Forestry fire prevention early warning method and system based on Internet of Things

Abstract

The invention discloses a forestry fireproof early warning method and system based on the Internet of Things, and relates to the technical field of forestry monitoring, prevention and control, a heterogeneous sensing network is constructed through millimeter wave radar and multispectral imaging, and omnibearing monitoring in a complex environment is realized by combining terrain shielding coefficient Dmxs and signal coverage intensity Xgql analysis. The thermal radiation intensity Rsqd, the combustible drying index Gzzs and the airflow disturbance coefficient Qlys are fused to improve the hidden fire source identification precision; a LoRa relay and a Mesh network are adopted to ensure communication stability of a weak signal area, data transmission is optimized through edge calculation, and high-risk area resources are preferentially scheduled based on hierarchical response of a joint early warning index Lyzb; the risk assessment module calculates a fire danger index Hxzs in real time to predict fire spreading, the intelligent decision-making module comprehensively assesses hidden fire source probability Yhgl and a positioning error Wcsp to generate an early warning strategy, and the emergency linkage module automatically executes hierarchical response to form an intelligent monitoring-early warning-disposal closed-loop management system.

Description

Technical Field

[0001] The present invention relates to the technical field of forestry monitoring and prevention and control, and specifically to a forest fire prevention early warning method and system based on the Internet of Things. Background Art

[0002] The forest fire prevention early warning system originated from traditional manual patrols and observation towers. With the frequent occurrence of forest fires in forest areas and the increasing difficulty of prevention and control, means such as video monitoring and satellite remote sensing have gradually emerged; the forest fire prevention early warning system has now gradually evolved towards automation, real-time, and intelligence, forming a comprehensive prevention and control system with multi-source perception and multi-point linkage.

[0003] However, when the forest fire prevention early warning system is applied to the shelter forests of important infrastructure, the following technical problems often exist: 1. Monitoring blind spots caused by complex environments: The shelter forests around important infrastructure usually have complex terrains and dense forests. Traditional monitoring equipment (such as single smoke sensors or infrared detectors) is easily blocked, resulting in monitoring dead spots and making it difficult to timely and comprehensively perceive initial fire sources or potential hazards.

[0004] 2. The early warning timeliness is limited by communication capabilities: Most of the shelter forests are in remote areas or areas with weak signals. The existing wireless transmission or remote communication has insufficient stability, and there is a risk of transmission delay or loss of early warning information, affecting the timeliness and reliability of fire reporting.

[0005] 3. Lack of intelligent analysis and dynamic command capabilities: The current systems mostly rely on manual judgment or single fire indicators, lacking multi-source data fusion and intelligent algorithm support, and unable to dynamically predict the fire situation trend, spread path, influence range, etc., limiting the efficiency of scientific decision-making and command and dispatch. Summary of the Invention

[0006] In view of the deficiencies of the prior art, the present invention provides a forest fire prevention early warning method and system based on the Internet of Things, which solves the technical drawbacks mentioned in the background art.

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions: A forest fire prevention early warning system based on the Internet of Things, including a multi-dimensional perception module, an edge computing module, a risk assessment module, an intelligent decision-making module, and an emergency linkage module; The multi-dimensional perception module is used to construct a heterogeneous Internet of Things network through anti-blocking millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes, and collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content, and micro-meteorological data in real time to eliminate monitoring blind spots; An edge computing module for calibrating, abnormally eliminating, and compensating for missing collected data, dividing grid cells according to the terrain of the shelter forest, constructing a digital elevation model of the shelter forest, and calculating the terrain shielding coefficient Dmxs and signal coverage intensity Xgql of each grid cell; A risk assessment module for fusing the terrain shielding coefficient Dmxs and signal coverage intensity Xgql and generating a real-time fire risk index Hxzs; extracting the thermal radiation intensity Rsqd, combustible dryness index Gzzs, and air flow disturbance coefficient Qlys in the grid cell, and calculating the probability of hidden fire sources Yhgl; An intelligent decision-making module for predicting the fire spread path and the impact coefficient of critical infrastructure based on the spatio-temporal evolution of the real-time fire risk index Hxzs; and performing real-time evaluation on the probability of hidden fire sources Yhgl, and generating a joint warning index value Lyzb when the probability of hidden fire sources Yhgl exceeds the threshold; An emergency linkage module for evaluating the joint warning index value Lyzb and initiating a hierarchical response strategy, ensuring communication in weak signal areas by combining low-earth orbit satellites and Mesh networks, and dynamically pushing the resource scheduling priority and evacuation path optimization plan.

