Non-navigation route planning method and system based on fire rescue virtual reality scene

Abstract

The application provides a non-navigation route planning method and system based on a fire rescue virtual reality scene, comprising the following steps: receiving environment data of a forest fire area in real time through a Beidou navigation signal transceiver carried on a UAV, and calculating fire spread data according to the environment data by a computing node configured on the UAV; constructing a predicted virtual environment of the forest fire area through a VR imaging device carried on the UAV based on the fire spread data; dividing a survey point working area of the forest fire area based on the predicted virtual environment of the forest fire area, and planning a non-navigation route of the UAV according to real-time coordinates of the UAV and the divided survey point working area; and selecting an optimal route according to a distribution weight algorithm. The application fully considers the influence of environmental factors, improves the accuracy of virtual environment modeling, increases the objectivity of interest point selection, makes the planned non-navigation route of the UAV more scientific and reasonable, and greatly improves the survey effect of the UAV.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of unmanned aerial vehicle exploration, and particularly relates to a non-navigation route planning method and system based on a virtual reality scene for fire rescue. BACKGROUND

[0002] The existing forest fire exploration methods mainly include ground exploration and air exploration. The ground exploration mainly monitors through sensors arranged on the ground, and the monitoring range is relatively small. A large number of sensors are needed to cover a large area of forest, and the sensors are easily affected by adverse environments such as high temperature, lightning, strong wind, and heavy rain, which may cause sensor failure or false alarm. The air exploration has relatively high flexibility, is not limited by terrain, can penetrate into complex terrain areas, and can adjust the flight route and height at any time according to the needs, quickly reach the designated area for detailed exploration, and obtain high-resolution image and video data to facilitate accurate judgment of the fire and the fire source position. Unmanned aerial vehicles have strong practical value in forest fire air exploration, but the endurance time of most unmanned aerial vehicles is only within 1 hour, the flight time is relatively short, and the work efficiency is low, which makes it difficult to play a role in complex and time-consuming tasks. Therefore, the path planning of the unmanned aerial vehicle must be accurate, and the key exploration points must be effectively explored in a short time. However, the existing unmanned aerial vehicle air exploration path planning technology is relatively single, and the path cannot be dynamically adjusted according to the environmental conditions and flight conditions, which makes it difficult to achieve the expected effect of unmanned aerial vehicle forest fire air exploration. SUMMARY

[0003] The present application provides a non-navigation route planning method and system based on a virtual reality scene for fire rescue to overcome the deficiencies in the prior art.

[0004] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0005] In a first aspect, the present application provides a non-navigation route planning method based on a virtual reality scene for fire rescue. In the fire rescue, an unmanned aerial vehicle is used to explore a forest fire area, which includes the following steps:

[0006] Step 1: Real-time receiving the environmental data of the forest fire area through the Beidou navigation signal transceiver mounted on the unmanned aerial vehicle. The computing node configured on the unmanned aerial vehicle calculates the fire spread data according to the environmental data;

[0007] Step 2: Based on the fire spread data, a predicted virtual environment of the forest fire area is constructed through the VR imaging device mounted on the unmanned aerial vehicle, and is visualized on the ground;

[0008] Step 3: Based on the predicted virtual environment of the forest fire area, divide the forest fire area into exploration point working areas, and plan the non-navigation route of the UAV according to the real-time coordinates of the UAV and the divided exploration point working areas.

[0009] Step 4: For the planned multiple unnavigation routes for drones, select the optimal route based on the distribution weight algorithm.

[0010] Optionally, in step 1, the environmental data includes wind speed, air temperature, humidity, combustible material moisture content, and digital elevation model data of the ground; the fire spread data includes fire spread rate, fire spread slope and aspect of the forest surface.

[0011] Optionally, the fire spread rate is calculated by establishing the following multiple linear regression model:

[0012] v = β0 + β1v wind +β2p+ε;

[0013] In the formula, v represents the fire spread rate, β0, β1, and β2 are model coefficients, and β2 increases with the fire spread rate. wind Let f represent wind speed, p represent the combined influencing factors consisting of air temperature T, humidity H and water content of combustibles M, p = f(T, H, M), f represent an empirical function, and ε represent empirical error.

