adjusted based on the ambient temperature gradient to reduce false alarms.
[0035] The localization module estimates the distance to the fire source using pre-trained models such as MiDaS. It then converts the fire source location (u,v) in the image coordinate system to the world coordinate system (x,y,z) using calibration parameters.
[0036] The ballistic calculation module incorporates wind speed sensor data and combines it with parabolic equations to calculate the launch angle and initial velocity of the fire extinguishing projectile. Furthermore, it uses a Kalman filter to predict the movement trend of the fire source (such as the spread of oil fires). It can also dynamically adjust the launch strategy based on reinforcement learning (PPO algorithm), prioritizing attacks on key fire points.
[0037] In this embodiment, reference Figure 3 , Figure 3 This is a schematic diagram of the system of the present invention.
[0038] The hardware deployment of the fire extinguishing grenade positioning device can be as follows: 1. Sensor Installation and Calibration: Component List: Visible light camera (FLIR Blackfly S USB3), infrared thermal imager (FLIR A65sc), Livox Mid-40 LiDAR, custom mounting bracket (3D printed PA6-GF30 material).
[0039] Assembly process: The lidar is fixed to the front of the chassis via a shock-absorbing adapter plate, and the pitch angle is adjusted to +15° (scanning range covers 1-20m).
[0040] - Camera layout - Install a visible light camera and an infrared thermal imager inside the protective cover, with a spacing of ≤10cm to reduce parallax.
[0041] - Use Zhang Zhengyou's calibration method to calibrate the camera's intrinsic parameters (focal length, distortion coefficient) and extrinsic parameters (rotation and translation matrix). The calibration board temperature needs to cover 50℃~800℃ to match the fire scene environment.
[0042] - Auxiliary sensor integration - Mount the LiDAR (such as Livox Mid-40) on the top of the device, aligning the coordinate system with the camera with an error of ≤2mm.
[0043] - Install an IMU (such as BMI088) and an ultrasonic anemometer, and connect them to the main controller via an RS485 interface.
[0044] 2. Mechanical structure assembly - Launching mechanism - The pitch / yaw axis is driven by a servo motor (such as Dynamixel XM540) with an angular resolution of 0.1°.
[0045] - A pressure sensor (0-10MPa range) is integrated into the launch tube to provide real-time feedback on the projectile's initial velocity.
[0046] - Protective design - The sensor module is equipped with high-temperature resistant quartz glass (temperature resistance ≥1200℃), and the air jet dust removal device prevents smoke and dust from adhering.
[0047] - Installation of the omnidirectional mobile platform (Mecanum wheel chassis): - Use the Mecanum wheel omnidirectional moving chassis as the moving base for the entire system.
[0048] - Securely install the assembled launch mechanism system onto the mobile chassis, ensuring a reasonable center of gravity distribution.
[0049] - Connect the chassis drive motor controller to test the chassis's omnidirectional movement (forward, backward, lateral, and stationary rotation) function, and verify its flexibility and positioning accuracy in narrow spaces (such as requiring a turning radius of <0.5m).
[0050] In this embodiment, the software development for the fire extinguishing bomb positioning device can be as follows: 1. Training the fire detection model in the fire detection module can include: - Dataset Construction - Collect publicly available datasets (such as FireNet, FLAME) and supplement them with self-collected data: - Label flame / smoke bounding boxes in visible light images (LabelImg tool).
[0051] - Annotate the temperature gradient matrix in an infrared image (generated using Python + OpenCV script).
[0052] - Data augmentation: Simulate dense smoke obscuration (random noise) and dynamic blur (motion PSF kernel).
[0053] - Model training - Improve YOLOv8 using the PyTorch framework: - The input layer is expanded to 4 channels (RGB + infrared).
[0054] - Add a temperature distribution consistency constraint (custom L_temp) to the loss function.
[0055] - Quantized deployment: Convert to the TensorRT engine to achieve <30ms inference latency on Jetson AGX Orin.
[0056] 2. Development of positioning and ballistic algorithms: - 3D coordinate calculation # Monocular Depth Estimation + Coordinate Transformation Example Code Def world_coord_trans(u, v, depth_map, cam_matrix, T_cam_to_world): z = depth_map[v, u] # Get depth from the MiDaS model x_cam = (u - cam_matrix[0,2]) z / cam_matrix[0,0] y_cam = (v - cam_matrix[1,2]) z / cam_matrix[1,1] point_world = T_cam_to_world @ [x_cam, y_cam, z, 1] # Homogeneous coordinate transformation return point_world[:3] - Ballistic compensation - A motion model considering air resistance can be established as follows: ```math \theta = \arctan\left(\frac{v_y + \sqrt{v_y^2 + 2g(z_{\text{target}}- z_0)}}{v_x}\right) - \alpha_{\text{wind}} The alpha_{wind} is dynamically calculated from the anemometer data.
[0057] 3. Real-time system development - ROS 2-node design - Key Topic: - ` / thermal_image` (sensor_msgs / Image) - ` / fire_detection_result` (Custom message containing BBox and temperature) - ` / launch_command` (controls the servo motor and pressure valve) - Multi-threaded processing: Image acquisition (100Hz) and ballistic calculation (50Hz) are separated.
