Mountain forest fire-fighting inspection method and system based on unmanned aerial vehicle cluster collaborative operation

Abstract

The invention provides a mountain forest fire-fighting inspection method based on unmanned aerial vehicle cluster collaborative operation, and the method comprises the following steps: 1, carrying out region division and equipment initialization, dividing a forest region into lanes, walkways and a stream zone based on a GIS map and field exploration, and inputting the coordinate information of the lanes, the walkways and the stream zone into an unmanned aerial vehicle and inspection vehicle control system; 2, setting inspection grades according to forest region division, and planning an inspection path and inspection time according to the inspection grades to detect fire behaviors; 3, fire behavior grades are divided according to the sizes of the fire behaviors, and the corresponding unmanned aerial vehicle clusters are dispatched to extinguish the fire behaviors according to the fire behavior grades; and 4, material supply and fire scene review. According to the invention, real-time routing inspection can be carried out on the mountain forest without depending on fixed monitoring equipment, the deployment cost is low, and the monitoring reliability is higher.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) technology, and in particular to a method and system for forest fire patrol based on UAV swarm collaborative operation. Background Technology

[0002] With socio-economic development and the enhancement of people's spiritual needs, forests surrounding cities have become important venues for holiday leisure and recreation as well as traditional ancestral worship activities. However, the open flame activities such as smoking, picnicking, and ancestral worship in forest areas, coupled with the weak safety awareness and complacency of some people, have resulted in human-caused forest fires accounting for more than 90% of all forest fires.

[0003] Patent No. 202510337573.1 discloses a deep learning-based inspection drone fire emergency system. This system uses drone thermal imaging to monitor fires and provides real-time feedback on forest fire conditions, predicting fire spread trends and improving the targeting and effectiveness of emergency response. However, forest environments are characterized by sparse populations, complex terrain, and dense vegetation, making early fire detection difficult. Fires are often delayed until they become uncontrollable, causing irreversible damage to the ecological environment, seriously threatening human and livestock lives, and resulting in huge economic losses.

[0004] Patent application number 202410751491.7 discloses a fire suppression system based on a drone swarm. This system attempts to improve fire suppression efficiency through a coordinated approach involving fixed-wing reconnaissance drones for inspection, multi-rotor fire suppression drones for bombing, and ground support teams for ammunition replenishment. However, this solution relies on pre-deployed monitoring cameras in the forest as the front-end for fire monitoring, which has the following drawbacks: High deployment costs: The procurement, installation, and maintenance costs of large-scale forest area monitoring equipment are enormous; Poor environmental adaptability: The monitoring lines are easily damaged by natural factors such as tree root growth and branch friction, resulting in insufficient stability of the monitoring system; Meanwhile, existing drone technology lacks a clear focus for forest area inspections, but forest areas are vast, which poses significant challenges for drone inspections. If drones conduct numerous inspections frequently, the inspection cost is high; conversely, if they conduct few inspections frequently, fires cannot be detected in a timely manner. Summary of the Invention

[0005] Given the significant limitations of existing technologies in real-time monitoring, low-cost deployment, and emergency response to forest fires, this application proposes a forest fire patrol method based on drone swarm collaborative operations. This method can reduce fire hazards in the forest floor and is a technical solution that can autonomously complete fire patrols and extinguishing without relying on fixed monitoring equipment.

[0006] This application also proposes a forest fire patrol system based on drone swarm collaborative operation, corresponding to the above method.

[0007] A method for forest fire patrol based on drone swarm collaborative operation includes the following steps: Step 1: Area division and device initialization Based on GIS maps and on-site surveys, the forest area was divided into driveways, footpaths, and stream zones, and the coordinate information of the driveways, footpaths, and stream zones was entered into the control system of drones and inspection vehicles. Step 2: Set patrol levels according to forest area divisions, and plan patrol routes and patrol times according to patrol levels to detect fires; Step 3: Classify the fire severity level according to the size of the fire, and dispatch the corresponding drone swarm to extinguish the fire according to the severity level; Step 4: Replenish supplies and re-inspect the fire site.