[0014] The slope and aspect of the forest surface where fire spreads are calculated based on digital elevation model data.

[0015] Optionally, in step 2, the process of constructing a predictive virtual environment for forest fire areas specifically involves:

[0016] Based on digital elevation model data of forest fire areas, 3D modeling software is used to create the original virtual environment of forest fire areas when no fire occurred.

[0017] By dividing the time period and rendering the original virtual environment based on the current fire spread data, a predicted virtual environment of the forest fire area for the next time period is obtained.

[0018] Optionally, during the process of constructing a predictive virtual environment for forest fire areas, the current fire spread data is updated at the end of each time period, and the original virtual environment is re-rendered to obtain the predictive virtual environment for forest fire areas in the next time period.

[0019] Optionally, in step 3, the planning process for the unmanned aerial vehicle's non-navigation route is specifically as follows:

[0020] Step 3.1: based on the predicted virtual environment of the forest fire area, the forest fire area is divided into exploration point work areas, and after the division, the interest points are marked on the exploration point work areas;

[0021] Step 3.2: obtain the real-time 3D coordinates of the unmanned aerial vehicle and map them into the predicted virtual environment of the forest fire area as the starting point;

[0022] Step 3.3: according to the starting point and the marked interest points, a plurality of unmanned aerial vehicle non-navigation routes are generated, so that each unmanned aerial vehicle non-navigation route passes through the starting point and the marked interest points.

[0023] Optionally, in step 3.1, the forest fire area is divided into exploration point work areas, specifically:

[0024] Step 3.1.1: uniformly divide the predicted virtual environment of the forest fire area into 3D grids;

[0025] Step 3.1.2: the divided 3D grids are clustered as follows using a feature clustering algorithm:

[0026] extract the terrain feature vector of each 3D grid and the vegetation feature vector k represents the 3D grid number, m represents the terrain feature, n represents the vegetation feature;

[0027] The adjacent 3D grids with the similarity of terrain features and vegetation features within the set threshold are clustered, and the similarity of terrain features and vegetation features is calculated as follows:

[0028]

[0029] In the formula, d1(x ki , x li ) and d2(y kj , y lj ) represent the Minkowski distance between the terrain features and the vegetation features of the kth 3D grid and the lth 3D grid, respectively, which represents the similarity of the terrain features and the vegetation features, and p represents the set similarity measure parameter; if d1

[0030] Step 3.1.3: the 3D grid area corresponding to each class after clustering represents each exploration point work area.

[0031] Optionally, in step 4, the process of selecting the optimal route according to the distribution weight algorithm is specifically:

[0032] According to the planned plurality of unmanned aerial vehicle non-navigation route, the network topology graph is drawn, the node of the network topology graph represents the starting point and the marked point of interest, and the edge of the network topology graph represents the flight path between the nodes;

[0033] The theoretical passing time of each edge in the network topology graph is calculated by the following formula:

[0034]

[0035] In the formula, t s [j] and L[j] are the theoretical passing time and path distance of the jth edge in the unmanned aerial vehicle non-navigation route, v UAV is the preset flight speed of the unmanned aerial vehicle;

[0036] According to the theoretical passing time of each edge, the initial weight value of each edge is assigned, and the theoretical passing time and the weight are proportional;

[0037] According to the weight of each edge in the unmanned aerial vehicle non-navigation route, the current optimal route is obtained, so that the sum of the weights of each edge on the optimal route is minimum.

[0038] Optionally, for the flight path of the unmanned aerial vehicle in the network topology graph, the weight of the edge is updated in the following manner:

[0039]

[0040] In the formula, w next [j] represents the updated weight of the jth edge, T UAV [j] represents the actual passing time of the unmanned aerial vehicle through the jth edge, w now [j] represents the current weight of the jth edge, P fa represents the false alarm probability of the unmanned aerial vehicle flying at the preset flight speed on the jth edge next time;

[0041] According to the updated weight of each edge, the current optimal route is reobtained.

