An inspection UAV fire emergency system based on deep learning

Abstract

The present invention relates to the field of data analysis, and particularly to an inspection UAV fire emergency system based on deep learning. The present invention collects forest image data and forest environment data; determines the location of the fire source, identifies the tree planting distribution characteristics of the area where the fire source is located, and analyzes the characterization value of the spread trend of the fire; predicts the diffusion coverage area of the fire based on the spread trend characterization value combined with the wind direction data, determines the spread characterization parameter based on the terrain characteristics of the diffusion coverage area to determine the diffusion warning level of the diffusion coverage area; determines the terrain steepness amplitude corresponding to several feasible paths based on a predetermined distance of the diffusion coverage area to determine the available fire fighting path; and issues an alarm signal of the corresponding level according to the diffusion warning level determined by the diffusion analysis module. The present invention can improve the efficiency of UAV inspection, provide real-time feedback on forest fires, predict the spread trend of fires, and improve the pertinence and effectiveness of emergency response.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to an inspection UAV fire emergency system based on deep learning. Background Art

[0002] Traditional forest fire inspection mainly relies on methods such as manual patrol and watchtower monitoring. Among them, manual patrol has low efficiency, limited coverage, and it is difficult to carry out work under complex terrain and bad weather conditions; although watchtower monitoring can cover a certain range, there are monitoring blind spots and it is impossible to grasp the dynamics of forest fires in real time and comprehensively.

[0003] UAVs have the advantages of strong mobility, high flexibility, and can quickly reach the designated area, and can play an important role in forest fire fighting. By carrying devices such as high-definition cameras and infrared thermal imagers, UAVs can conduct all-round and real-time monitoring of the forest from the air, can timely detect fire hazards and the location of the fire source, and overcome the limitations of traditional monitoring methods. At the same time, UAVs can also conduct reconnaissance at the fire scene and provide detailed information about the fire scene for ground firefighters, such as the size of the fire, the spreading direction, the terrain and landforms, etc., to help firefighters formulate scientific and reasonable fire extinguishing plans.

[0004] As an important technology in the field of artificial intelligence, deep learning has powerful feature extraction and pattern recognition capabilities. In forest fire inspection, deep learning can analyze and process the image and video data collected by UAVs, automatically identify fire features such as flames and smoke, and improve the accuracy and timeliness of fire monitoring. In addition, deep learning can also predict the development trend of fires and provide a scientific basis for fire fighting decisions. For example, by learning historical fire data and real-time monitoring data, a fire spreading model is established to predict the spreading direction and speed of fires under different meteorological and terrain conditions, to help the fire department make early prevention and response measures.

[0005] With the continuous development of UAV technology and deep learning technology, a forest fire emergency system that combines the two can give full play to the monitoring advantages of UAVs and the intelligent analysis capabilities of deep learning, and achieve early detection, accurate warning, scientific decision-making, and efficient extinguishing of forest fires.

[0006] Chinese Patent Application Publication No.: CN119396184A discloses a forest fire prevention drone automatic inspection system based on meteorological factors, including: a meteorological data collection device, an unmanned airport and an edge computing device, a remote data analysis and flight management system, and drones; the meteorological data collection device is used to collect meteorological factor data under different meteorological conditions, special holidays or different seasons in real time; the unmanned airport and the edge computing device are used to receive the meteorological factor data, clean and process it to obtain the processed data, and control the drones to cruise the inspection area; the remote data analysis and flight management system is used to analyze the processed data, judge the forest fire risk meteorological levels in different situations, formulate cruise tasks for different drones, and send the cruise tasks of the drones; the drones automatically execute the cruise tasks according to the control commands; this system can automatically increase the frequency of forest fire prevention drone inspections according to special nodes, meet the actual inspection needs, and make forest fire prevention more efficient and intelligent.

[0007] However, the following problems still exist in the prior art.

[0008] During the process of using drones for forest fire prevention inspections, the inspections and warnings are carried out only based on the areas where fires occur in real time, without considering the influence of factors such as meteorological conditions, forest terrain and tree distribution on the spread of fires. Furthermore, the inspections and warnings for other areas affected by the spread of fires are ignored, resulting in incomplete inspections and low efficiency. Summary of the Invention

[0009] Therefore, the present invention provides a fire emergency system for inspection drones based on deep learning to overcome the problems in the prior art that during the process of using drones for forest fire prevention inspections, the inspections and warnings are carried out only based on the areas where fires occur in real time, without considering the influence of factors such as meteorological conditions, forest terrain and tree distribution on the spread of fires. Furthermore, the inspections and warnings for other areas affected by the spread of fires are ignored, resulting in incomplete inspections and low efficiency.