[0008] A forest fire patrol system based on drone swarm collaborative operation includes a patrol vehicle, a drone swarm, a controller, and a communication transmission network; The inspection vehicle is equipped with a drone swarm, a communication transmission network, and a controller. The drone swarm is located inside the inspection vehicle, and the drone swarm includes: The reconnaissance team consists of lightweight fixed-wing drones and small multi-rotor drones. The lightweight fixed-wing drones are responsible for patrolling above the forest area; the small multi-rotor drones are used for patrolling narrow walking paths or stream areas. Spraying group: Uses multi-rotor drones, carrying dry powder or water-based extinguishing agents, for suppressing small fires; works in conjunction with bombing group for medium fires; assists bombing group and water intake group for large fires; Bombing team: uses multi-rotor drones, carrying 1-2 dry powder fire extinguishing bombs per flight, for medium-sized fires; auxiliary water intake team for large fires; Water intake team: Uses multi-rotor drones equipped with water tanks to draw water from streams, reservoirs, or other water sources via water pumps for use in large-scale fires; Warning Team: Equipped with a loudspeaker for use in warning personnel; The controller is capable of controlling the collaborative operation of the drone swarm; The communication transmission network is used for communication between the UAV cluster and the controller, as well as communication between the controller and the monitoring terminal.

[0009] The forest fire patrol system based on drone swarm collaborative operation in this application also includes a fixed-point supply station, which is used for the supply of fire-fighting materials, including fire suits, fire extinguishers, water-collecting drones, and fire extinguishing bombs.

[0010] The inspection vehicle in this application is also equipped with solar panels and a charging power supply. The solar panels are electrically connected to the charging power supply, which is used to replenish the power of the drone swarm.

[0011] The technical advantages of this application are as follows: Compared with existing drone high-altitude forest patrols, this application can detect hidden temperature anomalies under the forest canopy, promptly detect fires and illegal fire use, reduce fire losses, and regulate people's behavior in forest areas.

[0012] Since over 90% of forest fires are caused by human activity, this application prioritizes preventing human-caused fires. Based on human activity patterns in the forest, the area is divided into driveways, footpaths, and stream areas, with patrol levels established, patrol routes and times planned for fire detection. This prioritized approach effectively reduces the patrol area, saving patrol time and costs. High-frequency patrols in areas with high activity and population density, such as footpaths and stream areas, make it easier to detect and extinguish fires early, thus reducing fire damage. Furthermore, when a fire occurs, deploying appropriate drone swarms based on the fire severity level can reduce firefighting costs and conserve firefighting resources.

[0013] This application enables real-time inspection of forests without relying on fixed monitoring equipment, resulting in lower deployment costs and higher monitoring reliability. Attached Figure Description

[0014] Figure 1 This is a diagram illustrating the extinguishing of a small fire as described in this application; Figure 2 This is a diagram illustrating the firefighting process for a medium-sized fire as described in this application. Figure 3 This is a diagram illustrating the firefighting efforts for the large-scale fire described in this application. Figure 4 This is a schematic diagram of the forest fire patrol system based on drone swarm collaborative operation according to this application; Explanation of reference numerals in the attached diagram: Inspection vehicle 10, UAV swarm 20, reconnaissance group 201, spraying group 202, bombing group 203, water collection group 204, warning group 205, controller 30, communication transmission network 40. Detailed Implementation

[0015] The embodiments of the technical solution of this application will now be described in detail with reference to the accompanying drawings. These embodiments are only used to more clearly illustrate the technical solution of this application and are therefore merely examples, and should not be used to limit the scope of protection of this application.

[0016] A method for forest fire patrol based on 20-drone swarm collaborative operation includes the following steps: Step 1: Area division and device initialization Based on GIS maps and field surveys, the forest area is divided into driveways, walking paths, and stream areas, and the coordinate information of driveways, walking paths, and stream areas is entered into the control system of the UAV cluster 20 and the inspection vehicle 10. The lanes described in this application include, but are not limited to, asphalt roads, cement roads, gravel roads, dirt roads, wading roads, riverbeds, and drainage ditches. Any road that the inspection vehicle 10 can pass through is considered a lane. Lanes are the core channels for ground mobility and material transportation. Mountain roads and winding mountain roads can also serve as lanes.

[0017] The pedestrian paths described in this application are artificial or natural paths that are inaccessible to vehicles, such as mountain trails, stone slab paths, wooden boardwalks, paths formed by rainwater erosion, and dry riverbeds. These areas have frequent human activity and are highly concealed, and are difficult to detect due to the obstruction of trees.