[0058] In this embodiment, the system integration and commissioning of the fire extinguishing bomb positioning device can be as follows: 1. Functional Testing - Fire detection verification - Release different fuels (wood / gasoline / lithium battery) in a simulated fire scene and test the detection rate (Recall≥95%).
[0059] - Extreme scenario: Verify the robustness of the infrared channel when strong light shines directly on the camera.
[0060] - Positioning accuracy test - Use the Vicon motion capture system as the ground truth to measure the fire source location error within 5m (RMSE ≤ 0.3m required).
[0061] 2. Dynamic scenario drills - Mobile fire source tracking - Set up a moving fire pit (speed 1m / s) on a conveyor belt to test the Kalman filter prediction effect.
[0062] - Multi-launch coordination - Deploy 2 devices and implement fire zone division (Voronoi graph algorithm) via MQTT communication.
[0063] 3. Long-term stability testing - Continuously running for 72 hours, monitoring: - Memory leak (detected by Valgrind tool).
[0064] - Thermal imager temperature drift (manual blackbody calibration once per hour).
[0065] In this embodiment, the deployment optimization of the fire extinguishing bomb positioning device can be as follows: 1. Edge computing optimization - The model was pruned using the NVIDIA TAO toolkit, reducing the number of parameters by 40%.
[0066] 2. Fault Recovery Mechanism - A watchdog timer monitors the process status and switches to a backup model (lightweight MobileNetV3) when an anomaly occurs.
[0067] 3. User Interface - Develop a PAD-based control app to display real-time heatmaps and ballistic trajectories (Unity 3D rendering).
[0068] Optionally, this embodiment employs a liquid lens (Optotune EL-16-40) for dynamic focusing to overcome the problem of camera defocusing caused by high temperatures. Furthermore, it integrates short-wave infrared (SWIR) images to enhance edge features and overcome the problem of low contrast at night.
[0069] The present invention has the following beneficial effects: I. Enhanced Precision Firefighting Capabilities Centimeter-level positioning accuracy By employing multi-sensor fusion (visible light + infrared + depth), the fire source location error is ≤0.3m (compared to >2m with traditional technology), avoiding "flooding" style fire suppression and reducing water damage by more than 90%.
[0070] Case study: In a lithium battery warehouse fire, the system can precisely target a single burning battery cell, whereas a traditional sprinkler system would affect the entire shelf.
[0071] Dynamic target tracking By using Kalman filtering to predict the trajectory of moving fire sources (such as vehicle fires), the hit rate of fire extinguishing bombs is increased to 92% (compared to only 40% for manual fire extinguishing).
[0072] II. Response Speed Revolution Millisecond-level detection The lightweight YOLOv8 model achieves a flame recognition latency of <30ms on Jetson AGX Orin (traditional smoke detectors require 30-60 seconds).
[0073] In comparison: the entire process from the discovery of the fire to the launch of the fire extinguishing bomb takes less than 500ms, which is 50 times faster than manual operation.
[0074] All-time domain combat capability The technology combines infrared thermal imaging with short-wave infrared (SWIR) fusion to achieve effective detection at night and in dense smoke environments (it can still work when the visible light camera fails).
[0075] III. Adaptability to Complex Scenarios IV. Resource Optimization and Security Reduced extinguishing agent usage: By calculating the minimum effective extinguishing dose through three-dimensional positioning, chemical consumption is reduced by 70% (compared to traditional gas extinguishing systems).
[0076] Zero contact with personnel: It can replace firefighters in extreme environments such as nuclear contamination and chemical explosions. In 2023, it successfully carried out unmanned firefighting in a chemical plant fire in Hangzhou.
[0077] based on Figure 1 The fire extinguishing projectile positioning device shown can detect fires based on multi-source data and calculate the three-dimensional fire source center point. Furthermore, it can optimize the trajectory of the fire extinguishing projectile based on the movement of the three-dimensional fire source center point, achieving precise location and extinguishing of the fire source. This invention can be operated manually or automatically. In manual operation, the device automatically locks the muzzle to the fire source location based on fire source identification, and fine-tuning can be performed manually. In automatic fire extinguishing mode, simply place the device at the fire scene; it will automatically locate the fire source and calculate and predict the impact point of the fire extinguishing projectile.
[0078] The above are one or more embodiments of the fire extinguishing shell positioning device provided in this specification. Based on the same idea, this specification also provides corresponding fire extinguishing shell positioning methods, such as... Figure 2 As shown.
[0079] Figure 2 A flowchart illustrating a method for positioning fire extinguishing bombs provided in this specification includes: Multi-source data is acquired through a multi-sensor module; the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data. The positioning module converts image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and calculates the three-dimensional fire source center point. The launch position of the fire extinguishing projectile positioning device is obtained through the ballistic calculation module; the coordinates of the launch position and the center point of the three-dimensional fire source are input into the ideal ballistic model, and the launch trajectory is calculated through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; Safety verification for obstacle collision detection is performed based on point cloud data. When the verification passes, the launch trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set. The trajectory point set is then used for calculation. N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Fire extinguishing projectiles are launched from the launch mechanism according to the final launch trajectory.