[0010] To achieve the above object, the present invention provides a fire emergency system for inspection drones based on deep learning, which includes:

[0011] A fire situation collection module, which includes an image collection unit for obtaining forest image data and an environment collection unit for collecting forest environment data;

[0012] A fire source analysis module, which is connected to the fire situation collection module, used to determine the fire source location, identify the tree distribution characteristics of the area where the fire source is located, and analyze the spread trend characterization value for the fire;

[0013] A diffusion analysis module, which is connected to the fire source analysis module, is used to predict the diffusion coverage area of the fire situation based on the spread trend characterization value in combination with wind direction data, determine the spread characterization parameters based on the terrain features of the diffusion coverage area, so as to determine the diffusion warning level of the diffusion coverage area;

[0014] A path matching module, which is respectively connected to the fire situation acquisition module and the diffusion analysis module, is used to determine the terrain steepness amplitude corresponding to a number of feasible paths based on a predetermined distance in the diffusion coverage area, so as to determine available fire-fighting paths;

[0015] An alarm emergency module, which is connected to the diffusion analysis module, is used to issue an alarm signal of the corresponding level according to the diffusion warning level determined by the diffusion analysis module;

[0016] Among them, the tree planting distribution characteristics include the accumulation density of vegetation in the area and the average distance between trees, and the terrain characteristics include the valley depth and the slope perpendicularity.

[0017] Further, the fire situation analysis module is used to determine the fire source position, including

[0018] It is used to call the forest image data collected by the image acquisition unit to determine the smoke distribution and the change of smoke concentration, so as to locate the fire source direction;

[0019] It is used to extract the heat distribution along the fire source direction, and determine the distribution position corresponding to the highest heat as the fire source position.

[0020] Further, the fire situation analysis module is used to analyze the spread trend characterization value for the fire situation, including

[0021] It is used to take the ratio of the accumulation density of vegetation in the area to the accumulation density threshold as the first spread trend feature;

[0022] It is used to take the ratio of the average distance threshold to the average distance between trees in the area as the second spread trend feature;

[0023] It is used to perform weighted summation on the first spread trend feature and the second spread trend feature as the spread trend characterization value.

[0024] Further, the diffusion analysis module is used to predict the diffusion coverage area of the fire situation based on the spread trend characterization value in combination with wind direction data, including,

[0025] It is used to obtain the wind direction in the area where the fire source is located;

[0026] It is used to determine the spread trend characterization values corresponding to a number of areas along the wind direction starting from the area where the fire source is located;

[0027] If there exists a spread trend characterization value corresponding to any area that is greater than or equal to the spread trend characterization threshold, then the area is predicted as the diffusion coverage area of the fire.

[0028] Further, the diffusion analysis module is used to determine the spread diffusion characterization parameters based on the terrain features of the diffusion coverage area, including,

[0029] using the ratio of the valley depth to the valley depth threshold as the first spread diffusion characteristic;

[0030] using the ratio of the slope perpendicularity to the slope perpendicularity threshold as the second spread diffusion characteristic;

[0031] using the sum of the first spread diffusion characteristic and the second spread diffusion characteristic as the spread diffusion characterization parameter.

[0032] Further, the diffusion analysis module is used to determine the diffusion alarm level of the diffusion coverage area, including,

[0033] presetting the correspondence between the diffusion alarm level and the preset spread diffusion characterization parameter interval;

[0034] determining the preset spread diffusion characterization parameter interval to which the spread diffusion characterization parameter belongs;

[0035] setting the diffusion alarm level corresponding to the preset spread diffusion characterization parameter interval as the diffusion alarm level of the diffusion coverage area;

[0036] wherein, the diffusion alarm level and the preset spread diffusion characterization parameter interval are in one-to-one correspondence.

[0037] Further, the path matching module is used to determine the feasible path, including,

[0038] used to determine several paths and determine the paths that meet the road condition benchmark conditions as the feasible paths;

[0039] wherein, the road condition benchmark conditions include that the road surface width is greater than the path width threshold and the bend radius is greater than the bend radius threshold.

[0040] Further, the path matching module is used to determine the available fire paths, including,

[0041] If there exists a terrain steepness amplitude of any feasible path that is less than the terrain steepness amplitude threshold, then the feasible path is determined as the available fire path.

[0042] Further, the predetermined distance is determined according to the range of the fire-fighting equipment.

[0043] Further, the alarm emergency module is used to issue an alarm signal of the corresponding level according to the diffusion alarm level determined by the diffusion analysis module, including

[0044] Pre-set the corresponding relationship between the alarm signal and the diffusion alarm level;

[0045] Among them, the alarm signal corresponds one-to-one to the diffusion alarm level, and the alarm signal includes the sounding of an alarm.