[0018] The stream zone described in this application is a strip-shaped area extending a predetermined distance (e.g., 10 meters) to both sides of the stream's centerline, taking into account both the characteristics of water source areas and dense human activity.

[0019] Step 2: Set patrol levels according to forest area divisions, and plan patrol routes and patrol times according to patrol levels to detect fires; In a preferred embodiment, a three-tiered inspection system is established based on forest area divisions, wherein, Level 1 Inspection: The inspection area covers the pedestrian walkway and stream area. Centered on inspection vehicle 10, the inspection extends to the pedestrian walkway and stream area on both sides of the road. For single pedestrian walkways or stream areas, a reconnaissance drone (UAV) of reconnaissance team 201 is dispatched to patrol along a zigzag path within a range of 20-40 meters on both sides, monitoring heat sources with temperatures >60℃ and human activity. After reaching the mountaintop, it returns to the inspection area along the same route. This process is repeated until the UAV's battery level drops below a predetermined value, at which point it enters inspection vehicle 10 to replenish its power.

[0020] In a preferred embodiment, high-frequency activity points such as camping areas, viewing platforms, ancestral cemeteries, and rest pavilions are pre-marked on a GIS map, and a key monitoring circle with a radius of 10 to 50 meters is set up. For key monitoring areas, warning teams deploy drones (205) to remain stationed for at least 10 minutes to continuously monitor for heat sources exceeding 60°C. Open flame activity is confirmed through remote manual identification and control system-assisted judgment. If suspected or confirmed open flame activity is detected, the drone broadcasts warnings such as "Smoking is prohibited in the forest area; please comply." The activity is also recorded by taking photos. If open flame activity is detected again during a patrol and facial recognition confirms it is the same person, the drone transmits both records to patrol vehicle 10, arranging for the nearest patrol personnel to persuade the tourist to leave. The incident is also registered and recorded for future monitoring.

[0021] During Level 1 inspections, key prevention and control measures are implemented at densely populated areas, with the following specific measures: The AI ​​video analysis module on the 201 reconnaissance drone is used to identify crowd gatherings in real time. Based on the crowd gathering behavior, new temporary activity points are marked, such as roadside picnic areas. Then, the 205 warning drone is dispatched to stay for more than 10 minutes. If abnormal temperature is detected, the drone will broadcast warning messages such as "Smoking is prohibited in the forest area. Please abide by the rules."

[0022] Level 2 Inspection: The inspection range is within 20 meters on both sides of the lane. One reconnaissance drone is dispatched on each side to patrol along the lane, monitoring heat sources with temperatures above 60°C and personnel activities. After the inspection mission is completed, the drone returns to the inspection vehicle 10 to replenish power. The inspection frequency is 1 to 2 round trips per day for the same lane.

[0023] Level 3 Inspection: The inspection area covers the area above the forest. Fixed-wing drones conduct a large-scale patrol once a day, generating regional temperature heat maps. The inspection frequency can be increased during holidays.

[0024] During the inspection, the drone transmits the coordinates of abnormal heat sources, temperature values, and on-site images to the inspection vehicle's control system via 5G / LoRa and other communication networks, generating a dynamic monitoring dashboard and automatically highlighting abnormal data in red for early warning.

[0025] In order to promptly detect temperature anomalies in the forest under cover, in a preferred embodiment, the UAV of this application flies at an altitude of 3 to 10 meters during inspections. Since this application inspects along driveways, pedestrian paths, and streams, it can avoid obstacles such as trees, thereby enabling ultra-low-altitude flight.

[0026] Step 3: Classify the fire severity level according to the size of the fire, and dispatch a drone swarm of 20 to extinguish the fire according to the severity level; The unmanned aerial vehicle cluster 20 described in this application is divided into the following categories according to function: Reconnaissance Team 201 includes lightweight fixed-wing drones and small multi-rotor drones. The lightweight fixed-wing drones are responsible for patrolling above the forest area; the small multi-rotor drones are used for patrolling narrow walking paths or stream areas; the drones in Reconnaissance Team 201 carry non-contact, long-range temperature sensors such as thermal imaging sensors or infrared temperature sensors.

