Real-time monitoring system for forestry fire prevention

Abstract

The invention relates to the technical field of forestry monitoring, in particular to a real-time monitoring system for forestry fire prevention, which comprises a ground fixed monitoring terminal, an unmanned aerial vehicle mobile monitoring terminal, a data processing module, a dynamic networking transmission module, a cloud management platform and a linkage response module. The ground and the unmanned aerial vehicle terminal are integrated with a multi-source sensor, and accurate recognition is realized through a flame-smoke multi-feature fusion recognition algorithm improved based on YOLOv8; the transmission module adopts LoRa + 5G + satellite dynamic networking to ensure data transmission without a signal blind area; the cloud platform is fused with the GIS to realize early warning grading and accurate positioning; and the linkage response module is linked with the forest rangers, the fire-fighting resources and the channel guide equipment according to the early warning level. According to the system, the problems of many monitoring blind areas, high false alarm rate and slow response in the prior art are solved, comprehensive, real-time and accurate monitoring of forestry fire prevention is realized, and the fire early warning and fighting efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of forestry monitoring technology, specifically a real-time monitoring system for forestry fire prevention. Background Technology

[0002] Forestry resources are a core component of the ecological environment, bearing crucial ecological functions such as water conservation, air purification, and biodiversity maintenance. They are also an important material foundation for national economic development. However, forest fires, as the primary threat to forestry resources, are characterized by their suddenness, rapid spread, wide destructive range, and difficulty in firefighting. Once they occur, they not only lead to large-scale vegetation destruction and ecosystem imbalance but also directly threaten the lives and property of the people, causing incalculable economic losses and ecological damage. Therefore, constructing a real-time, accurate, and comprehensive forestry fire monitoring system is a key measure to prevent fires and reduce fire losses, and a core requirement of modern forestry management.

[0003] Currently, forestry fire prevention monitoring methods have gradually shifted from traditional manual monitoring to technological monitoring, forming a monitoring system that combines various methods such as manual patrols, fixed sensor monitoring, drone patrols, and satellite remote sensing monitoring. However, each monitoring method still has many technical shortcomings and cannot meet the fire prevention monitoring needs of complex forest areas.

[0004] Manual patrol monitoring, as a traditional method, relies on forest rangers to patrol on foot or by vehicle. In mountainous and remote forest areas with complex terrain, the patrol range is limited and the efficiency is low, making 24-hour uninterrupted monitoring impossible. At night, in heavy rain, or in dense fog, the safety and effectiveness of manual patrols are significantly reduced, making it difficult to detect early small fires or hidden fire hazards. Furthermore, the coverage density of manual patrols is limited by labor costs, easily creating monitoring blind spots in large forest areas, leading to the failure to detect fires in their early stages and missing the optimal time for firefighting.

[0005] Fixed sensor monitoring is currently the most widely used technological monitoring method, achieving fixed-point monitoring by deploying single-type sensors such as temperature and smoke sensors in forest areas. However, it has significant limitations. Firstly, the limited sensor types result in insufficient monitoring dimensions. Relying solely on temperature or smoke indicators makes it susceptible to environmental interference. For example, direct sunlight in summer may trigger false alarms from temperature sensors, while fog, dust, and industrial exhaust gases can cause smoke sensors to misinterpret, with false alarm rates generally exceeding 10%, resulting in substantial wasted patrol costs. Secondly, sensor deployment lacks scientific planning, often employing a uniform arrangement without considering the impact of terrain differences on the monitoring range. This easily creates monitoring blind spots in complex terrain areas such as valleys and ravines. Furthermore, the lack of data exchange between sensors prevents multi-source data complementarity, further reducing monitoring accuracy. In addition, fixed sensors often use wired or single wireless communication methods, making data transmission prone to interruption and data loss in remote mountainous areas with weak signals.

[0006] Drone patrol monitoring has been rapidly adopted due to its high mobility and wide coverage, but existing technologies still face several bottlenecks. First, there is a lack of coordination with ground monitoring equipment. Data transmission between drones and ground stations relies on a single 4G / 5G communication network, which is prone to data delays or loss in signal-dead areas such as mountainous regions, making real-time monitoring impossible. Second, patrol paths lack intelligent planning, often employing fixed routes that cannot dynamically adjust key monitoring areas based on ground monitoring data, leading to duplicate or missed detections. Third, fire identification algorithms lack accuracy, relying heavily on traditional image recognition methods, which struggle to effectively extract flame and smoke features against complex backgrounds (such as vegetation shadows, cloud cover, and complex terrain), resulting in low accuracy, especially in identifying early-stage small fires.

[0007] Satellite remote sensing can achieve large-scale macroscopic monitoring, but it is limited by satellite orbit period and spatial resolution, resulting in low monitoring frequency (generally 1-2 times per day), slow response speed, and difficulty in capturing the rapid spread of fire in its early stages. Furthermore, satellite remote sensing data is easily affected by weather conditions, such as clouds and rain, which can obstruct the monitoring line of sight, leading to a decrease in monitoring accuracy and failing to meet the needs of refined monitoring.

[0008] Furthermore, existing monitoring systems generally suffer from insufficient data fusion and a disconnect between early warning and coordinated response. Data collected by various monitoring devices is independent, lacking a unified data analysis platform and hindering deep integration of multi-source data, resulting in fragmented monitoring information. Early warning systems often only issue simple alarm signals, lacking a scientific risk grading mechanism and failing to provide accurate data for fire resource dispatch. The coordinated response mechanism is inadequate, preventing the rapid transmission of early warning information to responsible entities such as forest rangers and fire departments. Coordinated measures such as fire resource dispatch and forest fire prevention access guidance are delayed, hindering rapid firefighting efforts after a fire breaks out and further exacerbating fire damage.

[0009] In summary, existing forestry fire monitoring technologies have significant shortcomings in terms of monitoring coverage, identification accuracy, data transmission stability, and early warning linkage efficiency, making it difficult to meet the requirements of modern forestry fire prevention for real-time performance, precision, and comprehensiveness. Therefore, developing a real-time forestry fire monitoring system that can integrate multi-source monitoring equipment, achieve intelligent identification, stable transmission, and efficient linkage is of significant practical importance and application value for improving forestry fire prevention capabilities and protecting forestry resources and the ecological environment. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of existing forestry fire prevention monitoring systems, such as numerous monitoring blind spots, high false alarm rates, unstable data transmission, and untimely early warning linkage. It provides a real-time monitoring system for forestry fire prevention that achieves comprehensive coverage, real-time monitoring, accurate identification, and efficient linkage for forestry fire prevention through multi-source sensing fusion, intelligent algorithm recognition, dynamic network transmission, and precise linkage response, thereby minimizing the probability of forest fires and losses.