[0046] Compared with the prior art, the present invention is provided with a fire condition acquisition module, which includes an image acquisition unit for acquiring forest image data and an environment acquisition unit for acquiring forest environment data; a fire source analysis module for determining the fire source location, identifying the tree planting distribution characteristics of the area where the fire source is located, and analyzing the spread trend characterization value for the fire condition; a diffusion analysis module for predicting the diffusion coverage area of the fire condition based on the spread trend characterization value in combination with wind direction data, and determining the spread diffusion characterization parameter based on the terrain characteristics of the diffusion coverage area to determine the diffusion alarm level of the diffusion coverage area; a path matching module for determining the terrain steepness amplitude corresponding to several feasible paths based on a predetermined distance of the diffusion coverage area to determine the available fire fighting path; an alarm emergency module for issuing an alarm signal of the corresponding level according to the diffusion alarm level determined by the diffusion analysis module. The present invention can improve the efficiency of UAV patrol, provide real-time feedback on forest fires, predict the spread trend of forest fires, and improve the pertinence and effectiveness of emergency response.

[0047] In particular, the present invention is provided with a fire source analysis module, which analyzes the spread trend of a fire by determining the fire source location and combining the tree planting characteristics of the corresponding area. In actual situations, dense vegetation distribution means a large number of combustibles per unit area. When a fire occurs, a large amount of vegetation can continuously provide fuel for combustion, making it easier for the fire to spread and burn more violently. At the same time, due to the short distance between the densely distributed vegetation, heat is more easily transferred between the vegetation through conduction, radiation, and convection, enabling the fire to spread to the surrounding areas faster and expanding the scope of the fire's impact. Similarly, if the distance between the trees in the area where the fire source is located is too close, the flame is more likely to spread directly from one tree to another, expanding the scope of the fire's impact in a short time. At the same time, the effect of thermal radiation generated in the area where the trees are too close is more significant. Each burning tree radiates a large amount of heat to the surrounding trees, rapidly increasing the temperature of the adjacent trees and greatly increasing the possibility of reaching the ignition point. This chain reaction of thermal radiation will cause more trees to catch fire successively, further promoting the spread of the fire. Therefore, the present invention determines the spread trend characterization value of the fire by the vegetation accumulation density in the area where the fire source is located and the average distance between the trees, so as to characterize the degree of influence of the distribution of the trees and vegetation in the area where the fire source is located on the spread of the fire, providing data support for predicting the spread coverage area of the fire in the future. The present invention can improve the efficiency of drone patrol, provide real-time feedback on forest fires, predict the spread trend of the fire, and improve the pertinence and effectiveness of emergency response.

[0048] In particular, the present invention is provided with a diffusion analysis module, which predicts the spread coverage area of the fire based on the tree planting distribution characteristics and wind direction data. The vegetation and trees that can act as combustibles and affect the severity of the fire in a forest fire, their distribution determines the material basis for the spread of the fire. At the same time, combined with the wind direction, which is a factor that can directly affect the advancing direction of the fire, a comprehensive analysis is carried out to accurately predict the area that the fire may affect. Furthermore, combined with the terrain and topography of the corresponding area, the spread degree of the fire is determined. Therefore, the present invention determines the spread characterization parameter based on the terrain characteristics of the spread coverage area, so as to characterize the influence of the internal terrain structure of the area affected by the fire on the severity of the fire, providing data support for determining the spread warning level in the future. The present invention can improve the efficiency of drone patrol, provide real-time feedback on forest fires, predict the spread trend of the fire, and improve the pertinence and effectiveness of emergency response.

[0049] In particular, the present invention is provided with a path matching module. On the premise that the prediction of the spread coverage area of the fire is completed, the terrain steepness of the path at a certain distance from the corresponding area is considered to determine an available fire path that can allow fire-fighting equipment to pass smoothly and provide sufficient space for fire-fighting operations. Description of the Drawings

[0050] Figure 1 It is a functional module diagram of the inspection UAV fire emergency system based on deep learning in the invention embodiment;

[0051] Figure 2 It is a logical step diagram for predicting the diffusion coverage area of a fire in the invention embodiment;

[0052] Figure 3 It is a logical decision diagram for determining a feasible path in the invention embodiment;

[0053] Figure 4 It is a logical decision diagram for determining an available fire path in the invention embodiment. Detailed implementation manners

[0054] In order to make the objectives and advantages of the present invention clearer, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] The preferred implementation manners of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these implementation manners are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0056] It should be noted that in the description of the present invention, the terms indicating directions or positional relationships such as "upper", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0057] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] Please refer to Figure 1 As shown, it is a functional module diagram of the inspection UAV fire emergency system based on deep learning in the embodiment of the present invention. The inspection UAV fire emergency system based on deep learning in the embodiment of the present invention includes:

[0059] A fire situation acquisition module, which includes an image acquisition unit for acquiring forest image data and an environment acquisition unit for collecting forest environment data;

[0060] A fire source analysis module, which is connected to the fire situation acquisition module, is used to determine the fire source location, identify the tree planting distribution characteristics of the area where the fire source is located, and analyze the spread trend characterization value for the fire situation;