[0008] Preferably, the multi-dimensional perception module constructs a heterogeneous monitoring network by deploying millimeter-wave radar nodes, multi-spectral imaging nodes, and LoRa relay nodes. Among them, the millimeter-wave radar nodes obtain three-dimensional thermal distribution data of the vegetation-covered area by using a penetration detection method, the multi-spectral imaging nodes collect surface temperature and humidity gradient and combustible moisture content data through the cooperation of visible light and infrared bands, and the LoRa relay nodes realize low-power remote transmission of monitoring data; each node forms an adaptive network based on the spatial topology relationship and automatically enables multi-hop relay transmission when local communication is blocked; Unify the time scales of multi-source perception data through a time synchronization mechanism to form a three-dimensional environmental data set including spatial coordinates, timestamps, and monitoring values, and upload it to the edge computing module in real time.

[0009] Preferably, the edge computing module includes a data preprocessing unit and a grid analysis unit; The data preprocessing unit is used to detect and eliminate outliers from the collected three-dimensional thermal images and surface temperature and humidity gradient data, and compensate for missing data by using an adaptive Kalman filtering algorithm; establish a data correlation model based on time series analysis to perform spatio-temporal alignment on multi-source heterogeneous data.

[0010] Preferably, the grid analysis unit is used to divide the monitoring area into several dynamic grid units numbered FH1, FH2, ..., FHn according to the digital elevation model of the shelter forest and the three-dimensional environmental data set. The terrain shielding coefficient Dmxs of each grid unit is calculated by the line-of-sight analysis algorithm, taking into account the slope, vegetation density, and obstacle height; the signal coverage intensity Xgql is obtained by fusing the measured 5G signal intensity and the ray tracing model.

[0011] Preferably, the risk assessment module includes a fire risk index calculation unit and a concealed fire source calculation unit; The fire risk index calculation unit is used to perform the following calculation process: First, extract the terrain shielding coefficient Dmxs and the signal coverage intensity Xgql and perform normalization processing, and then calculate and output the real-time fire risk index Hxzs using the following formula: 。

[0012] Preferably, the concealed fire source calculation unit is used to obtain the following parameters in real time from each grid unit: The thermal radiation intensity Rsqd collected by the infrared thermal imager, the combustible dryness index Gzzs calculated based on the soil moisture sensor and meteorological station data, and the air flow disturbance coefficient Qlys calculated by the wind speed sensor and terrain data; After extracting the thermal radiation intensity Rsqd, the combustible dryness index Gzzs, and the air flow disturbance coefficient Qlys and performing dimensionless processing, obtain the concealed fire source probability Yhgl by calculating using the following formula: 。

[0013] Preferably, the intelligent decision-making module includes a spatio-temporal evolution analysis unit; The spatio-temporal evolution analysis unit is used to construct a spatio-temporal propagation model of the fire spread based on the dynamic change characteristics of the real-time fire risk index Hxzs, combined with the spatial position data and historical evolution rules of the monitoring area; and use the multi-path analysis algorithm and the key influence domain identification method to predict the possible fire spread paths.

[0014] Preferably, the intelligent decision-making module further includes a joint warning assessment unit; The joint warning assessment unit is used to evaluate the concealed fire source probability Yhgl; compare the concealed fire source probability Yhgl with the preset concealed fire source probability threshold Y; the specific assessment content is as follows: When the concealed fire source probability Yhgl ≤ the concealed fire source probability threshold Y, it is determined that the concealed fire source risk in the current monitoring area is in a controllable state; When the probability of hidden fire source Yhgl > the probability threshold of hidden fire source Y, the over-threshold response mechanism for hidden fire sources is automatically triggered, including recording the event of the probability of hidden fire source exceeding the limit, extracting the fire source positioning error Wcsp and the timeliness of emergency response Yxsx during the over-threshold period, and after dimensionless processing; the combined early warning index value Lyzb is calculated using the following formula: 。