[0011] The technical solution adopted by the present invention to solve its technical problem is: a real-time monitoring system for forestry fire prevention, including a ground fixed monitoring terminal, a drone mobile monitoring terminal, a data processing module, a dynamic networking transmission module, a cloud management platform, and a linkage response module; The ground-based fixed monitoring terminals are arranged in the monitoring area at a preset density, and the UAV mobile monitoring terminals cruise along a preset path. Both the ground-based fixed monitoring terminal and the UAV mobile monitoring terminal integrate infrared thermal imaging sensors, smoke sensors, carbon monoxide sensors, temperature and humidity sensors, and wind speed and direction sensors. The data processing module is connected to each sensor and uses a flame-smoke multi-feature fusion recognition algorithm based on YOLOv8 to extract the color, shape and motion features of flames and smoke for intelligent recognition. The dynamic networking transmission module includes a LoRa communication unit, a 5G communication unit, and a satellite communication unit, and automatically switches the communication mode according to the signal strength of the monitoring area; The cloud management platform integrates a GIS geographic information system, receives data sent by the transmission module, and performs real-time display, risk assessment, and early warning classification. The linkage response module is connected to the cloud management platform and, based on the warning level, links the forest ranger terminal, the fire command system, and the forest fire prevention passage indicator equipment. Each module establishes data interaction through a communication protocol, realizing full-process automation of monitoring, identification, transmission, early warning, and linkage response.

[0012] Specifically, the ground-based fixed monitoring terminals are deployed at a density of 500-800 meters per terminal in plain areas, and at a density of 300-400 meters per terminal in ridges, valleys, and forest edges. They adopt a dual power supply mode of 300W solar panels + 200Ah lithium batteries and can work normally for more than 72 hours in continuous rainy weather.

[0013] Specifically, the improved YOLOv8 algorithm introduces an attention mechanism and optimizes the dataset through sample expansion, achieving an accuracy of no less than 98.5% and a false alarm rate of less than 1.2%.

[0014] Specifically, the switching mechanism of the dynamic networking transmission module is as follows: 5G communication is used when the 5G signal is ≥-70dBm; LoRa communication is switched when -90dBm < 5G signal < -70dBm; and satellite communication is started when the 5G signal is ≤-90dBm.

[0015] Specifically, the warning classification includes three levels: Level 1 warning corresponds to a smoke concentration ≥0.3 mg / m³. 3 Furthermore, there is no flame signal; a level-two warning corresponds to a smoke concentration ≥0.5 mg / m³. 3 Furthermore, localized flames were detected, and a level-three warning corresponds to a flame area ≥10m². 2 Or the carbon monoxide concentration is ≥50ppm.

[0016] Specifically, the UAV mobile monitoring terminal is based on GIS grid-based cruise, avoids obstacles with an elevation difference of more than 50 meters, cruises at a speed of 15-25 km / h, has a flight time of ≥2 hours, and supports data interaction and collaborative blind spot filling with ground terminals.

[0017] Specifically, the linkage response module pushes information to the forest ranger's terminal during a Level 1 warning; during a Level 2 warning, it links with the fire station and activates the passageway indicator lights; and during a Level 3 warning, it connects with the fire command center to synchronize fire information and dispatch fire resources.

[0018] Specifically, the infrared thermal imaging sensor has a detection range of ≥1500 meters and a resolution of ≥640×512, while the smoke sensor has a detection range of 0.01-5 mg / m³. 3 Response time ≤ 3 seconds.

[0019] Specifically, it also includes dual storage modules: local storage retains 30 days of raw data, while cloud storage retains 1 year of historical data and early warning records.

[0020] Specifically, it also includes a self-testing and maintenance module, which periodically calibrates sensors, checks communication and power supply status, and pushes alarms and fault location information when equipment fails.

[0021] The beneficial effects of this invention are: (1) The real-time monitoring system for forestry fire prevention described in this invention can achieve comprehensive coverage of the monitoring area and effectively eliminate monitoring blind spots. By deploying ground-based fixed monitoring terminals according to terrain characteristics—conventional layout in plain areas and denser deployment in complex terrain and high-risk fire areas—and combining them with GIS grid-based patrol and collaborative blind spot filling of UAV mobile monitoring terminals, it can achieve monitoring without blind spots, whether it is a mountain ridge and valley with complex terrain, the edge of forest land, or a remote forest area with weak signal, thus completely solving the problem of incomplete coverage caused by terrain limitations and insufficient manpower in traditional monitoring.

[0022] (2) The real-time monitoring system for forestry fire prevention described in this invention significantly improves the accuracy of fire identification and reduces the risk of false alarms. The system integrates multi-source sensors to collect environmental parameters and combines a flame-smoke multi-feature fusion recognition algorithm based on YOLOv8. By introducing an attention mechanism and optimizing the dataset, it can accurately extract the color, shape, and motion characteristics of flames and smoke. At the same time, it integrates multi-dimensional data such as gas, temperature, and humidity. It can not only effectively distinguish between real fire signals and environmental interference (such as direct sunlight, clouds, fog, dust, etc.), but also keenly capture early fire hazards (such as low-concentration smoke and local small flames), avoiding ineffective inspection costs and gaining critical time for early fire response.

[0023] (3) The real-time monitoring system for forestry fire prevention described in this invention ensures stable transmission of monitoring data and prevents data interruption or loss. The dynamic networking transmission module adopts a three-mode collaborative mode of LoRa, 5G and satellite communication, which can automatically switch the communication mode according to the signal strength of the monitoring area. Whether in the edge of the forest area with good signal or in the deep mountain area with weak signal, it can ensure continuous data transmission. This flexible transmission mechanism solves the problem of easy interruption of traditional single communication mode in complex environment, and provides coherent and complete data support for subsequent early warning and linkage response.