[0027] Spraying Group 202: Uses a multi-rotor drone, carrying dry powder or water-based extinguishing agents, for suppressing small fires; works in conjunction with Bomb Drop Group 203 for medium fires; assists Bomb Drop Group 203 and Water Collection Group 204 for large fires. Bombing Team 203: Employs a multi-rotor drone, carrying 1-2 dry powder fire extinguishing bombs per flight, for medium-sized fires; Auxiliary Water Supply Team 204 is used for large-scale fires; Water intake team 204: Uses a multi-rotor drone equipped with a water tank to draw water from streams, reservoirs, or other water sources via a water pump for use in large-scale fires; Warning Group 205: Equipped with a loudspeaker for personnel warning.

[0028] The inspection vehicle 10 of this application serves as a mobile base station, mobile supply station, and drone dispatch station, and is equipped with a controller 30, communication relay equipment, drones and their take-off and landing platforms, as well as fire extinguishing bombs, solar panels, and charging power supplies. The inspection vehicle 10 moves along the lane.

[0029] Step 4: Replenish supplies and re-inspect the fire site.

[0030] Material replenishment falls into two categories. When there is no fire, the supplies in the patrol vehicle are inspected and replenished promptly. For example, fire extinguishing bombs and dry powder sprayers are checked to ensure they are not expired and need replacement. The water level in the reservoir is also checked to see if it needs replenishing. The water source in the stream area is also checked to ensure it meets the water intake requirements. Finally, the supplies at the fixed supply points are checked to see if they need replenishing.

[0031] After a fire breaks out and is extinguished, in addition to the above-mentioned measures, a follow-up inspection of the fire scene is necessary. In a preferred implementation, after the fire is extinguished, inspection personnel first follow up and ensure that any open flames within sight are extinguished by means of trampling, watering, and covering with soil. Half an hour after the personnel leave, a reconnaissance team dispatches a 201 drone to conduct a thermal imaging inspection to see if the fire scene temperature has dropped to a normal level. Afterward, an inspection is conducted every two hours until the end of the day's inspections.

[0032] Since temperature sensors determine the presence of a fire by detecting temperature, in a preferred embodiment, a temperature of 50-100 degrees Celsius is considered abnormal, and the drone needs to remain stationary and monitor whether the temperature continues to rise. A temperature of 100-200 degrees Celsius indicates a suspected open flame; if there are no tourists nearby, the spraying drone group 202 will be deployed to cool the suspected fire area. If there are tourists nearby, they will be warned first, and the suspected fire area will be cooled down. A temperature above 200 degrees Celsius confirms an open flame, and the fire will be extinguished according to its size.

[0033] In a preferred embodiment, the fires described in this application are categorized by size as follows: Small-scale fire: fire area < 10㎡; Medium-sized fire: fire area 10~50㎡; Large-scale fire: Fire area > 50㎡.

[0034] In a preferred embodiment, the handling methods for various types of fires are as follows: For small fires, dispatch 1 to 5 spray-group 202 drones to spray dry powder or water for fire extinguishing.

[0035] When the spraying drones 202 work together, in a preferred embodiment, the fire area is divided into several grids of equal size. Each spraying drone is responsible for 1 to 4 grids, and coordinate boundaries are assigned through a vehicle-mounted system or terminal system to avoid duplication of work.

[0036] For example, please see Figure 1 Based on the area and shape of the fire, the fire area is divided into two grids, left and right, with jet drone A and jet drone B each responsible for extinguishing the fire in one grid.

[0037] For medium-sized fires, deploy 1-3 spray-type 202 drones and 2-5 bomb-dropping 203 drones to extinguish the fire using fire extinguishing bombs in conjunction with dry powder and water sprays. Specifically, the bomb-dropping 203 drones drop fire extinguishing bombs in the core area of ​​the fire, while the spray-type drones spray dry powder and water to extinguish the fire in the outer areas and areas in the center of the fire not covered by fire extinguishing bombs. Afterwards, the spray-type 202 drones spray water to cool down the high-temperature areas based on thermal imaging.

[0038] When the spraying drone group 202 and the bomb-dropping drone group 203 operate in coordination, in a preferred embodiment, the fire area is divided into several non-interfering circular areas based on the operating radius of the fire extinguishing bombs, and coordinates are generated at the center of each circle. One fire extinguishing bomb is dropped into each circular area, and coordinate boundaries are assigned via the vehicle-mounted control system or terminal control system to avoid redundant operations. Simultaneously, the fire area outside the circular areas is divided into several zones using a grid, with each spraying drone responsible for one zone. Coordinate boundaries are assigned via the vehicle-mounted control system or terminal control system to avoid redundant operations.