[0042] In a second aspect, the present application provides a non-navigation route planning system based on a virtual reality scene of fire rescue, which uses an unmanned aerial vehicle to explore a forest fire area, comprising:

[0043] The Beidou navigation signal transceiver carried on the unmanned aerial vehicle is used to receive environmental data of the forest fire area in real time;

[0044] The computing node configured on the unmanned aerial vehicle is used to calculate fire spread data according to the environmental data;

[0045] The VR imaging equipment carried on the unmanned aerial vehicle is used for constructing a predicted virtual environment of a forest fire area based on fire spread data and visualizing on the ground;

[0046] The ground computing center divides the forest fire area into exploration point work areas based on the predicted virtual environment of the forest fire area, plans a non-navigation route of the unmanned aerial vehicle by calculating the distance between the real-time positioning coordinates of the unmanned aerial vehicle and the divided exploration point work areas, and selects an optimal route according to a link distribution weight algorithm.

[0047] The present application has the following advantages:

[0048] (1) The present application uses the unmanned aerial vehicle to explore the forest fire area, effectively uses the real-time received environmental data to calculate the fire spread data, fully considers the influence of environmental factors, can be adjusted in time according to the change of the environment, and improves the accuracy of the fire information acquisition;

[0049] (2) The present application introduces the fire spread data in the VR imaging process, and updates the virtual environment in time periods, on the one hand, fully considers the fire factor in the virtual environment modeling process, improves the accuracy of modeling, and on the other hand, avoids the influence of the change of the fire on the virtual environment, and can update the virtual environment in time;

[0050] (3) The present application considers the topographic features and vegetation features of the forest fire area, combines the improved clustering algorithm to divide the exploration point work areas, the divided exploration point work areas have topographic feature similarity and vegetation feature similarity, effectively increase the objectivity of the interest point selection, make the planned non-navigation route of the unmanned aerial vehicle more scientific and reasonable, and avoid the repeated flight of the unmanned aerial vehicle in similar areas and the mistaken entry into the dangerous area;

[0051] (4) The present application combines the directed network topology graph and the distribution weight algorithm to plan the optimal route of the non-navigation route of the unmanned aerial vehicle, skillfully designs the dynamic update of the distribution weight, fully considers the historical flight data of the unmanned aerial vehicle and the false alarm probability of the normal flight of the unmanned aerial vehicle in the update process, makes the selection of the optimal route consider the actual flight situation and the predicted flight situation of the unmanned aerial vehicle, and can greatly improve the exploration effect of the unmanned aerial vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 It is a flow chart of a non-navigation route planning method based on a fire rescue virtual reality scene.

[0053] Figure 2 It is an architecture diagram of a non-navigation route planning system based on a fire rescue virtual reality scene.

[0054] Figure 3 It is a planning process flow chart of a non-navigation route of an unmanned aerial vehicle. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be clearly and completely described in connection with the drawings in the embodiments of the present application.

[0056] Embodiment one

[0057] The embodiment proposes a non-navigation route planning method based on a fire rescue virtual reality scene, and the specific process is as shown in the figure. Figure 1 In the fire rescue, a UAV is used to explore the forest fire area, and the airborne equipment includes a Beidou navigation signal transceiver, a VR imaging device, various types of sensors, etc. The method includes the following steps:

[0058] Step 1: Real-time receive the environmental data of the forest fire area through the Beidou navigation signal transceiver carried on the UAV, and the computing node configured on the UAV calculates the fire spread data according to the environmental data.

[0059] The environmental data mainly includes wind speed, air temperature, humidity, water content of combustible materials, and digital elevation model (DEM) data of the ground, wherein the DEM data can be directly obtained from the ground base station, and is generally stored in the local terrain database for subsequent calculation of the fire spread slope and slope direction of the forest ground; other data can be collected through various types of sensors. The fire spread data mainly includes the fire spread speed, the fire spread slope and slope direction of the forest ground, which are used to simulate the geographic information of the forest ground fire spread.