[0015] Preferably, after the emergency linkage module obtains the combined early warning index value Lyzb in real time and evaluates it, preset response level division criteria are set, including the first response level division threshold L1 and the second response level division threshold L2; and the first response level division threshold L1 is greater than the second response level division threshold L2; The combined early warning index value Lyzb is classified and judged according to the response level division criteria; the specific content is as follows: When the combined early warning index value Lyzb is less than or equal to the second response level division threshold L2, it is determined that the current monitoring area is in a low-risk state, and the basic monitoring and information notification mechanism is started; When the combined early warning index value Lyzb is greater than L2 and less than or equal to the first response level division threshold L1, it is determined that the current monitoring area is in a medium-risk state, and the resource scheduling preparation and key area risk early warning process are automatically triggered; When the combined early warning index value Lyzb is greater than the first response level division threshold L1, it is determined that the current monitoring area is in a high-risk state, and comprehensive emergency response measures are immediately started, including linkage disposal strategies such as priority allocation of emergency resources, communication reinforcement in weak signal areas, safety control of key areas, and pushing of the optimal evacuation path.

[0016] The forest fire prevention early warning method based on the Internet of Things includes the following steps: Step 1: Construct a heterogeneous Internet of Things network through anti-occlusion millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes, and collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content, and micro-meteorological data in real time to eliminate monitoring blind spots; Step 2: Calibrate, abnormally eliminate, and compensate for missing data in the collected data, divide grid units according to the terrain of the shelter forest, and construct a digital elevation model of the shelter forest to calculate the terrain shielding coefficient Dmxs and signal coverage intensity Xgql of each grid unit; Step 3: Integrate the terrain shielding coefficient Dmxs and the signal coverage intensity Xgql, and generate a real-time fire risk index Hxzs; extract the heat radiation intensity Rsqd, fuel dryness index Gzzs, and air flow disturbance coefficient Qlys in the grid unit, and calculate the probability of hidden fire source Yhgl; Step 4: Based on the spatio-temporal evolution of the real-time fire risk index Hxzs, predict the fire spread path and the impact coefficient of key infrastructure; and conduct real-time assessment of the probability of hidden fire sources Yhgl, and generate the combined early warning index value Lyzb when the probability of hidden fire sources Yhgl exceeds the threshold. Step 5: Evaluate the combined early warning index value Lyzb and initiate a hierarchical response strategy, combine low-earth orbit satellites and Mesh networks to ensure communication in weak signal areas, and dynamically push the resource scheduling priority and the optimized evacuation path plan.

[0017] The present invention provides a forest fire prevention early warning method and system based on the Internet of Things, which has the following beneficial effects: (1) The forest fire prevention early warning method and system based on the Internet of Things construct a heterogeneous Internet of Things network through anti-occlusion millimeter-wave radar and multi-spectral imaging to achieve multi-dimensional perception of three-dimensional thermal images, surface temperature and humidity gradients, and fuel moisture content; combined with the dynamic analysis of the terrain shielding coefficient Dmxs and the signal coverage intensity Xgql, eliminate the monitoring dead corners in vegetation-dense areas and complex terrains; the calculation of the probability of hidden fire sources Yhgl integrates the heat radiation intensity Rsqd, the fuel dryness index Gzzs, and the air flow disturbance coefficient Qlys to improve the sensitivity of early fire source identification. (2) The forest fire prevention early warning method and system based on the Internet of Things use LoRa relay nodes and Mesh networks to achieve low-power long-distance data transmission to ensure the communication stability in weak signal areas; the edge computing module reduces data loss through Kalman filtering and spatio-temporal alignment; after the combined early warning index value Lyzb is triggered, it links low-earth orbit satellites to strengthen communication to ensure the real-time upload of early warning information and the issuance of instructions; the emergency linkage module conducts hierarchical responses based on Lyzb, preferentially schedules resources to high-risk areas (Lyzb > L1), and optimizes the timeliness Yxsx of the evacuation path push. (3) The forest fire prevention early warning method and system based on the Internet of Things, the risk assessment module calculates the fire risk index Hxzs in real time, and predicts the fire spread path in combination with the spatio-temporal evolution model; the intelligent decision-making module generates a scientific early warning strategy through the combined evaluation of the probability of hidden fire sources Yhgl and the fire source positioning error Wcsp; the emergency linkage module presets the response level division standard and compares and evaluates it with the combined early warning index value Lyzb, and automatically initiates hierarchical responses, including resource priority scheduling and key area control, to improve the command efficiency; multi-source data fusion and adaptive algorithms significantly reduce the need for manual intervention and form a closed-loop management of monitoring - early warning - disposal. Description of the Drawings