[0061] A diffusion analysis module, which is connected to the fire source analysis module, is used to predict the diffusion coverage area of the fire situation based on the spread trend characterization value combined with wind direction data, determine the spread diffusion characterization parameter based on the terrain characteristics of the diffusion coverage area, so as to determine the diffusion warning level of the diffusion coverage area;

[0062] A path matching module, which is respectively connected to the fire situation acquisition module and the diffusion analysis module, is used to determine the terrain steepness amplitude corresponding to several feasible paths based on a predetermined distance in the diffusion coverage area, so as to determine the available fire fighting paths;

[0063] An alarm emergency module, which is connected to the diffusion analysis module, is used to send out alarm signals of corresponding levels according to the diffusion warning level determined by the diffusion analysis module;

[0064] Among them, the tree planting distribution characteristics include the accumulation density of vegetation in the area and the average distance between trees, and the terrain characteristics include the valley depth and the slope perpendicularity.

[0065] Specifically, the specific structure of the image acquisition unit is not limited, as long as it has the function of collecting and obtaining forest image data.

[0066] In some possible implementations, a lidar is carried by a drone. The lidar emits a laser beam to the forest and measures the time from the emission of the laser to the reception of the reflection by the receiver, calculates the distance between the laser beam and the target object, obtains a large number of three-dimensional space coordinate points, and forms point cloud data. Based on the analysis of the point cloud data, the height and distribution of the vegetation are obtained, and then the vegetation accumulation density is calculated. For example, the vegetation coverage area can be obtained by calculating the projected area of the vegetation points on the horizontal plane. Among them, for a regular grid area, the number of grids occupied by the projected vegetation points in the grid can be counted, and then multiplied by the area of a single grid; for an irregular area, a polygon area calculation method can be used to connect the boundaries of the projected vegetation points into a polygon and calculate its area. Furthermore, the ratio of the vegetation coverage area to the total area of the region is used as the vegetation accumulation density;

[0067] The distance between trees is determined by identifying the positions and shapes of the trees, a high-precision digital terrain model is constructed to obtain the valley depth and the morphological information of the slope, so as to calculate the slope perpendicularity; and on the basis of the digital terrain model, by calculating the elevation difference and the horizontal distance between adjacent points, the slope value of the road area is obtained, and the slope value is determined as the terrain steepness amplitude. Of course, other methods can also be used to obtain forest image data, which will not be elaborated here;

[0068] Among them, the forest image data includes tree planting distribution characteristics, terrain characteristics, and terrain steepness amplitude, etc.

[0069] Specifically, the specific structure of the environmental acquisition unit is not limited, as long as it has the function of collecting forest environmental data.

[0070] In some possible implementations, a wind direction sensor is carried by a drone to obtain the wind direction information at the location of the drone in real time.

[0071] In some possible implementations, the drone is equipped with a high-precision anemometer. By measuring the flight speed changes of the drone in different directions, combining with the flight attitude and speed data of the drone itself, the wind direction is deduced using relevant algorithms, which will not be elaborated here.

[0072] Specifically, the method of dividing the forest area is not limited, as long as it meets the requirement of the drone to efficiently inspect the fire. For example, several square areas can be divided according to the area of the forest, and the drone is controlled to preferentially inspect the area where the fire source is located, and then combined with the wind direction data to inspect the adjacent surrounding areas of this area, which will not be elaborated here.

[0073] Specifically, the specific structures of the fire source analysis module, the diffusion analysis module, the path matching module, and the alarm and emergency module are not limited. It itself or each unit therein can be composed of logic components or a combination of logic components. The logic components include a field programmable processor, a computer, or a microprocessor in the computer.

[0074] Specifically, the fire situation analysis module is used to determine the location of the fire source, including

[0075] It is used to call the forest image data collected by the image acquisition unit to determine the smoke distribution and the change of smoke concentration, so as to locate the direction of the fire source.

[0076] It is used to extract the heat distribution along the direction of the fire source, and determine the distribution position corresponding to the highest heat as the location of the fire source.

[0077] It can be understood that the distribution of the smoke diffuses from the fire source to the surrounding under the action of the wind. Generally speaking, the closer to the fire source, the higher the smoke concentration. As the distance increases, the smoke will gradually dilute and the concentration will decrease. Observe the change of the smoke concentration along the direction of the smoke diffusion. At the same time, combined with the heat change caused by the fire, based on the fact that the heat at the fire source is much higher than the heat of the surrounding environment, the location of the fire source can be accurately determined, which will not be elaborated here.

[0078] In this embodiment, based on the learning of a large amount of forest fire smoke image data by a deep learning algorithm to distinguish smoke from other interfering data in the forest, such as shadows, fog, etc., a pre-trained deep learning model can automatically extract the features of smoke to identify the smoke;

[0079] For example, in implementation, the Mask RCNN model can be used. Mask RCNN is a powerful object detection and segmentation model that can accurately identify the smoke area in the image. Of course, other forms of models can also be used, and those skilled in the art can select according to requirements, which will not be elaborated here.