[0039] For example, please see Figure 2 The fire area was first divided into five circular zones according to the working radius of the fire extinguishing bombs, and five fire extinguishing bombs were dropped according to the coordinates. The remaining area not covered by the fire extinguishing bombs was then divided into two zones, left and right, using a grid. Reconnaissance Team 201 UAV A and Reconnaissance Team 201 UAV B were each responsible for one fire extinguishing zone.

[0040] For large-scale fires, more than three spraying drones (202), more than five bomb-dropping drones (203), and all water-collecting drones (204) are deployed for firefighting. Specifically, bomb-dropping drones (203) fly to the fire scene first and drop fire extinguishing bombs from the outside in. After dropping all the fire extinguishing bombs, if the fire is still not extinguished, they immediately fly back to patrol vehicle (10) to reload fire extinguishing bombs. Simultaneously, water-collecting drones (204) collect water from the stream, then fly to the fire area to extinguish the fire or cool it down. After emptying the water, they immediately fly to the water source to collect more water. During this process, spraying drones (202) spray the remaining embers to prevent reignition in the surrounding area until the open flames are extinguished. After the fire is extinguished, water-collecting drones (204) and spraying drones (202) spray water on the high-temperature areas to cool them down based on thermal imaging.

[0041] For example, please see Figure 3 The bombing group 203 drone dropped bombs to extinguish the fire on the periphery. The spraying group 202 drone sprayed water or dry powder extinguishing agent to extinguish the fire in the edge areas not covered by the fire extinguishing bombs. The water intake group 204 drone sprayed water to extinguish the fire in the inner areas not covered by the fire extinguishing bombs.

[0042] For medium and large fires, the fire department must be notified promptly to prevent the fire from escalating beyond the control of drones. Simultaneously, under safe conditions, the warning drone group 205 will organize nearby volunteers to participate in firefighting. If the fire becomes uncontrollable, the warning drone group 205 will issue an alarm to evacuate relevant personnel from the forest area.

[0043] When the spraying drone 202, bomb-dropping drone 203, and water-collecting drone 204 work together, the primary task is to prevent the fire from spreading. Therefore, the bomb-dropping drone 203 needs to immediately rush to the fire scene to control the fire. In a preferred embodiment, based on the burned area collected by the reconnaissance drone, the control system first divides the perimeter of the fire scene into multiple continuous and non-interfering circular areas according to the operating radius of the fire extinguishing bombs. One fire extinguishing bomb is dropped into each circular area, and coordinate boundaries are assigned through the vehicle-mounted system or terminal system to avoid duplicate operations. Simultaneously, the outer portion between two adjacent circular areas is divided into several zones and separated by a grid. Each spraying drone is responsible for one zone, and coordinate boundaries are assigned through the vehicle-mounted control system or terminal control system to avoid duplicate operations. Then, based on the fire scene conditions, the perimeter of the unextinguished fire area is divided into several fire-fighting zones and their coordinates are determined according to the water storage capacity of the water-collecting drone 204 and the distribution of the surrounding fire scene. Each water-collecting drone 204 is responsible for one fire-fighting zone, and coordinate boundaries are assigned through the vehicle-mounted control system or terminal system to avoid duplicate operations. After dropping its bombs, bombing team 203 immediately flew away to replenish its ammunition. Water collection team 204 continued firefighting efforts according to pre-defined designated areas. After running out of water, the water collection team 204 would collect water from nearby streams, reservoirs, or other water sources and then return to firefighting operations. If the water source was close to the fire area, it could be connected to a water pipe for continuous firefighting. When bombing team 203 reloaded and returned to the field, the control system similarly divided the fire area's perimeter into multiple continuous and non-interfering circular zones based on the fire extinguishing bomb's operating radius. One fire extinguishing bomb was dropped into each circular zone, with coordinate boundaries assigned via the vehicle-mounted control system or terminal control system to avoid redundant operations. This process continued until the fire was extinguished.

[0044] This application's forest fire patrol method based on 20 drone swarms also implements the following firefighting strategies according to wind direction: When the wind force is less than level 3, drones will circle the fire site to fight the fire. When the wind force is level 3 to 5, the drone will extinguish the fire by partially surrounding the fire site upwind, and set up an isolation zone at a predetermined distance from the fire site downwind.