[0080] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
Claims
1. A fire extinguishing bomb positioning device, characterized in that, include: A multi-sensor module is used to acquire multi-source data; wherein, the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data; The positioning module is used to convert image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and to calculate the three-dimensional fire source center point. The ballistic calculation module is used to obtain the launch position of the fire extinguishing projectile positioning device; it is used to input the coordinates of the launch position and the center point of the three-dimensional fire source into the ideal ballistic model, and calculate the launch trajectory through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; This is used for safety verification of obstacle collision detection based on point cloud data. When the verification passes, the launched trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set; after calculating the trajectory point set... N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Launching mechanism, used to launch fire extinguishing projectiles according to the final launch trajectory.
2. The fire extinguishing bomb positioning device as described in claim 1, characterized in that, The device also includes a fire detection module, which is used to detect fires using temperature data, and when the temperature data is higher than the temperature threshold, it uses a trained YOLOv8 model to detect flames on RGB-D image data and outputs bounding boxes and confidence scores. Used to identify pixel-level temperature gradient data from infrared images using the DeepSORT model, and to track dynamic fire sources; The positioning module is also used to activate the dynamic compensation mode when the moving speed of the fire source exceeds a preset threshold.
3. The fire extinguishing bomb positioning device as described in claim 2, characterized in that, The dynamic compensation mode includes updating the coordinates of the three-dimensional fire source center point at a frequency of 10Hz and predicting the coordinates of the three-dimensional fire source center point 0.5s later through extended Kalman filtering. The formula for predicting the coordinates of the three-dimensional fire source center point after 0.5 seconds using extended Kalman filtering is as follows: in, This represents the predicted fire source state vector after 0.5 seconds. This represents the current state vector of the fire source. Represents the current velocity vector. Indicates the prediction time step. This represents the acceleration compensation term caused by wind speed.
4. The fire extinguishing bomb positioning device as described in claim 1, characterized in that, The device also includes: The launch strategy adjustment module is used to adjust and analyze the fire distribution characteristics based on reinforcement learning algorithms, identify key fire points, and control the launch mechanism to attack the key fire points.
5. The fire extinguishing bomb positioning device as described in claim 1, characterized in that, The device also includes: The human-computer interaction module is used to switch the launch mode of the launch mechanism. When the launch mode is manual mode, after locking the fire source position, the launch mechanism will launch the fire extinguishing projectile after fine-tuning the launch position with manual intervention. When the firing mode is automatic firing, the firing mechanism fires the fire extinguishing projectile according to the final firing trajectory after locking the fire source location.
6. The fire extinguishing bomb positioning device as described in claim 1, characterized in that, The multi-sensor module includes: Multispectral and depth cameras are used to acquire RGB-D image data; LiDAR sensors are used to collect point cloud data; Infrared sensors are used to acquire infrared images; Wind speed sensor, used to collect real-time wind speed vectors; Temperature sensor, used to collect temperature data.
7. The fire extinguishing bomb positioning device as described in claim 1, characterized in that, The device also includes: The energy storage module is used to power the positioning device for fire extinguishing bombs; The control module is used to control the movement of the fire extinguishing bomb positioning device.
8. A method for positioning fire extinguishing bombs, characterized in that, include: Multi-source data is acquired through a multi-sensor module; the multi-source data includes RGB-D image data, point cloud data, infrared images, real-time wind speed vectors, and temperature data. The positioning module converts image coordinates into world coordinates based on the pre-calibrated camera intrinsic parameter matrix and lidar-camera extrinsic parameter matrix, and calculates the three-dimensional fire source center point. The launch position of the fire extinguishing projectile positioning device is obtained through the ballistic calculation module; the coordinates of the launch position and the center point of the three-dimensional fire source are input into the ideal ballistic model, and the launch trajectory is calculated through the ideal ballistic model; the launch trajectory includes the initial launch elevation angle and initial velocity, and the influencing factors in the ideal ballistic model include gravitational acceleration; Safety verification for obstacle collision detection is performed based on point cloud data. When the verification passes, the launch trajectory is used as the final launch trajectory; when the verification fails, the `drag_trajectory` function is used to simulate the launch trajectory, incorporating air resistance and real-time wind speed vectors during the simulation to obtain an optimized trajectory point set. The trajectory point set is then used for calculation. N The actual error distance between the data points and the three-dimensional fire source center point is calculated. When the actual error distance is less than the preset safety distance, the ballistic trajectory point set is output as the final launch trajectory. When the actual error distance is greater than or equal to the preset safety distance, the launch trajectory is corrected according to the preset trajectory correction method until the actual error distance is less than the preset safety distance. The corrected ballistic trajectory point set is then output as the final launch trajectory. The final launch trajectory includes the final launch elevation angle and the final initial velocity. Fire extinguishing projectiles are launched from the launch mechanism according to the final launch trajectory.