[0080] By using a drone to carry a thermal infrared camera, thermal infrared images of the forest area are collected from different heights and angles, and the heat distribution and changes in the thermal infrared images are analyzed to quickly locate the area with the highest heat, which will not be elaborated here.

[0081] Specifically, the fire situation analysis module is used to analyze the characterization value of the spread trend of the fire situation, including,

[0082] Using the ratio of the accumulation density of vegetation in the area to the accumulation density threshold as the first spread trend feature;

[0083] Using the ratio of the average spacing threshold to the average spacing between trees in the area as the second spread trend feature;

[0084] Using the weighted sum of the first spread trend feature and the second spread trend feature as the spread trend characterization value.

[0085] In an actual fire, the tree spacing has a greater impact on the spread of the fire. In implementation, the impact of the spacing between trees on the spread of the fire is considered preferentially. Therefore, a slightly higher weight is given to the second spread trend feature calculated based on the average spacing. When performing weighted summation, the weight of the first spread trend feature is set to 0.45, and the weight of the second spread trend feature is set to 0.55;

[0086] In this embodiment, the purpose of setting the accumulation density threshold of vegetation and the average spacing threshold between trees is to characterize the situation where the vegetation and tree distribution in each area are relatively dense and prone to aggravating the spread trend of the fire. Among them, the accumulation density threshold of vegetation is preset. By pre-obtaining a number of historical data related to forest fire records, calling the historical data of the accumulation density of vegetation, and solving the average value of the accumulation density of vegetation, the accumulation density threshold of vegetation is set as the product of the average value of the accumulation density and the offset coefficient, and the offset coefficient is selected within the interval [1.05, 1.1];

[0087] It can be understood that when the distance between trees is short, the branches and leaves of the trees are close to each other, and the flame is likely to spread from one tree to another, forming a crown fire. The burning speed of the crown fire is fast, the intensity is high, and it is difficult to extinguish, which will cause serious damage to the forest. Therefore, in this embodiment, the average spacing threshold between trees is selected within the interval [3m, 4m].

[0088] Specifically, the present invention sets a fire source analysis module to analyze the spread trend of the fire situation by determining the fire source position in combination with the tree planting characteristics of the corresponding area. In actual situations, dense vegetation distribution means a large number of combustibles per unit area. When a fire occurs, a large amount of vegetation can continuously provide fuel for combustion, making the fire easier to spread and burn more violently. At the same time, due to the short distance between the densely distributed vegetation, heat is more easily transferred between the vegetation through conduction, radiation, and convection, enabling the fire to spread to the surrounding areas faster and expanding the scope of the fire. For example, in areas with dense shrubs, the heat generated by the flame can quickly be transferred to adjacent shrubs, causing the shrubs to burn successively and the fire to spread rapidly;

[0089] Similarly, if the distance between the trees in the area where the fire source is located is too close, the flame will be more likely to spread directly from one tree to another, expanding the scope of the fire in a short time. For example, the branches and leaves of adjacent trees are in contact with each other or are extremely close, and the flame can easily jump from the crown of one tree to another, forming a large-area crown fire that quickly sweeps across the entire area. Moreover, the close distance between the trees will hinder the normal circulation of air, forming a relatively enclosed space environment. The heat and smoke generated by the burning of the trees accumulate in a limited space. Once the air circulation condition is improved due to the burning, collapse, etc. of the trees, a large amount of fresh air rushes in, which will trigger more intense combustion, the fire intensity will increase instantly and spread more rapidly. At the same time, the effect of thermal radiation in the area where the trees are distributed too closely is more significant. Each burning tree radiates a large amount of heat to the surrounding trees, rapidly increasing the temperature of the adjacent trees and greatly increasing the possibility of reaching the ignition point. This chain reaction of thermal radiation will cause more trees to catch fire successively, further promoting the spread of the fire;

[0090] Therefore, the present invention determines the spread trend characterization value of the fire situation through the accumulation density of the vegetation and the average spacing between the trees in the area where the fire source is located to characterize the degree of influence of the distribution of the trees and vegetation in the area where the fire source is located on the spread of the fire situation, providing data support for predicting the spread coverage area of the fire situation in the future. The present invention can improve the efficiency of drone patrol, provide real-time feedback on forest fires, predict the spread trend of the fire situation, and improve the pertinence and effectiveness of emergency response.

[0091] Specifically, please refer to Figure 2As shown, it is a logic step diagram for predicting the spread coverage area of a fire in an embodiment of the present invention. The spread analysis module is used to predict the spread coverage area of the fire based on the spread trend characterization value in combination with wind direction data, including,

[0092] To obtain the wind direction in the area where the fire source is located;

[0093] To determine the spread trend characterization values corresponding to several areas along the wind direction starting from the area where the fire source is located;

[0094] If there exists any area whose corresponding spread trend characterization value is greater than or equal to the spread trend characterization threshold, then the area is predicted as the spread coverage area of the fire.