[0045] When the wind force is greater than level 5, it is difficult for small drones to perform inspection and firefighting tasks, and inspections are generally not carried out in such weather.

[0046] In a preferred embodiment, in step 1, based on GIS maps and on-site surveys, the forest area is divided into high-voltage power line zones, and the understory area and high-voltage power lines within a 10-30 meter radius on both sides of the power lines are inspected. When people play on the edge of the forest area, they may fly kites, balloons, or discard plastic bags. Under the influence of wind, these items may float in the air and potentially cause a fire if they come into contact with the high-voltage power lines erected above the forest area. At the same time, animals such as snakes and birds may also be electrocuted, causing a fire.

[0047] For the forest underpass area under high-voltage power lines, a two-level inspection system is implemented, which involves a round trip inspection once a day to check for any abnormal temperatures in the forest. For the high-voltage power lines themselves, a three-level inspection system is implemented, with fixed-wing drones conducting large-scale inspections to check for any foreign objects along the power lines. If any foreign objects are found, the relevant departments are notified to remove them and eliminate potential hazards.

[0048] Please refer to Figure 4 A method for forest fire patrol based on 20-drone swarm collaborative operation includes the following steps: Step 1: Area division and device initialization Based on GIS maps and field surveys, the forest area was divided into driveways, walking paths, and stream belts, and the coordinate information of the driveways, walking paths, and stream belts was entered into the control system of the drone and inspection vehicle. Step 2: Set patrol levels according to forest area divisions, and plan patrol routes and patrol times according to patrol levels to detect fires; Step 3: Classify the fire severity level according to the size of the fire, and dispatch the corresponding drone swarm of 20 to extinguish the fire according to the severity level; Step 4: Replenish supplies and re-inspect the fire site.

[0049] A forest fire patrol system based on the collaborative operation of a drone swarm 20 includes a patrol vehicle 10, a drone swarm 20, a controller 30, and a communication transmission network 40. The inspection vehicle 10 is equipped with a drone cluster 20, a communication transmission network 40, and a controller 30. The drone swarm 20 is located inside the inspection vehicle 10, and the drone swarm 20 includes: Reconnaissance Team 201: Includes lightweight fixed-wing drones and small multi-rotor drones. Lightweight fixed-wing drones are responsible for patrolling above the forest area; small multi-rotor drones are used for patrolling narrow walking paths or stream areas. Spraying Group 202: Uses a multi-rotor drone, carrying dry powder or water-based extinguishing agents, for suppressing small fires; works in conjunction with Bomb Drop Group 203 for medium fires; assists Bomb Drop Group 203 and Water Collection Group 204 for large fires. Bombing Team 203: Employs a multi-rotor drone, carrying 1-2 dry powder fire extinguishing bombs per flight, for medium-sized fires; Auxiliary Water Supply Team 204 is used for large-scale fires; Water intake team 204: Uses a multi-rotor drone equipped with a water tank to draw water from streams, reservoirs, or other water sources via a water pump for use in large-scale fires; Warning Group 205: Equipped with a loudspeaker for personnel warning; The controller 30 is capable of controlling the collaborative operation of the drone swarm 20; The communication transmission network 40 is used for communication between the UAV cluster 20 and the controller 30, as well as communication between the controller 30 and the monitoring terminal.

[0050] The forest fire patrol system based on the collaborative operation of a drone swarm 20 in this application also includes a fixed-point supply station. This station is used for replenishing firefighting supplies, including fire suits, fire extinguishers, water-fetching drones (204), and fire extinguishing bombs. When the fire extinguishing bombs in the patrol vehicle 10 run out, it can be replenished at the fixed-point supply station.

[0051] The inspection vehicle 10 of this application is also equipped with a solar panel and a charging power supply. The solar panel and the charging power supply are electrically connected, and the charging power supply is used to replenish the power of the drone swarm 20.