[0024] (4) The real-time monitoring system for forestry fire prevention described in this invention achieves seamless integration of scientific early warning and efficient linkage, significantly improving the efficiency of fire response. The cloud management platform integrates a GIS geographic information system and a risk assessment model, which can accurately classify early warnings based on monitoring data and clarify the degree of fire risk; the linkage response module automatically matches response measures such as forest ranger verification, fire station standby, fire resource dispatch, and passage guidance according to the early warning level, avoiding the problems of "disconnection between early warning and response" and "blind response measures" in traditional monitoring, realizing rapid connection from early warning triggering to response execution, and reducing the risk of fire spread.

[0025] (5) The real-time monitoring system for forestry fire prevention described in this invention reduces system maintenance costs and ensures data security and long-term stable operation of the system. The self-checking and maintenance module equipped with the system can periodically calibrate sensors and detect communication and power supply status. When equipment fails, it can accurately push the fault location and type, which is convenient for quick repair. The dual storage module retains recent original data and stores long-term historical data and early warning records through local and cloud data backup. This not only facilitates subsequent data traceability and analysis, but also avoids the problems of "difficult equipment maintenance" and "easy data loss" in traditional monitoring, ensuring the continuous and reliable operation of the system.

[0026] (6) The real-time monitoring system for forestry fire prevention described in this invention is of great significance for protecting forestry ecology and reducing economic losses. Through comprehensive monitoring, accurate identification, and rapid response, it can be dealt with in a timely manner at the bud stage of a fire, minimizing the damage to vegetation caused by the fire and maintaining the ecological functions of forests in conserving water resources, purifying air, and protecting biodiversity. At the same time, it avoids the loss of timber resources, increased firefighting costs, and personnel safety risks caused by the large-scale spread of fires, providing a strong guarantee for the protection of forestry resources and the development of the national economy. Attached Figure Description

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] Figure 1 This invention provides an architecture diagram of a real-time monitoring system for forestry fire prevention. Detailed Implementation

[0029] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0030] like Figure 1 As shown, the real-time monitoring system for forestry fire prevention of the present invention includes a ground-based fixed monitoring terminal, a drone-based mobile monitoring terminal, a data processing module, a dynamic networking transmission module, a cloud management platform, and a linkage response module. These modules work collaboratively to achieve full automation of the monitoring, identification, transmission, early warning, and linkage response process, specifically including: (a) Ground-based fixed monitoring terminal Ground-based fixed monitoring terminals are the basic monitoring units of the system, used to achieve fixed-point, continuous, multi-source environmental parameter acquisition. The density of these terminals is scientifically planned according to the topographical characteristics of the monitoring area: in plains and forest areas, they are evenly distributed at intervals of 500-800 meters to ensure no overlap and full coverage of the monitoring range; in mountainous areas, ridges, valleys, forest edges, and areas prone to fires, they are densely distributed at intervals of 300-400 meters to eliminate monitoring blind spots caused by terrain obstruction.

[0031] Each ground-based fixed monitoring terminal integrates an infrared thermal imaging sensor, a smoke sensor, a carbon monoxide sensor, a temperature and humidity sensor, and a wind speed and direction sensor to achieve simultaneous acquisition of multi-dimensional fire-related parameters. The infrared thermal imaging sensor is a high-sensitivity model with a detection range of ≥1500 meters and a resolution of ≥640×512, with a temperature measurement range of -20℃ to 500℃, capable of accurately capturing high-temperature radiation signals in early-stage fires; the smoke sensor has a detection range of 0.01-5 mg / m³. 3 With a response time of ≤3 seconds, it can quickly detect low concentrations of smoke; the carbon monoxide sensor has a detection range of 10-1000ppm and a response time of ≤3 seconds, enabling it to capture carbon monoxide gas produced by incomplete combustion in fires; the temperature and humidity sensor has a temperature measurement accuracy of ±0.2℃ and a humidity measurement accuracy of ±2%RH; the wind speed and direction sensor has a wind speed measurement accuracy of ±0.1m / s and a wind direction measurement accuracy of ±3°, providing environmental parameter support for fire risk assessment.

[0032] The ground-based fixed monitoring terminal adopts a dual power supply mode of solar panels and lithium batteries. It is equipped with a 300W high-efficiency solar panel and a 200Ah high-capacity lithium battery. The solar panel faces due south, and its tilt angle is adjusted according to the latitude of the monitoring area (generally ±5° of the local latitude) to ensure maximum solar energy absorption. Under sufficient sunlight, the solar panel powers the terminal and charges the lithium battery; at night or in cloudy or rainy weather, it automatically switches to lithium battery power, ensuring that the terminal can still operate normally for more than 72 hours even during continuous cloudy or rainy weather. A steel bracket is installed at the bottom of the terminal, with an installation height of 5 meters. The bracket is equipped with lightning protection and anti-theft devices to adapt to the complex environment of forest areas.

[0033] (ii) Mobile monitoring terminal for unmanned aerial vehicles Mobile drone monitoring terminals serve as a supplement to fixed ground-based monitoring terminals, enabling mobile blind spot monitoring and precise verification of key areas. Multi-rotor drones are selected, equipped with the same type of multi-source sensors as the fixed ground-based monitoring terminals, along with high-definition cameras and GPS positioning modules (positioning accuracy ±1cm) to ensure data consistency and positioning accuracy during mobile monitoring.

[0034] The monitored forest area is divided into grids based on a GIS geographic information system, with each grid unit being 1km x 1km. The drone's patrol path is preset according to the terrain complexity and vegetation type of each grid unit. The patrol path avoids obstacles with an elevation difference exceeding 50 meters, employing a clockwise or counter-clockwise grid patrol mode. The patrol speed is set at 15-25km / h, dynamically adjusted according to the warning level (reduced to 15km / h in warning areas, and maintained at 25km / h in non-warning areas). The drone's endurance is ≥2 hours, supporting hovering observation. When a warning signal is issued by a fixed ground monitoring terminal, the cloud management platform automatically sends instructions to the drones in the corresponding area, adjusting their patrol paths and prioritizing the patrol of the warning area and three surrounding grid units, achieving collaborative ground and aerial monitoring to fill gaps in coverage.

[0035] The drone communicates with the ground-based fixed monitoring terminal via LoRa communication units, sharing monitoring data and avoiding duplicate monitoring. It also supports direct communication with the cloud management platform, uploading monitoring data and identification results in real time, improving data transmission efficiency. Each drone is equipped with a spare battery that can be quickly replaced, ensuring continuous operation around the clock.