[0095] The spread trend characterization threshold is selected within the range [1.23, 1.46].

[0096] Specifically, the spread analysis module is used to determine the spread characterization parameter based on the terrain features of the spread coverage area, including,

[0097] To use the ratio of the valley depth to the valley depth threshold as the first spread characteristic;

[0098] To use the ratio of the slope perpendicularity to the slope perpendicularity threshold as the second spread characteristic;

[0099] To use the sum of the first spread characteristic and the second spread characteristic as the spread characterization parameter.

[0100] It can be understood that generally, valleys with a depth of more than ten meters are prone to form local microclimates, with poor air circulation, difficult diffusion of heat and smoke, easy accumulation of heat, and difficult dissipation of the high temperature generated by the fire, thus facilitating the spread of the fire. Therefore, in practice, to characterize the situation of being prone to form local microclimates, the valley depth threshold is set to 20m. Of course, those skilled in the art can adjust the valley depth threshold, which will not be elaborated here.

[0101] Moreover, on relatively steep slopes, such as slopes exceeding 25°, the spread speed of the fire will increase significantly. Especially in the presence of wind, the flame is more likely to rapidly expand upward along the slope to form a "rushing fire", and its spread speed is much faster than that in flat terrain areas. Therefore, in practice, the slope perpendicularity threshold is set to 30°;

[0102] Among them, the angle between the slope surface and the vertical horizontal plane is used as the slope perpendicularity.

[0103] Specifically, the present invention provides a diffusion analysis module that predicts the diffusion coverage area of a fire based on the tree planting distribution characteristics in combination with wind direction data. Vegetation and trees that can act as combustibles and affect the severity of a fire in a forest fire, their distribution determines the material basis for the spread of the fire. At the same time, combined with the wind direction, which is a factor that can directly affect the advancing direction of the fire, a comprehensive analysis is carried out to accurately predict the area that the fire may affect. Furthermore, the degree of spread of the fire is determined in combination with the terrain and topography of the corresponding area, including,

[0104] In the case where there are similar terrain and topography structures such as valleys in the area where the fire spreads, deeper such structures will form unique air flow channels, and the internal air flow changes may affect the propagation path of flying fire. The flying fire may be carried by the air flow in the valley to a farther place, triggering new ignition points, thereby expanding the spread range of the fire; for slope structures such as slopes, the higher the verticality of the slope, that is, the steeper the slope, the faster the fire spreads upward. At the same time, based on the characteristic of hot air rising, it also makes the flame easier to expand upward. The vegetation covering such structures may be more likely to collapse due to gravity, resulting in more complete combustion. Since the fire spreads upward rapidly, the vegetation above may be strongly thermally radiated in a short time, accelerating combustion, which makes the fire in the steep slope area more intense and the fire spread faster;

[0105] Therefore, the present invention determines the spread characterization parameter based on the terrain characteristics of the diffusion coverage area to characterize the influence of the internal terrain structure on the severity of the fire in the area affected by the fire, providing data support for determining the diffusion warning level subsequently. The present invention can improve the efficiency of UAV inspection, provide real-time feedback on forest fires, predict the fire spread trend, and improve the pertinence and effectiveness of emergency response.

[0106] Specifically, the diffusion analysis module is used to determine the diffusion warning level of the diffusion coverage area, including,

[0107] Pre-set the corresponding relationship between the diffusion warning level and the preset spread characterization parameter interval;

[0108] Determine the preset spread characterization parameter interval to which the spread characterization parameter belongs;

[0109] Set the diffusion warning level corresponding to the preset spread characterization parameter interval as the diffusion warning level of the diffusion coverage area;

[0110] Among them, the diffusion warning level and the preset spread characterization parameter interval are in one-to-one correspondence.

[0111] In this embodiment, the diffusion warning level is determined in the following manner:

[0112] The spreading characterization parameter is divided into three preset intervals, corresponding to three spreading warning levels respectively;

[0113] If the spreading characterization parameter is within the first preset interval [2.24, 2.37), the spreading warning level corresponding to the spreading coverage area is the first spreading warning level;

[0114] If the spreading characterization parameter is within the second preset interval [2.37, 2.53], the spreading warning level corresponding to the spreading coverage area is the second spreading warning level;

[0115] If the spreading characterization parameter is within the third preset interval (2.53, +∞], the spreading warning level corresponding to the spreading coverage area is the third spreading warning level.