[0018] Figure 1 It is a schematic diagram of the framework structure of the forest fire prevention early warning system based on the Internet of Things of the present invention; Figure 2 It is a schematic diagram of the step flow of the forest fire prevention early warning method based on the Internet of Things of the present invention. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0020] Embodiment 1 Please refer to Figure 1 , the present invention provides a forest fire prevention early warning system based on the Internet of Things, including a multi-dimensional perception module, an edge computing module, a risk assessment module, an intelligent decision-making module and an emergency linkage module; The multi-dimensional perception module is used to construct a heterogeneous Internet of Things network through anti-occlusion millimeter wave radar, multi-spectral imaging and LoRa relay nodes, and collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content and micro-meteorological data in real time to eliminate monitoring blind spots; The edge computing module is used to calibrate, abnormally eliminate and missing compensate the collected data, divide grid units according to the terrain of the shelter forest, and construct a digital elevation model of the shelter forest to calculate the terrain shielding coefficient Dmxs and signal coverage intensity Xgql of each grid unit; The risk assessment module is used to fuse the terrain shielding coefficient Dmxs and signal coverage intensity Xgql, and generate a real-time fire risk index Hxzs; extract the heat radiation intensity Rsqd, fuel dryness index Gzzs and air flow disturbance coefficient Qlys in the grid unit, and calculate the probability of hidden fire source Yhgl; The intelligent decision-making module is used to predict the fire spread path and the influence coefficient of key infrastructure based on the spatio-temporal evolution of the real-time fire risk index Hxzs; and conduct real-time evaluation on the probability of hidden fire source Yhgl, and generate a joint early warning index value Lyzb when the probability of hidden fire source Yhgl exceeds the threshold; The emergency linkage module is used to evaluate the joint early warning index value Lyzb and start a hierarchical response strategy, ensure communication in weak signal areas by combining low-earth orbit satellites and Mesh networks, and dynamically push resource scheduling priorities and evacuation path optimization plans.

[0021] In this embodiment, the multi-dimensional perception module constructs a heterogeneous network through anti-occlusion millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes to collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content, and micro-meteorological data in real time, effectively eliminating monitoring blind spots; the edge computing module calibrates, removes anomalies, and compensates for missing data in the collected data, and calculates the terrain shielding coefficient Dmxs and signal coverage intensity Xgql based on the digital elevation model of the shelter forest to improve data reliability; the risk assessment module fuses Dmxs and signal coverage intensity Xgql to generate a real-time fire risk index Hxzs, and at the same time calculates the probability of hidden fire sources Yhgl by combining the heat radiation intensity Rsqd, fuel dryness index Gzzs, and air flow disturbance coefficient Qlys to achieve accurate risk assessment; the intelligent decision-making module predicts the fire spread path based on the spatio-temporal evolution of Hxzs, and generates a combined early warning index value Lyzb through real-time evaluation of the probability of hidden fire sources Yhgl to enhance the scientificity of decision-making; the emergency linkage module starts a hierarchical response strategy according to the combined early warning index value Lyzb, combines low-earth orbit satellites and Mesh networks to ensure smooth communication, dynamically optimizes the priority of resource scheduling and evacuation paths, and forms a closed-loop emergency management system.

[0022] Embodiment 2 The multi-dimensional perception module constructs a heterogeneous monitoring network by deploying millimeter-wave radar nodes, multi-spectral imaging nodes, and LoRa relay nodes. Among them, the millimeter-wave radar nodes use a penetrative detection method to obtain three-dimensional thermal distribution data of the vegetation-covered area, the multi-spectral imaging nodes collect surface temperature and humidity gradients and fuel moisture content data through the cooperation of visible and infrared bands, and the LoRa relay nodes realize low-power long-distance transmission of monitoring data; each node forms an adaptive network based on the spatial topological relationship and automatically enables multi-hop relay transmission when local communication is blocked; Through the time synchronization mechanism, the multi-source perception data is unified into a time standard, forming a three-dimensional environmental data set containing spatial coordinates, timestamps, and monitoring values, and uploading it to the edge computing module in real time.