[0036] (III) Data Processing Module The data processing module is the core of the system to achieve accurate identification. It establishes data connections with the sensors of the ground fixed monitoring terminal and the UAV mobile monitoring terminal, and adopts a flame-smoke multi-feature fusion identification algorithm based on YOLOv8 to achieve intelligent identification of fire signals.

[0037] The core process of the algorithm includes three stages: data preprocessing, feature extraction, and multi-source fusion recognition. 1) In the data preprocessing stage, the infrared thermal imaging data and visible light image data are denoised (using median filtering algorithm), enhanced (using histogram equalization algorithm), and registered (based on SIFT feature matching algorithm) to eliminate the influence of environmental interference factors on the recognition results. 2) In the feature extraction stage, an attention mechanism module (CBAM attention mechanism) is introduced to enhance the feature extraction capability of flame and smoke areas, focusing on capturing the high-temperature radiation characteristics of flames (high grayscale areas in infrared images), irregular shape characteristics (edge ​​contour complexity), and smoke diffusion motion characteristics (inter-frame pixel change rate). At the same time, the numerical features of smoke sensors and carbon monoxide sensors are extracted. 3) In the multi-source fusion recognition stage, the image features and sensor numerical features are input into the improved YOLOv8 neural network model, and feature fusion is achieved through fully connected layers to output the recognition results (no fire, smoke warning, flame warning).

[0038] The algorithm's training dataset includes flame and smoke samples, as well as interference samples (direct sunlight, fog, rain), under different weather conditions (sunny, cloudy, foggy, rainy), different time periods (daytime, nighttime), and different vegetation types (coniferous forest, broadleaf forest, shrubland), totaling over 100,000 samples. Dataset diversity was increased through sample augmentation techniques such as rotation, scaling, flipping, and adding noise to ensure the algorithm's robustness in complex environments. Testing showed that the algorithm's recognition accuracy is no less than 98.5%, with a false positive rate of less than 1.2%, reducing the false positive rate by more than 80% compared to traditional recognition algorithms.

[0039] (iv) Dynamic networking transmission module The dynamic networking transmission module is used to realize data transmission between each monitoring terminal and the cloud management platform. It adopts a three-mode dynamic networking method of LoRa+5G+satellite to ensure the stability and reliability of data transmission in different environments.

[0040] The module includes a LoRa communication unit, a 5G communication unit, and a satellite communication unit: 1) The LoRa communication unit uses the SX1278 chip, operates at a frequency of 433MHz, has a communication distance of 3-5 kilometers, supports low-power, long-distance data transmission, and is suitable for building local area networks between ground-based fixed monitoring terminals to achieve data relay forwarding; 2) The 5G communication unit uses the Huawei ME909s-821 module, supports SA / NSA dual-mode, has a transmission rate ≥100Mbps, and a latency ≤50ms, and is suitable for high-speed data transmission in areas with 5G signal coverage; 3) The satellite communication unit uses the Iridium 9602 module, supports global coverage, has a data transmission rate of 2.4kbps, and is suitable for remote mountainous areas without 5G and LoRa signals, ensuring data transmission without dead zones.

[0041] The switching mechanism of the transmission module is controlled in real time by the signal strength detection unit: when the 5G signal strength is ≥-70dBm, the 5G communication unit is used to directly upload data; when -90dBm < 5G signal strength < -70dBm, it automatically switches to LoRa communication, and data transmission is achieved through relay forwarding by adjacent ground fixed monitoring terminals; when the 5G signal strength is ≤-90dBm, the satellite communication unit is activated, and data is uploaded through low-orbit satellites. The switching process is completed automatically without manual intervention, ensuring the continuity of data transmission and avoiding data loss.

[0042] (v) Cloud Management Platform The cloud management platform is the core control unit of the system, integrating a GIS geographic information system, a data storage module, a risk assessment module, and an early warning display module to realize functions such as data reception, storage, analysis, and early warning.

[0043] GIS Geographic Information System: Loads geographic data such as 1:10000 topographic and geomorphological maps, vegetation distribution maps, water source distribution maps, forest fire prevention road maps, fire station location maps, and settlement distribution maps of the monitored forest area. It accurately marks the locations of ground fixed monitoring terminals and UAV mobile monitoring terminals (positioning accuracy ±5 meters), supports map zooming, panning, layer switching, and other operations, and intuitively displays the geographical environment and equipment distribution of the monitored area.

[0044] Data storage module: It adopts a dual storage mode of local storage + cloud backup. The local storage (64GB SD card) of the ground fixed monitoring terminal and the UAV mobile monitoring terminal retains the raw monitoring data of the most recent 30 days. The cloud management platform uses Alibaba Cloud server (configured with 8 cores, 16G memory, and 10TB storage capacity) to store one year of historical data and all early warning event records, and supports data query, export, statistical analysis and visualization (such as trend charts and heat maps).

[0045] Risk Assessment Module: Based on the identification results from the data processing module, and combined with parameters such as temperature and humidity, wind speed and direction, vegetation type, and terrain slope, a fire risk assessment model is constructed. The model employs a machine learning algorithm (random forest algorithm). Input parameters include smoke concentration, carbon monoxide concentration, flame area, temperature, humidity, wind speed, wind direction, vegetation flammability coefficient, and terrain slope. The outputs the probability of fire occurrence (0-100%) and the predicted spread rate (m / min), providing a scientific basis for early warning classification.

[0046] Early warning display module: Based on risk assessment results, early warnings are categorized into three levels: Level 1 (yellow), Level 2 (orange), and Level 3 (red). Different colors are used to mark the warning locations on a GIS map, and the module simultaneously displays the warning level, monitoring data (smoke concentration, carbon monoxide concentration, temperature, wind speed and direction), fire probability, and predicted spread trends. Early warning information is simultaneously pushed to the management center's monitoring screen, management personnel's mobile app, and SMS messages to ensure relevant personnel receive timely warning information.

[0047] (vi) Linkage Response Module The linkage response module is connected to the cloud management platform in real time and executes corresponding linkage response measures according to the warning level, so as to achieve efficient connection between warning and response.