[0116] The purpose of setting the warning level is to give an early warning when the spreading characterization parameter is relatively high. The spreading characterization parameter is calculated from the valley depth and the slope verticality. The standard normal value of the spreading characterization parameter is 2. Therefore, based on the deviation degree of the spreading characterization parameter from the standard normal value, the first preset interval, the second preset interval and the third preset interval are set. Each interval is continuous and represents an increasingly large deviation degree, so as to represent the deviation situation;

[0117] In implementation, the left interval endpoint of the first preset interval is between 1.1 times and 1.13 times of the standard normal value, and the right interval endpoint is set between 1.15 times and 1.2 times of the standard normal value.

[0118] The left interval endpoint of the second preset interval is continuous with the right interval endpoint of the first preset interval, and the right interval endpoint of the second preset interval is between 1.25 times and 1.3 times of the standard normal value;

[0119] The left interval endpoint of the third preset interval is continuous with the right interval endpoint of the second preset interval. Usually, the right interval endpoint of the third preset interval is set to positive infinity to cover all cases where the spreading characterization parameter has a large deviation.

[0120] Specifically, please refer to Figure 3 shown in the figure, which is the logical decision diagram for determining the feasible path in the embodiment of the present invention. The path matching module is used to determine the feasible path, including

[0121] used to determine a number of paths and determine the paths that meet the road condition benchmark conditions as feasible paths;

[0122] Among them, the road condition benchmark conditions include that the road surface width is greater than the path width threshold and the curve radius is greater than the curve radius threshold.

[0123] Specifically, for path recognition, the forest image data collected by the drone can be input into a pre-trained deep learning model. Through the deep learning model, pixel-level or point-level classification of the image or point cloud data is performed to segment the road from the forest background, so as to achieve the purpose of path recognition.

[0124] Specifically, the path width threshold can be determined according to the way of fire truck rescue. For example, in an application environment where fire trucks can drive smoothly in both directions, when multiple fire trucks need to operate simultaneously or leave enough space for passing other rescue vehicles, etc., the path width of the main road is generally not less than 6m. At the same time, for the branch road that satisfies the one-way driving of the fire truck and leaves a certain space to ensure that rescue personnel can carry out relevant operations beside the vehicle, the path width is generally not less than 4m, which will not be elaborated here;

[0125] The bend radius threshold can be determined according to the type of fire truck for rescue. For example, when the urban main battle fire truck is used for rescue, because the turning radius of this type of fire truck is relatively small, under the condition of allowing road conditions, a bend radius of generally 10 - 12m can meet the turning requirements of this type of fire truck; when the off-road fire truck is used for rescue, this type of fire truck is mainly used for fire fighting and rescue in complex terrains such as mountains and forests, and its bend radius is usually larger, generally about 15m. This is because the forest road conditions are poor, there are many obstacles and irregular bends, and a larger turning radius is required to ensure safe turning, which will not be elaborated here.

[0126] Specifically, please refer to Figure 4 As shown, it is the logical decision diagram for determining the available fire path in the embodiment of the present invention. The path matching module is used to determine the available fire path, including,

[0127] If the terrain steepness amplitude of any feasible path is less than the terrain steepness amplitude threshold, then the feasible path is determined as the available fire path.

[0128] It can be understood that generally, fire trucks used for special environment operations such as mountainous areas have higher climbing ability and can cope with steeper slopes, but generally should not exceed 30°. If it exceeds 30°, the fire truck may lose stability, cannot drive normally, and may even have accidents such as rollover or slipping. The terrain steepness amplitude threshold is set to 28°, which will not be elaborated here.

[0129] The present invention sets a path matching module. On the premise that the prediction of the diffusion coverage area of the fire is completed, the terrain steepness amplitude of the path at a certain distance from the corresponding area is considered to determine the available fire path that can provide a smooth passage for fire fighting equipment and sufficient space for fire fighting implementation.

[0130] Specifically, the predetermined distance is determined according to the range of the fire-fighting equipment, where the fire-fighting equipment includes large or special fire trucks.

[0131] It can be understood that for forest fires, fire-fighting operations should be carried out outside the safe distance to avoid harm to rescue equipment and rescue personnel. Therefore, in this embodiment, fire-fighting and rescue are carried out on the road at a predetermined distance from the diffusion coverage area according to the range of the fire-fighting equipment. For example, when the fire-fighting equipment used is equipped with long-range fire-extinguishing equipment, the predetermined distance can be set at about 50m - 90m, and the range of the equipment is used to extinguish the fire, which will not be elaborated here.

[0132] Specifically, the alarm and emergency module is used to issue an alarm signal of the corresponding level according to the diffusion alarm level determined by the diffusion analysis module, including

[0133] Pre-set the corresponding relationship between the alarm signal and the diffusion alarm level;

[0134] Among them, the alarm signal corresponds one-to-one to the diffusion alarm level, and the alarm signal includes the sounding of the alarm.