[0023] In this embodiment, penetrative detection of the vegetation-covered area is realized by obtaining three-dimensional thermal distribution data through millimeter-wave radar nodes; the monitoring accuracy is improved by using multi-spectral imaging nodes to collect surface temperature and humidity gradients and fuel moisture content data in cooperation; LoRa relay nodes are used to ensure low-power long-distance transmission of monitoring data; the adaptive networking mechanism based on the spatial topological relationship automatically enables multi-hop relay transmission to ensure data connectivity when communication is blocked; the time synchronization mechanism unifies the time standard of multi-source perception data to form a three-dimensional environmental data set containing spatial coordinates, timestamps, and monitoring values; it is uploaded to the edge computing module in real time to provide standardized data input for subsequent analysis; this module effectively solves the problems of data collection integrity and transmission reliability in complex forest areas.

[0024] Embodiment 3 The edge computing module includes a data preprocessing unit and a grid analysis unit; The data preprocessing unit is used to detect and remove outliers from the collected three-dimensional thermal images and surface temperature and humidity gradient data, and compensate for missing data using an adaptive Kalman filtering algorithm; establish a data correlation model based on time series analysis to perform spatio-temporal alignment of multi-source heterogeneous data; The grid analysis unit is used to divide the monitoring area into several dynamic grid units numbered FH1, FH2,..., FHn according to the digital elevation model of the shelter forest and the three-dimensional environmental data set. The terrain shielding coefficient Dmxs of each grid unit is calculated by a line-of-sight analysis algorithm, taking into account slope, vegetation density, and obstacle height; the signal coverage intensity Xgql is obtained by fusing the measured 5G signal intensity and the ray tracing model.

[0025] The risk assessment module includes a fire risk index calculation unit and a hidden fire source calculation unit; The fire risk index calculation unit is used to perform the following calculation process: First, extract the terrain shielding coefficient Dmxs and the signal coverage intensity Xgql and perform normalization processing, and then calculate and output the real-time fire risk index Hxzs using the following formula: .

[0026] The hidden fire source calculation unit is used to obtain the following parameters in real time from each grid unit: The thermal radiation intensity Rsqd collected by an infrared thermal imager, the combustible dryness index Gzzs calculated based on soil moisture sensor and meteorological station data, and the air flow disturbance coefficient Qlys calculated based on wind speed sensor and terrain data; After extracting the thermal radiation intensity Rsqd, the combustible dryness index Gzzs, and the air flow disturbance coefficient Qlys and performing dimensionless processing, calculate and obtain the hidden fire source probability Yhgl using the following formula: .

[0027] The intelligent decision-making module includes a spatio-temporal evolution analysis unit; The spatio-temporal evolution analysis unit is used to construct a spatio-temporal propagation model of fire spread based on the dynamic change characteristics of the real-time fire risk index Hxzs, combined with the spatial location data and historical evolution rules of the monitoring area; and use the multi-path analysis algorithm and the key influence domain identification method to predict the possible fire spread paths; The intelligent decision-making module also includes a joint warning evaluation unit; The joint warning evaluation unit is used to evaluate the hidden fire source probability Yhgl; compare the hidden fire source probability Yhgl with a preset hidden fire source probability threshold Y; the specific evaluation content is as follows: When the probability of hidden fire source Yhgl ≤ the threshold value of the probability of hidden fire source Y, it is determined that the risk of hidden fire source in the current monitoring area is in a controllable state; When the probability of hidden fire source Yhgl > the threshold value of the probability of hidden fire source Y, the over-threshold response mechanism for hidden fire sources is automatically triggered, including recording the event of the probability of hidden fire source exceeding the limit, extracting the fire source positioning error Wcsp and the timeliness of emergency response Yxsx during the over-threshold period and performing dimensionless processing; the combined early warning index value Lyzb is calculated using the following formula: 。