[0048] Level 1 Warning (Yellow): Corresponds to smoke concentration ≥ 0.3 mg / m³ 3Furthermore, with no flame signal, the probability of a fire is 30%-50%. The linkage response module pushes early warning information to the terminal of the forest ranger responsible for the grid within the monitoring area, including parameters such as the early warning location (GPS coordinates, map marking), smoke concentration, temperature and humidity, wind speed and direction, etc., notifying the forest ranger to arrive at the early warning location within 30 minutes for verification, and providing feedback on the verification results (normal, false alarm, small fire) through the APP.

[0049] Level II Warning (Orange): Corresponds to smoke concentration ≥ 0.5 mg / m³ 3 And localized flames were detected (flame area < 10m²). 2 The probability of a fire is 50%-80%. In addition to pushing information to the forest ranger's terminal, the linkage response module also links with the nearest fire station to simultaneously warn of the location, area of ​​the fire, spread trend and surrounding environment information. After receiving the information, the fire station prepares to dispatch. At the same time, the LED indicator lights along the forest fire prevention road (one every 500 meters) are turned on to guide fire trucks to pass quickly and shorten the response time.

[0050] Level 3 Warning (Red): Corresponding flame area ≥ 10m 2 If the carbon monoxide concentration is ≥50ppm, the probability of a fire is ≥80%. The linkage response module immediately connects to the regional fire command center to synchronize the fire location, spread rate, surrounding vegetation type, water source distribution, fire station location, and forest fire prevention access status in real time. It assists the fire command center in formulating firefighting plans and dispatching large-scale firefighting resources such as fire trucks, helicopters, and firefighters. At the same time, it pushes evacuation warning information to residents' terminals within the fire-affected area (mobile phone text messages and village broadcasts), clarifying evacuation routes and safe assembly points to ensure the safety of personnel.

[0051] (vii) Self-test and maintenance module The self-test and maintenance module is used to perform regular self-tests and fault alarms on system equipment, ensuring long-term stable operation of the system. Every 24 hours, the module calibrates the sensors of fixed ground monitoring terminals and UAV mobile monitoring terminals, adjusting the sensor detection accuracy by comparing with a standard signal source to ensure accurate data acquisition. Daily signal testing is performed on the communication module, checking the signal strength and transmission rate of each communication unit to ensure normal data transmission. Weekly power module power testing is performed, checking the charging efficiency of the solar panels and the remaining power of the lithium batteries to ensure stable power supply.

[0052] When a device malfunction is detected (such as sensor data deviation exceeding ±5%, communication transmission rate below 10Mbps, or remaining lithium battery power below 20%), the self-test and maintenance module sends an alarm message to the cloud management platform, clearly marking the location, device number, and malfunction type of the faulty device. After receiving the alarm message, the management personnel can quickly arrange for staff to go to repair or replace the device, reducing system maintenance costs and downtime.

[0053] Example 1 This embodiment uses forestry fire prevention monitoring in a mountain forest farm (total area of ​​500 square kilometers) as an application scenario to describe in detail the specific implementation of the present invention. The forest farm is located at an altitude of 800-1500 meters, with terrain mainly consisting of mountains and hills. The vegetation is mainly composed of coniferous forests such as Pinus tabuliformis and Larch, and broad-leaved forests such as Oak and Poplar. It is a high-risk area for fires and has experienced many forest fires in the past. The existing monitoring system has problems such as many blind spots, high false alarm rate, and slow response.

[0054] I. System Deployment (I) Deployment of Ground-based Fixed Monitoring Terminals Based on the terrain characteristics of the forest farm, a differentiated deployment plan was formulated: in plains and hilly areas below 1,000 meters above sea level, ground fixed monitoring terminals were evenly arranged at a spacing of 700 meters; in ridges, valleys and forest edges above 1,000 meters above sea level, they were densely arranged at a spacing of 350 meters, with a total of 1,200 ground fixed monitoring terminals deployed.

[0055] The installation process for each ground-mounted fixed monitoring terminal is as follows: 1) Site selection: Choose a high-lying, unobstructed location, avoiding low-lying waterlogged areas and steep slopes; 2) Foundation construction: Excavate a 1m×1m×1.2m foundation pit, pour C30 concrete foundation, and pre-embed anchor bolts; 3) Bracket installation: Fix the steel bracket on the concrete foundation, with a bracket height of 5 meters. During installation, ensure that the bracket is perpendicular to the ground (vertical deviation ≤0.5°); 4) Equipment installation: Fix the terminal host to the top of the bracket, and fix the solar panel to the side of the bracket, facing due south at an angle of 30° (the latitude of this forest farm is 35°N); 5) Lightning and anti-theft measures: Install a lightning rod on the top of the bracket, with a grounding resistance ≤10Ω, and add an anti-theft lock and protective box to the terminal host.

[0056] The specific parameters of the sensors integrated into the terminal are as follows: the infrared thermal imaging sensor is FLIRA655sc, with a detection distance of 1800 meters, a resolution of 640×512, and a temperature measurement range of -20℃ to 500℃; the smoke sensor is MQ-2, with a detection range of 0.01-5mg / m³. 3The system features a 2.5-second response time; a MQ-7 carbon monoxide sensor with a detection range of 10-1000ppm and a response time of 2.8 seconds; an SHT30 temperature and humidity sensor with a temperature measurement range of -40℃ to 125℃ and an accuracy of ±0.2℃, and a humidity measurement range of 0-100%RH with an accuracy of ±2%RH; and an FS300 wind speed and direction sensor with a wind speed measurement range of 0-60m / s and an accuracy of ±0.1m / s, and a wind direction measurement range of 0-360° with an accuracy of ±3°. Each terminal is equipped with a 300W monocrystalline silicon solar panel and a 200Ah lithium iron phosphate battery, along with an MPPT controller to improve solar energy conversion efficiency.

[0057] (II) Deployment of UAV Mobile Monitoring Terminals DJI M300 RTK multi-rotor drones were selected as mobile monitoring terminals, with a total of 20 drones deployed in 5 drone base stations located in the east, south, west, north and central areas of the forest farm. Each base station deployed 4 drones, 2 operators and 10 spare batteries.