[0052] To verify the effectiveness of the inspection implementation proposed in this application, a simulated inspection experiment was conducted in a forest area of ​​Helan Mountain. The experiment took place on June 21, 2025. A 10-kilometer section of the road was selected for verification, extending 5 kilometers to each side of the road as the central axis, totaling 100 square kilometers. Along this section of the road, there were 24 walking trails and 3 streams. 37 locations were defined using natural open spaces along the walking trails, including viewing platforms, campsites, cemeteries, and rest pavilions. Forty people were mixed in with tourists and randomly simulated picnicking and smoking on the road, walking trails, and streams. Twenty people carried electric heaters to simulate picnicking, and 20 people carried electric heaters to simulate smoking. For safety reasons, the temperature of the electric heaters was uniformly set to 100 degrees Celsius. The size of the heat source of the electric heaters was similar to that of the actual heat source in the scene. These 40 people were randomly paired up and chose to participate in activities on the road, walking trails, and streams, randomly selecting any location along the route to simulate picnicking and smoking. Under natural conditions, picnics and smoking sessions cannot last for several hours. If we completely simulate real-world picnics and smoking, the chances of being detected by drones are low, making it difficult to compare the advantages and disadvantages of different implementation methods. To increase the chances of being detected while picnicking and smoking, electric heating is applied every 10 minutes for 5 minutes during the simulation.

[0053] The patrol began at 4 p.m. in the designated area, and by 7 p.m., the number of fires, such as those caused by picnicking and smoking, and the average time taken to identify them were tallied.

[0054] Example 1 The method for forest fire prevention and inspection based on a swarm of 20 UAVs in this application was used to control UAV patrols. A total of 12 reconnaissance UAVs (201) were deployed, with a flight speed of 30 km / h. Among them, one lightweight fixed-wing UAV was used for patrolling above the forest area at an altitude of 100 meters; one small multi-rotor UAV was used for patrolling within a 20-meter range on both sides of the road at an altitude of 5 meters; four small multi-rotor UAVs were used for patrolling footpaths and stream areas at an altitude of 5 meters; and six small multi-rotor UAVs were used for monitoring key monitoring areas and densely populated areas at altitudes of 3-10 meters.

[0055] Comparative Example 1 The same drone was used for the routine inspection. That is, a lightweight fixed-wing or multi-rotor drone was dispatched at a speed of 30 km / h and an altitude of 100 meters.

[0056] Both Example 1 and Comparative Example 1 implemented their inspection plans simultaneously at 4 PM. The statistical results are shown in Table 1. Table 1: Types, Number, and Duration of Fires, and Drone Travel Distance The average time in Table 1 refers to the average time from the drone's departure to the discovery of the fire.

[0057] As shown in Table 1, Example 1 was able to detect almost all fires, with an average detection time of 0.4 hours for picnics and 0.8 hours for smoking. The larger the area of ​​the temperature anomaly, the shorter the detection time. While Comparative Example 1 could detect a certain number of fires through high-altitude reconnaissance, it still lagged significantly behind Example 1, especially in detecting smoking. When smoking, the fire source area is very small, making it virtually undetectable by high-altitude reconnaissance. However, this application utilizes ultra-low-altitude focused inspections in high-traffic areas of forest areas, such as driveways, pedestrian paths, and stream areas, to promptly detect potential fire hazards, and the average time to detect smoking is much shorter than in Comparative Example 1.

[0058] Meanwhile, Table 1 shows that Example 1's average picnic time was longer than that of Comparative Example 1. Subsequent investigation revealed that 7 out of the 10 picnic scenarios were in open areas, and all of these were detected by aerial reconnaissance. The remaining three picnics occurred under trees, which were more concealed and were not detected by the aerial drone. In Example 1, the later detection time of the forest-based picnics increased the average detection time.

[0059] The experimental results above demonstrate that this application has a strong detection capability against small fire sources such as smoking and fire sources hidden in the forest.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for forest fire patrol based on unmanned aerial vehicle (UAV) swarm collaborative operation, characterized in that: Includes the following steps: Step 1: Area division and device initialization Based on GIS maps and field surveys, the forest area was divided into driveways, footpaths, and stream zones, and the coordinate information of the driveways, footpaths, and stream zones was entered into the control system of the drone cluster and inspection vehicle. Step 2: Set patrol levels according to forest area divisions, and plan patrol routes and patrol times according to patrol levels to detect fires; Step 3: Classify the fire severity level according to the size of the fire, and dispatch the corresponding drone swarm to extinguish the fire according to the severity level; Step 4: Replenish supplies and re-inspect the fire site.

2. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 1, characterized in that: In step 1, the GIS map pre-marks high-frequency activity points such as camping areas, viewing platforms, ancestral cemeteries, and rest pavilions, and sets up a key monitoring circle with a radius of 10 to 50 meters.

3. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 1, characterized in that: In step 1, the drone swarm is divided into: The reconnaissance team consists of lightweight fixed-wing drones and small multi-rotor drones. The lightweight fixed-wing drones are responsible for patrolling above the forest area; the small multi-rotor drones are used for patrolling narrow walking paths or stream areas. Spraying team: Employs multi-rotor drones equipped with dry powder or water-based extinguishing agents for suppressing small fires; The bomb-dropping team is used in conjunction with the bomb-dropping team for medium-sized fires; the auxiliary bomb-dropping team and water-fetching team are used for large-scale fires. Bombing team: uses multi-rotor drones, carrying 1-2 dry powder fire extinguishing bombs per flight, for medium-sized fires; auxiliary water intake team for large fires; Water intake team: Uses multi-rotor drones equipped with water tanks to draw water from streams, reservoirs, or other water sources via water pumps for use in large-scale fires; Warning Team: Equipped with a loudspeaker for use in warning personnel.

4. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 3, characterized in that: In step 2, a three-level inspection system is established based on the forest area division, among which, Level 1 Inspection: The inspection area is the pedestrian walkway and stream area. Centered on the inspection vehicle, it radiates to the pedestrian walkway and stream area on both sides of the road. For single pedestrian walkways or stream areas, a reconnaissance drone is dispatched to patrol along a zigzag path within a range of 20-40 meters on both sides to monitor heat sources with temperatures >60℃ and human activities. Level 2 inspection: The inspection range is within 20 meters on both sides of the lane. One reconnaissance drone is dispatched on each side to patrol along the lane and monitor heat sources with a temperature > 60℃ and human activities. Level 3 Inspection: The inspection area is above the forest area. Fixed-wing drones conduct a large-scale patrol once a day to generate regional temperature heat maps.

5. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 1, characterized in that: The drone flies at an altitude of 3 to 10 meters during inspections.

6. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 1, characterized in that: In step 3, the fire is categorized by size according to its severity: Small-scale fire: fire area < 10㎡; Medium-sized fire: fire area 10~50㎡; Large-scale fire: Fire area > 50㎡.

7. The forest fire patrol method based on UAV swarm collaborative operation as described in claim 1, characterized in that: In step 1, based on GIS maps and field surveys, the forest area is divided into high-voltage power line zones, and the understory area and high-voltage power lines within a range of 10 to 30 meters on both sides of the high-voltage power lines are inspected.

8. A forest fire patrol system based on drone swarm collaborative operation, characterized in that: Includes inspection vehicle (10), drone swarm (20), controller (30), and communication transmission network (40); The inspection vehicle (10) is equipped with a drone cluster (20), a controller (30), and a communication transmission network (40). The drone swarm (20) is located inside the inspection vehicle (10), and the drone swarm (20) includes: Reconnaissance Team (201): Includes lightweight fixed-wing UAVs and small multi-rotor UAVs. Lightweight fixed-wing UAVs are responsible for patrolling above the forest area; small multi-rotor UAVs are used for patrolling narrow walking paths or stream areas. Spraying Group (202): Uses a multi-rotor drone equipped with dry powder or water-based extinguishing agents for small fire suppression; works in conjunction with the bombing group for medium fires; and assists the bombing group and water intake group for large fires. Bombing team (203): Uses multi-rotor UAVs, carrying 1-2 dry powder fire extinguishing bombs per flight, for medium-sized fires; auxiliary water intake team for large fires; Water intake team (204): Uses a multi-rotor drone equipped with a water tank to draw water from streams, reservoirs, or other water sources via a water pump for use in large-scale fires; Warning group (205): Equipped with a loudspeaker for personnel warning; The controller (30) is capable of controlling the collaborative operation of the drone swarm (20); The communication transmission network (40) is used for communication between the UAV cluster (20) and the controller (30), as well as communication between the controller (30) and the monitoring terminal.

9. A forest fire patrol system based on UAV swarm collaborative operation as described in claim 8, characterized in that: It also includes designated supply stations, which are used for the supply of fire-fighting materials, including fire suits, fire extinguishers, water intake drones, and fire extinguishing bombs.

10. A forest fire patrol system based on drone swarm collaborative operation as described in claim 8, characterized in that: The inspection vehicle (10) is also equipped with a solar panel and a charging power supply. The solar panel and the charging power supply are electrically connected. The charging power supply is used to supplement the power of the drone swarm (20).

Images