[0028] After the emergency linkage module obtains the combined early warning index value Lyzb in real time and evaluates it, preset the response level division criteria, including the first response level division threshold L1 and the second response level division threshold L2; and the first response level division threshold L1 is greater than the second response level division threshold L2; The combined early warning index value Lyzb is classified and judged according to the response level division criteria; the specific content is as follows: When the combined early warning index value Lyzb is less than or equal to the second response level division threshold L2, it is determined that the current monitoring area is in a low-risk state, and the basic monitoring and information notification mechanism is started; When the combined early warning index value Lyzb is greater than L2 and less than or equal to the first response level division threshold L1, it is determined that the current monitoring area is in a medium-risk state, and the resource scheduling preparation and key area risk early warning process are automatically triggered; When the combined early warning index value Lyzb is greater than the first response level division threshold L1, it is determined that the current monitoring area is in a high-risk state, and comprehensive emergency response measures are immediately started, including linkage disposal strategies such as preferential allocation of emergency resources, communication reinforcement in weak signal areas, safety control of key areas, and pushing of the optimal evacuation path.

[0029] In this embodiment, the multi-dimensional perception module uses millimeter-wave radar nodes to obtain three-dimensional thermal distribution data, multi-spectral imaging nodes to collect surface temperature and humidity gradients and combustible moisture content data, and LoRa relay nodes to ensure low-power transmission, and constructs a high-precision environmental monitoring network; The edge computing module processes 3D thermal images and surface temperature and humidity gradient data through the data preprocessing unit using the adaptive Kalman filtering algorithm. The grid analysis unit divides dynamic grid cells FH1 - FHn based on the digital elevation model and calculates the terrain occlusion coefficient Dmxs and the signal coverage intensity Xgql, providing accurate data support for risk assessment; the risk assessment module generates a real-time fire risk index Hxzs by fusing Dmxs and Xgql through the fire risk index calculation unit, and the hidden fire source calculation unit calculates the hidden fire source probability Yhgl by combining the heat radiation intensity Rsqd, the combustible dryness index Gzzs, and the air flow disturbance coefficient Qlys, realizing quantitative risk assessment; the intelligent decision-making module uses the spatio-temporal evolution analysis unit to construct a fire spread model based on Hxzs, and the joint warning assessment unit generates a joint warning index value Lyzb by comparing Yhgl with the threshold Y, and realizes scientific warning by combining the fire source positioning error Wcsp and the emergency response timeliness Yxsx; The emergency linkage module starts a hierarchical response strategy according to the comparison results of Lyzb with the hierarchical thresholds L1 and L2, including preferential resource allocation and optimal evacuation path push, forming a complete monitoring-warning-disposal closed loop; the collaborative work of each module effectively solves industry pain points such as difficult environmental monitoring in complex forest areas, inaccurate risk assessment, and untimely emergency response. Embodiment 4 Please refer to Figure 2 , a forest fire prevention warning method based on the Internet of Things, including the following steps: Step 1: Construct a heterogeneous Internet of Things network through anti-occlusion millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes to collect 3D thermal images, surface temperature and humidity gradients, combustible moisture content, and micro-meteorological data in real time, eliminating monitoring blind spots; Step 2: Calibrate, abnormally eliminate, and compensate for the collected data, divide grid cells according to the terrain of the shelter forest, construct a digital elevation model of the shelter forest, and calculate the terrain occlusion coefficient Dmxs and the signal coverage intensity Xgql of each grid cell; Step 3: Fuse the terrain occlusion coefficient Dmxs and the signal coverage intensity Xgql, and generate a real-time fire risk index Hxzs; extract the heat radiation intensity Rsqd, the combustible dryness index Gzzs, and the air flow disturbance coefficient Qlys in the grid cell, and calculate the hidden fire source probability Yhgl; Step 4: Based on the spatio-temporal evolution of the real-time fire risk index Hxzs, predict the fire spread path and the influence coefficient of key infrastructure; and conduct real-time assessment of the hidden fire source probability Yhgl, and generate a joint warning index value Lyzb when the hidden fire source probability Yhgl exceeds the threshold; Step 5: Evaluate the joint warning index value Lyzb and start a hierarchical response strategy, ensure communication in weak signal areas by combining low-earth orbit satellites and Mesh networks, and dynamically push the resource scheduling priority and the optimized evacuation path plan.