[0058] The modification and configuration of the drone are as follows: 1) Sensor integration: The same type of infrared thermal imaging sensor, smoke sensor, carbon monoxide sensor, temperature and humidity sensor and wind speed and direction sensor as the ground fixed monitoring terminal are integrated on the payload mount of the UAV. The sensor data is connected to the UAV flight control system through a serial port. 2) Communication module installation: Install a LoRa communication unit (SX1278) and a 5G communication unit (Huawei ME909s-821) inside the drone to achieve dual communication with ground terminals and cloud platforms; 3) GPS positioning optimization: Add an RTK differential positioning module to improve positioning accuracy to ±1cm; 4) Optimized battery life: Equipped with 4 intelligent flight batteries, each with a capacity of 5935mAh, providing a flight time of 2.5 hours, and supporting hot-swappable quick replacement.

[0059] The forest farm was divided into 500 square kilometers using a GIS geographic information system, comprising 500 1km x 1km grid units. Each drone base station was responsible for patrolling 100 grid units. The patrol route was preset via a cloud management platform, planned clockwise according to the grid units, avoiding obstacles with elevation differences exceeding 50 meters (such as steep cliffs and deep valleys). The patrol speed was set at 20km / h. During the patrol, the drone interacted with a fixed ground monitoring terminal every 30 minutes to synchronize monitoring data and avoid redundant monitoring.

[0060] (III) Deployment of Dynamic Network Transmission Module Each ground-based fixed monitoring terminal and UAV mobile monitoring terminal integrates a LoRa communication unit, a 5G communication unit, and a satellite communication unit, as specifically deployed as follows: 1) LoRa communication network construction: The LoRa communication units of the ground fixed monitoring terminal are networked at a frequency of 433MHz, and communication connections are established between adjacent terminals to form a local area network, supporting data relay forwarding; 2) 5G signal coverage test: Before deployment, a comprehensive test of the 5G signal in the forest farm was conducted, marking the 5G signal coverage area (mainly concentrated on the edge and central areas of the forest farm) and the weak signal area (mainly concentrated in the remote mountainous areas in the north and west). 3) Satellite communication activation: Global data service is activated for the satellite communication unit (Iridium 9602) of each terminal to ensure normal data transmission in areas without signal. 4) Switching mechanism configuration: Automatic switching of communication methods can be achieved by using the signal detection module built into the terminal and setting a preset signal strength threshold.

[0061] Hardware connection method of the transmission module: The output end of the data processing module is connected to the LoRa communication unit, 5G communication unit and satellite communication unit through RS485 interface. The communication unit realizes data transmission and reception through antenna. The antenna is installed on the top of the terminal, higher than the sensor, to ensure that the signal transmission is not blocked.

[0062] (iv) Deployment of cloud management platform The cloud management platform is deployed in the computer room of the forest farm management center. The hardware configuration is as follows: the server is an Alibaba Cloud ECS server, model g7.xlarge (8 cores, 16GB memory, 10TB SSD storage), equipped with 2 servers to achieve primary and backup switching to ensure stable platform operation; the monitoring screen is a 3×3 spliced ​​55-inch LCD screen with a resolution of 3840×2160, used to display GIS maps, monitoring data and early warning information in real time; the management personnel are equipped with 20 tablets and 50 smartphones, with a dedicated APP installed to realize mobile office.

[0063] The platform software deployment is as follows: 1) The operating system used is CentOS 7.9, and the database used is MySQL 8.0, which is used to store historical data and early warning records; 2) The GIS geographic information system uses ArcGIS Server, loading 1:10000 topographic and geomorphological maps, vegetation distribution maps, water source distribution maps, forest fire prevention road maps, location maps of 3 surrounding fire stations, and distribution data of 12 residential areas of the forest farm, and realizes map visualization through WebGIS technology; 3) The risk assessment module is developed using Python, and the assessment model is built based on the random forest algorithm. The model is trained and deployed using the TensorFlow framework. 4) The warning display module uses Vue.js to develop the web interface and React Native to develop the mobile APP, supporting functions such as warning information push, data query, and report generation.

[0064] (v) Deployment of the linkage response module The hardware connection and software configuration of the linkage response module are as follows: 1) Forest Ranger Terminal: Equip 80 forest rangers with smartphones and install a dedicated system APP. The APP supports functions such as receiving early warning information, location navigation, and feedback of verification results. Each forest ranger is responsible for patrolling 5-8 grid units and binds to the corresponding grid units through the APP. 2) Forest fire prevention road indicator lights: LED indicator lights will be installed along the three main forest fire prevention roads in the forest farm, one every 500 meters, for a total of 300 lights. The indicator lights will be waterproof and lightning-proof, with a rated power of 10W and a brightness of ≥5000cd. They will be connected to the linkage response module via LoRa communication module and will support remote control to turn on and off. 3) Fire station linkage: Establish a data interface with the dispatch system of 3 fire stations around the forest farm (5-15 kilometers away from the forest farm), and use the RESTful API protocol to realize data interaction and synchronize early warning information and fire situation; 4) Regional fire command center linkage: Connect with the provincial fire command center through a dedicated network to upload real-time fire data and receive resource dispatch instructions.

[0065] (vi) Deployment of self-inspection and maintenance module The self-test and maintenance module's software is integrated into the ground-based fixed monitoring terminal and the cloud management platform, with the following specific configuration: 1) Sensor calibration parameters: Preset standard thresholds for each sensor, such as temperature calibration error ≤ ±0.5℃ for infrared thermal imaging sensors and concentration calibration error ≤ ±0.02mg / m³ for smoke sensors. 3 ; 2) Communication testing frequency: Signal testing of the communication module is performed at 0:00 and 12:00 every day to test the signal strength and transmission rate of each communication unit; 3) Power supply detection frequency: The power supply module is tested every Sunday at 1:00 AM to record the charging efficiency of the solar panel and the remaining power of the lithium battery; 4) Fault alarm settings: When a device fault is detected, an alarm message is immediately sent to the cloud management platform. The alarm message includes the faulty device number, location coordinates, fault type and fault occurrence time. At the same time, a pop-up notification is displayed on the monitoring screen and the management personnel's APP.

[0066] II. System Workflow (a) Data collection phase Each sensor on the ground-based fixed monitoring terminal collects monitoring data in real time at a frequency of 1 time per second. The collected data includes infrared thermal imaging frame data, smoke concentration value, carbon monoxide concentration value, temperature value, humidity value, wind speed value, and wind direction angle value. The collected data is first stored in a local 64GB SD card and then transmitted to the data processing module for real-time analysis.