[0135] In this embodiment, the corresponding relationship between the alarm signal and the diffusion alarm level is determined in the following way:

[0136] If the diffusion alarm level is the first diffusion alarm level, then alarm signal 1 is issued;

[0137] If the diffusion alarm level is the second diffusion alarm level, then alarm signal 2 is issued;

[0138] If the diffusion alarm level is the third diffusion alarm level, then alarm signal 3 is issued;

[0139] Among them, alarm signal 1 includes that the alarm sounds for 30s every 10 minutes;

[0140] Alarm signal 2 includes that the alarm sounds for 30s every 5 minutes;

[0141] Alarm signal 3 includes that the alarm sounds for 30s every 2 minutes.

[0142] So far, the technical solution of the present invention has been described in combination with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A patrol drone fire emergency system based on deep learning, characterized in that: include: A fire situation collection module, which includes an image collection unit for acquiring forest image data and an environment collection unit for acquiring forest environment data; A fire source analysis module, which is connected to the fire situation collection module, is used to determine the location of the fire source, identify the tree distribution characteristics in the area where the fire source is located, and analyze the fire spread trend characterization value; a diffusion analysis module connected to the fire source analysis module, for predicting the diffusion coverage area of ​​the fire according to the diffusion trend characterization value combined with wind direction data, and determining the diffusion characterization parameters based on the terrain characteristics of the diffusion coverage area to determine the diffusion alarm level of the diffusion coverage area; A path matching module, which is connected to the fire situation collection module and the diffusion analysis module respectively, and is used to determine the terrain steepness corresponding to a number of feasible paths based on the predetermined distance of the diffusion coverage area, so as to determine the available fire fighting path; An alarm emergency module, connected to the diffusion analysis module, for issuing an alarm signal of a corresponding level according to the diffusion alarm level determined by the diffusion analysis module; The tree distribution characteristics include the density of vegetation in the area and the average spacing between trees, and the terrain characteristics include the depth of valleys and the verticality of slopes; The diffusion analysis module is used to predict the diffusion coverage area of ​​the fire according to the spreading trend characterization value combined with wind direction data, including: To obtain the wind direction in the area where the fire source is located; To determine the spreading trend characterization values ​​corresponding to several areas along the wind direction starting from the area where the fire source is located; If there is any area whose corresponding spreading trend characterization value is greater than or equal to the spreading trend characterization threshold, the area is predicted as the spreading coverage area of ​​the fire; The path matching module is used to determine that the feasible path includes: To determine several paths and determine the path that meets the road condition benchmark conditions as a feasible path; The road condition reference condition includes that the road width is greater than a path width threshold and the curve radius is greater than a curve radius threshold.

2. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The fire situation analysis module is used to determine the location of the fire source, including Using the forest image data collected by the image acquisition unit to determine smoke distribution and smoke concentration changes, so as to locate the direction of the fire source; It is used to extract the heat distribution along the direction of the fire source, and determine the distribution position corresponding to the highest heat as the fire source position.

3. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The fire situation analysis module is used to analyze the fire spreading trend representation value, including: The ratio of the accumulation density of vegetation in the area to the accumulation density threshold is used as the first spreading trend feature; The ratio of the average spacing threshold to the average spacing between trees in the area is used as the second spreading trend feature; The first spreading trend feature and the second spreading trend feature are weightedly summed to obtain the spreading trend representation value.

4. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The diffusion analysis module is used to determine the spreading and diffusion characterization parameters based on the terrain characteristics of the diffusion coverage area, including: The ratio of the valley depth to the valley depth threshold is used as the first spreading and diffusion feature; The ratio of the slope verticality to the slope verticality threshold is used as the second spreading and diffusion feature; The sum of the first spreading diffusion characteristic and the second spreading diffusion characteristic is used as the spreading diffusion characterization parameter.

5. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The diffusion analysis module is used to determine the diffusion alarm level of the diffusion coverage area, including: Preset the corresponding relationship between the diffusion alarm level and the preset diffusion characterization parameter range; Determining a preset spread diffusion characterization parameter interval to which the spread diffusion characterization parameter belongs; Setting the diffusion alarm level corresponding to the preset spreading diffusion characterization parameter interval as the diffusion alarm level of the diffusion coverage area; Among them, the diffusion alarm level corresponds one-to-one to the preset spread characterization parameter range.

6. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The path matching module is used to determine the available fire fighting paths, including: If there is any feasible path whose terrain steepness is less than the terrain steepness threshold, the feasible path is determined as the available fire fighting path.

7. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The predetermined distance is determined according to the range of the fire-fighting equipment.

8. The deep learning-based inspection drone fire emergency system according to claim 1 is characterized in that: The alarm emergency module is used to send an alarm signal of a corresponding level according to the diffusion alarm level determined by the diffusion analysis module, including: Pre-set the corresponding relationship between the alarm signal and the diffusion alarm level; The alarm signal corresponds to the diffusion alarm level one by one, and the alarm signal includes the sounding of an alarm.

Images