[0030] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The forestry fire prevention early warning system based on the Internet of Things is characterized by: It includes a multi-dimensional perception module, an edge computing module, a risk assessment module, an intelligent decision-making module, and an emergency linkage module; The multi-dimensional perception module is used to construct a heterogeneous Internet of Things network through anti-occlusion millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes, and collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content, and micro-meteorological data in real time to eliminate monitoring blind spots; The edge computing module is used to calibrate, eliminate anomalies, and compensate for missing data in the collected data, divide the grid units according to the terrain of the shelter forest, construct a digital elevation model of the shelter forest, and calculate the terrain occlusion coefficient Dmxs and signal coverage intensity Xgql of each grid unit; The risk assessment module is used to fuse the terrain occlusion coefficient Dmxs and signal coverage intensity Xgql, and generate a real-time fire risk index Hxzs; extract the heat radiation intensity Rsqd, fuel dryness index Gzzs, and air flow disturbance coefficient Qlys in the grid unit, and calculate the probability of hidden fire sources Yhgl; The intelligent decision-making module is used to predict the fire spread path and the impact coefficient of key infrastructure based on the spatio-temporal evolution of the real-time fire risk index Hxzs; and conduct a real-time assessment of the probability of hidden fire sources Yhgl, and generate a joint warning index value Lyzb when the probability of hidden fire sources Yhgl exceeds the threshold; The emergency linkage module is used to evaluate the joint warning index value Lyzb and initiate a hierarchical response strategy, ensure communication in weak signal areas by combining low-earth orbit satellites and Mesh networks, and dynamically push the resource scheduling priority and evacuation path optimization plan.

2. The forest fire prevention early warning system based on the Internet of Things according to claim 1, wherein: The multi-dimensional perception module constructs a heterogeneous monitoring network by deploying millimeter-wave radar nodes, multi-spectral imaging nodes, and LoRa relay nodes. Among them, the millimeter-wave radar nodes use a penetrating detection method to obtain three-dimensional thermal distribution data in the vegetation-covered area. The multi-spectral imaging nodes collect surface temperature and humidity gradients and fuel moisture content data through the cooperation of visible light and infrared bands. The LoRa relay nodes realize low-power long-distance transmission of monitoring data; each node forms an adaptive network based on the spatial topological relationship and automatically enables multi-hop relay transmission when local communication is blocked; The multi-source perception data is unified in time scale through a time synchronization mechanism to form a three-dimensional environmental data set containing spatial coordinates, time stamps, and monitoring values, and is uploaded to the edge computing module in real time.

3. The forest fire prevention early warning system based on the Internet of Things according to claim 1, characterized in that: The edge computing module includes a data preprocessing unit and a grid analysis unit; The data preprocessing unit is used to detect and eliminate outliers in the collected three-dimensional thermal images and surface temperature and humidity gradient data, and compensate for missing data using an adaptive Kalman filter algorithm; establish a data correlation model based on time series analysis to perform spatio-temporal alignment of multi-source heterogeneous data.

4. The forest fire prevention early warning system based on the Internet of Things according to claim 1, wherein: The grid analysis unit is used to divide the monitoring area into several dynamic grid units and number them FH1, FH2,..., FHn according to the digital elevation model of the shelter forest and the three-dimensional environmental data set. The terrain occlusion coefficient Dmxs of each grid unit is calculated by a line-of-sight analysis algorithm, taking into account slope, vegetation density, and obstacle height; the signal coverage intensity Xgql is obtained by fusing the measured 5G signal intensity and the ray tracing model.

5. The forest fire prevention warning system based on the Internet of Things according to claim 1, characterized in that: The risk assessment module includes a fire risk index calculation unit and a concealed fire source calculation unit; The fire risk index calculation unit is used to perform the following calculation process: First, extract the terrain shielding coefficient Dmxs and the signal coverage intensity Xgql and perform normalization processing. Then, use the following formula to calculate and output the real-time fire risk index Hxzs: 。 6. The forest fire prevention early warning system based on the Internet of Things according to claim 1, characterized in that: The concealed fire source calculation unit is used to obtain the following parameters from each grid cell in real time: The thermal radiation intensity Rsqd collected by the infrared thermal imager, the combustible dryness index Gzzs calculated based on the soil moisture sensor and meteorological station data, and the air flow disturbance coefficient Qlys calculated through the wind speed sensor and terrain data; After extracting the thermal radiation intensity Rsqd, the combustible dryness index Gzzs, and the air flow disturbance coefficient Qlys and performing dimensionless processing, use the following formula to calculate and obtain the concealed fire source probability Yhgl: 。 7. The forest fire prevention early warning system based on the Internet of Things according to claim 1, characterized in that: The intelligent decision-making module includes a spatio-temporal evolution analysis unit; The spatio-temporal evolution analysis unit is used to construct a spatio-temporal propagation model of fire spread based on the dynamic change characteristics of the real-time fire risk index Hxzs, combined with the spatial position data and historical evolution rules of the monitoring area; and use the multi-path analysis algorithm and the key influence domain identification method to predict the possible fire spread paths.