[0067] The drone mobile monitoring terminal cruises along a preset grid path, and the sensors collect data at a frequency of 2 times per second. The collected data includes infrared thermal imaging frame data, smoke concentration value, carbon monoxide concentration value, temperature value, humidity value, wind speed value, wind direction angle value, and high-definition image data. Part of the collected data is transmitted to a nearby ground fixed monitoring terminal through the LoRa communication unit, and the other part is directly uploaded to the cloud management platform through the 5G communication unit. At the same time, the collected data of the most recent 12 hours is stored locally.

[0068] Under special weather conditions (such as heavy rain or dense fog), the system automatically adjusts the data acquisition frequency, increasing the acquisition frequency of the ground-based fixed monitoring terminal to 2 times / second and the acquisition frequency of the drone to 3 times / second, to ensure that weak fire signals can be captured.

[0069] (II) Data Processing Stage The data processing module of the ground-based fixed monitoring terminal processes the collected multi-source data in real time: 1) Data preprocessing: Median filtering algorithm is used to remove noise from infrared thermal imaging data, and histogram equalization algorithm is used to enhance image contrast; outlier removal is performed on sensor numerical data (using the 3σ criterion) to ensure data validity; 2) Feature extraction: High grayscale regions (flame candidate regions) in infrared thermal imaging data are extracted through the CBAM attention mechanism module. The edge contour complexity of the region (flame morphology features) and the inter-frame pixel change rate (smoke motion features) are calculated. At the same time, the numerical features of smoke concentration and carbon monoxide concentration are extracted. 3) Fusion recognition: Input image features and numerical features into the improved YOLOv8 model, and the model outputs the recognition result. If the recognition result is "smoke warning" or "flame warning", it is marked as a suspected fire signal, and the probability of fire occurrence is calculated at the same time.

[0070] The data processing module of the UAV mobile monitoring terminal adopts the same processing flow to identify the collected data in real time. When a suspected fire signal is identified, the identification result and the collected data are immediately transmitted to the cloud management platform and synchronized to the nearby ground fixed monitoring terminal to achieve complementary data verification.

[0071] (III) Data Transmission Stage After the data processing module of the ground-based fixed monitoring terminal outputs the identification result, the dynamic network transmission module selects the communication method to upload the data based on the current signal strength: 1) In the edge and central areas of the forest farm (5G signal strength ≥ -70dBm), data is directly uploaded using 5G communication units with a transmission rate of approximately 120Mbps and a latency of approximately 30ms. The data includes raw collected data, preprocessed image data, recognition results, and device status information. 2) In remote mountainous areas in the north and west of the forest farm (-90dBm < 5G signal strength < -70dBm), switch to LoRa communication unit to forward data to adjacent ground fixed monitoring terminals, which then relay the data to the cloud management platform. The transmission delay is about 100ms. 3) In the deep mountainous area in the southwest of the forest farm (5G signal strength ≤ -90dBm), the satellite communication unit is activated to upload data through the Iridium system. The transmission delay is about 10 seconds. The data is compressed using an algorithm (LZ77 algorithm) to reduce the amount of data transmitted, ensuring the effective transmission of critical data under low bandwidth.

[0072] During data transmission, the AES-256 encryption algorithm is used to encrypt the data to prevent it from being tampered with or stolen. At the same time, the CRC32 check algorithm is used to verify the transmitted data to ensure data integrity. If the verification fails, the data will be automatically retransmitted, and the number of retransmissions will not exceed 3.

[0073] (iv) Early warning classification stage After receiving the data, the risk assessment module of the cloud management platform performs further analysis on the data: 1) Input parameters: smoke concentration, carbon monoxide concentration, flame area, temperature, humidity, wind speed, wind direction, vegetation flammability coefficient (0.8 for coniferous forests, 0.5 for broad-leaved forests, and 0.6 for shrub forests), and terrain slope; 2) Model calculation: The probability of fire occurrence and the rate of fire spread are calculated using a trained random forest model; 3) Early warning classification: Early warning is classified according to the calculation results. Level 1 warning (yellow) corresponds to a fire occurrence probability of 30%-50%, Level 2 warning (orange) corresponds to 50%-80%, and Level 3 warning (red) corresponds to ≥80%.

[0074] The warning information is displayed in real time on the monitoring screen and the management personnel's APP: The left side of the monitoring screen displays a GIS map, with the warning location marked in yellow, orange and red, and the right side displays the warning details (warning level, equipment number, location coordinates, smoke concentration, carbon monoxide concentration, temperature, wind speed and direction, probability of fire occurrence, and spread speed); after receiving the warning push, the management personnel's APP displays the warning information in a pop-up window and supports clicking to navigate to the warning location.

[0075] (V) Joint Response Phase Level 1 Warning Response: The cloud management platform will push the warning information to the forest ranger's APP responsible for that grid unit. The information includes: "[Level 1 Warning] Smoke concentration of 0.35 mg / m³ was detected in grid XX (coordinates: XXX.XXXX, XXX.XXXX)." 3 "Temperature 28℃, wind speed 2.5m / s, fire probability 35%. Please verify within 30 minutes." Upon receiving the message, forest rangers will navigate to the warning location via the app. Upon arrival, they will take photos or videos of the scene using the app and report the verification results (normal, false alarm, small fire). If the report indicates a small fire (such as burning fallen leaves), the system will automatically escalate the warning to Level 2.

[0076] Level II Early Warning Response: In addition to sending information to forest rangers, the cloud management platform immediately sends early warning information to the nearest fire station dispatch system, simultaneously notifying the location of the warning and the flame area (5m²). 2 The system displays information on the fire's spread rate (0.5 m / min) and the distribution of surrounding water sources. Upon receiving the information, the fire station initiates its emergency response preparation procedure, and fire trucks and firefighters assemble within 15 minutes. Simultaneously, the system activates LED indicator lights along the forest firebreaks, which remain constantly lit in orange, guiding fire trucks quickly to the warning location. Forest rangers on-site provide real-time feedback on the fire situation via an app, offering decision-making support to the fire station.

[0077] Level 3 Early Warning Response: The cloud management platform immediately connects to the regional fire command center, synchronizing real-time fire data (flame area 15m²). 2 The fire was detected at a rate of 2 m / min, with a carbon monoxide concentration of 60 ppm, and a southwest wind direction of 4 m / s. Information on the distribution of surrounding fire-fighting resources was also provided. Based on this information, the fire command center formulated a firefighting plan, dispatching 3 fire trucks, 2 firefighting helicopters, and 50 firefighters to the affected area. Simultaneously, evacuation warning text messages were sent to 3 residential areas within the affected area, stating: "[Emergency Evacuation] A forest fire has occurred in the XX area and is expected to spread to surrounding areas within 1 hour. Please immediately evacuate northwest along the XX forest firebreak and proceed to the XX safe assembly point." The linkage response module continuously synchronized the fire's spread, providing real-time data support to the fire command center.

[0078] (vi) Self-inspection and maintenance phase The system automatically starts a self-check program at 2 AM every day: 1) Sensor calibration: Send standard signals to each sensor, compare the sensor output value with the standard value, and if the deviation exceeds the preset threshold (e.g., temperature sensor deviation > ±0.5℃), automatically adjust the sensor parameters; if adjustment is not possible, mark it as a fault. 2) Communication detection: Test the signal strength and transmission rate of each communication unit. If the transmission rate of a 5G communication unit is <10Mbps, it is marked as a communication anomaly. 3) Power supply detection: Detect the remaining power of the lithium battery. If the remaining power is less than 30%, check the charging efficiency of the solar panel. If the charging efficiency is less than 50%, mark it as a power supply fault.

[0079] After the self-inspection is completed, the cloud management platform generates an "Equipment Self-Inspection Report," recording the faulty equipment number, location, fault type, and handling suggestions. Based on the report, management personnel arrange for maintenance staff to handle the issue, and the faulty equipment is repaired or replaced within 24 hours.

[0080] III. System Performance The system has been deployed and operated in the forest farm for one year and has achieved the following results: 1) Monitoring coverage: 100% coverage of the 500 square kilometer forest area was achieved, with no monitoring blind spots, and a total of 32 early fire hazards were detected; 2) Recognition accuracy: The recognition accuracy rate is 99.2%, and the false alarm rate is 0.8%, which is 93.3% lower than the false alarm rate (12%) of the original monitoring system; 3) Response speed: The average verification time for Level 1 early warning is 22 minutes, the fire mobilization time for Level 2 early warning is 13 minutes, and the resource dispatch time for Level 3 early warning is 28 minutes, which is 55% faster than the original system. 4) Fire control: 32 fire hazards were successfully dealt with, of which 30 were controlled in the early stage and did not cause large-scale spread. Two fires that were upgraded to level three warning were extinguished within 2 hours, reducing economic losses by approximately 50 million yuan. 5) System stability: Annual failure rate of 2.5%, average repair time of faulty equipment of 18 hours, and average fault-free operation time of the system of more than 120 days.

[0081] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A real-time monitoring system for forestry fire prevention, characterized in that, It includes a ground-based fixed monitoring terminal, an unmanned aerial vehicle (UAV) mobile monitoring terminal, a data processing module, a dynamic networking transmission module, a cloud management platform, and a linkage response module; The ground-based fixed monitoring terminals are arranged in the monitoring area at a preset density, and the UAV mobile monitoring terminals cruise along a preset path. Both the ground-based fixed monitoring terminal and the UAV mobile monitoring terminal integrate infrared thermal imaging sensors, smoke sensors, carbon monoxide sensors, temperature and humidity sensors, and wind speed and direction sensors. The data processing module is connected to each sensor and uses a flame-smoke multi-feature fusion recognition algorithm based on YOLOv8 to extract the color, shape and motion features of flames and smoke for intelligent recognition. The dynamic networking transmission module includes a LoRa communication unit, a 5G communication unit, and a satellite communication unit, and automatically switches the communication mode according to the signal strength of the monitoring area; The cloud management platform integrates a GIS geographic information system, receives data sent by the transmission module, and performs real-time display, risk assessment, and early warning classification. The linkage response module is connected to the cloud management platform and, based on the warning level, links the forest ranger terminal, the fire command system, and the forest fire prevention passage indicator equipment. Each module establishes data interaction through a communication protocol, realizing full-process automation of monitoring, identification, transmission, early warning, and linkage response.

2. The real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The ground-based fixed monitoring terminals are deployed at a density of 500-800 meters per terminal in plain areas, and at a density of 300-400 meters per terminal in ridges, valleys, and forest edges. They adopt a dual power supply mode of 300W solar panels + 200Ah lithium batteries and can work normally for more than 72 hours in continuous rainy weather.

3. The real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The improved YOLOv8 algorithm introduces an attention mechanism and optimizes the dataset through sample augmentation.

4. The real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The switching mechanism of the dynamic networking transmission module is as follows: 5G communication is used when the 5G signal is ≥-70dBm; LoRa communication is switched when -90dBm < 5G signal < -70dBm; and satellite communication is started when the 5G signal is ≤-90dBm.

5. The real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The early warning classification includes three levels: A Level 1 warning corresponds to a smoke concentration ≥ 0.3 mg / m³. 3 And there was no flame signal; A Level II alert corresponds to a smoke concentration of ≥0.5 mg / m³. 3 And localized flames were detected; Level 3 warning corresponds to a flame area ≥ 10m² 2 Or the carbon monoxide concentration is ≥50ppm.

6. The real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The UAV mobile monitoring terminal is based on GIS grid-based cruise, avoids obstacles with an elevation difference of more than 50 meters, cruises at a speed of 15-25 km / h, has a flight time of ≥2 hours, and supports data interaction and collaborative blind spot filling with ground terminals.

7. A real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The linkage response module pushes information to the forest ranger's terminal during a Level 1 warning; during a Level 2 warning, it links with the fire station and activates the passageway indicator lights; and during a Level 3 warning, it connects with the fire command center to synchronize fire information and dispatch fire resources.

8. A real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: The infrared thermal imaging sensor has a detection range of ≥1500 meters and a resolution of ≥640×512, while the smoke sensor has a detection range of 0.01-5 mg / m³. 3 Response time ≤ 3 seconds.

9. A real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: It also includes dual storage modules: local storage retains 30 days of raw data, while cloud storage retains 1 year of historical data and warning records.

10. A real-time monitoring system for forestry fire prevention according to claim 1, characterized in that: It also includes a self-testing and maintenance module, which periodically calibrates sensors, checks communication and power supply status, and pushes alarms and fault location information when equipment fails.

Images