8. The forest fire prevention early warning system based on the Internet of Things according to claim 6, characterized in that: The intelligent decision-making module also includes a joint warning assessment unit; The joint warning assessment unit is used to evaluate the concealed fire source probability Yhgl; compare the concealed fire source probability Yhgl with the preset concealed fire source probability threshold Y; the specific assessment content is as follows: When the concealed fire source probability Yhgl ≤ the concealed fire source probability threshold Y, it is determined that the concealed fire source risk in the current monitoring area is in a controllable state; When the concealed fire source probability Yhgl > the concealed fire source probability threshold Y, automatically trigger the concealed fire source over-threshold response mechanism, including recording the concealed fire source probability over-limit event, extracting the fire source positioning error Wcsp and the emergency response timeliness Yxsx during the over-threshold period and performing dimensionless processing; Use the following formula to calculate and obtain the joint warning index value Lyzb: 。 9. The forest fire prevention early warning system based on the Internet of Things according to claim 8, characterized in that: After the emergency linkage module obtains the joint warning index value Lyzb in real time and evaluates it, preset the response level division standard, including the first response level division threshold L1 and the second response level division threshold L2; and the first response level division threshold L1 is greater than the second response level division threshold L2; Perform hierarchical determination on the joint warning index value Lyzb according to the response level division standard; the specific content is as follows: When the joint warning index value Lyzb is less than or equal to the second response level division threshold L2, it is determined that the current monitoring area is in a low-risk state, and start the basic monitoring and information notification mechanism; When the joint warning index value Lyzb is greater than L2 and less than or equal to the first response level division threshold L1, it is determined that the current monitoring area is in a medium-risk state, and automatically trigger the resource scheduling preparation and key area risk warning process; When the combined early warning index value Lyzb is greater than the first response level division threshold L1, it is determined that the current monitoring area is in a high-risk state, and comprehensive emergency response measures are immediately initiated, including linkage disposal strategies such as prioritized allocation of emergency resources, communication reinforcement in weak signal areas, security control of key areas, and pushing of the optimal evacuation path.

10. A forest fire prevention early warning method based on the Internet of Things, according to the forest fire prevention early warning system based on the Internet of Things described in any one of claims 1-9, characterized in that: It includes the following steps: Step 1: Construct a heterogeneous Internet of Things network through anti-occlusion millimeter-wave radar, multi-spectral imaging, and LoRa relay nodes to collect three-dimensional thermal images, surface temperature and humidity gradients, fuel moisture content, and micro-meteorological data in real time, eliminating monitoring blind spots; Step 2: Calibrate, abnormally eliminate, and compensate for missing the collected data, divide grid units according to the terrain of the shelter forest, and construct a digital elevation model of the shelter forest to calculate the terrain shielding coefficient Dmxs and signal coverage intensity Xgql of each grid unit; Step 3: Integrate the terrain shielding coefficient Dmxs and signal coverage intensity Xgql, and generate a real-time fire risk index Hxzs; extract the thermal radiation intensity Rsqd, fuel dryness index Gzzs, and air flow disturbance coefficient Qlys in the grid unit, and calculate the probability of hidden fire sources Yhgl; Step 4: Based on the spatio-temporal evolution of the real-time fire risk index Hxzs, predict the fire spread path and the influence coefficient of key infrastructure; and conduct real-time evaluation of the probability of hidden fire sources Yhgl, and generate a combined early warning index value Lyzb when the probability of hidden fire sources Yhgl exceeds the threshold; Step 5: Evaluate the combined early warning index value Lyzb and initiate a hierarchical response strategy, ensure communication in weak signal areas by combining low-earth orbit satellites and Mesh networks, and dynamically push the resource scheduling priority and evacuation path optimization plan.

Images