[0060] The traditional fire spread speed calculation method is relatively simple, assuming that the fire spread distance between two time points T1 and T2 is d1, and the time interval is t1; the fire spread distance between time points T2 and T3 is d2, and the time interval is t2. Then the fire spread speeds v1 and v2 can be calculated by the following formula:

[0061]

[0062] This single linear calculation method does not consider the influence of environmental factors at all, and the error is large.

[0063] In this embodiment, in view of the influence of wind speed, air temperature, humidity, combustible materials and other factors, the fire spread speed is calculated by establishing a multiple linear regression model as follows:

[0064] v = β0 + β1v + β2p + ε wind

[0065] In the formula, v represents the fire spread speed, β0, β1 and β2 are model coefficients, and β2 increases with the increase of the fire spread speed, v wind ​represents the wind speed, p represents a comprehensive influence factor composed of air temperature T, humidity H and combustible moisture content M, p = f (T, H, M), f represents an empirical function, and ε represents an empirical error; the model coefficients, the empirical function, and the empirical error can be obtained according to historical experience data simulation.

[0066] Step 2: Based on the fire spread data, a predicted virtual environment of the forest fire area is constructed by a VR imaging device carried on the unmanned aerial vehicle, and ground imaging visualization is performed.

[0067] Firstly, based on the DEM data of the forest fire area, a 3D modeling software is used to create an original virtual environment of the forest fire area without fire; then, time periods are divided, and the original virtual environment is rendered according to the current fire spread data to obtain a predicted virtual environment of the forest fire area in the next time period.

[0068] In the process of constructing the predicted virtual environment of the forest fire area, the current fire spread data is updated at the end of each time period, and the original virtual environment is re-rendered to obtain a predicted virtual environment of the forest fire area in the next time period. In this way, the virtual environment applied in each time period is ensured to consider the current fire situation, the effectiveness of the model data is ensured, the influence of fire change on the virtual environment is avoided, and the virtual environment can be updated in a timely manner.

[0069] In terms of ground imaging visualization, firstly, the visualization content and the visualization UI of a device in the ground working area are designed. According to the visualization design, a virtual visualization model can be constructed in Unity, and finally, the indoor positioning technology and AR are used to realize the mobile monitoring function of the device.

[0070] Step 3: Based on the predicted virtual environment of the forest fire area, the forest fire area is divided into exploration point working areas, and the non-navigation route of the unmanned aerial vehicle is planned according to the real-time coordinates of the unmanned aerial vehicle and the divided exploration point working areas.

[0071] In this embodiment, as shown in Figure 3 , the planning process of the non-navigation route of the unmanned aerial vehicle is specifically as follows:

[0072] Step 3.1: Based on the predicted virtual environment of the forest fire area, the forest fire area is divided into exploration point working areas, and interest points are marked on the obtained exploration point working areas after the division.

[0073] In step 3.1, the forest fire area is divided into exploration point working areas, which is specifically as follows:

[0074] First, the prediction virtual environment of the forest fire area is uniformly divided into 3D grid. The prior art often directly uses the divided 3D grid as the exploration point, which is too large in division error without considering the terrain and vegetation factors, and the selected interest point is too random, resulting in unreasonable planning route and increasing the difficulty of exploration. This embodiment is based on the uniform division of 3D grid, and the divided 3D grid is clustered as follows by using a feature clustering algorithm:

[0075] Extracting the terrain feature vector of each 3D grid And the vegetation feature vector k represents the 3D grid number, m represents the terrain feature, n represents the vegetation feature;

[0076] The adjacent 3D grids with the similarity of terrain features and vegetation features within the set threshold are clustered, and the similarity calculation formula of terrain features and vegetation features is as follows:

[0077]

[0078] In the formula, d1(x ki , x li ) and d2(y kj , y lj ) represent the Minkowski distance between the kth 3D grid and the ith 3D grid in the terrain feature and vegetation feature, which is used to represent the similarity of the terrain feature and the vegetation feature, and p represents the set similarity measurement parameter, which can be set according to the calculation requirement; if d1< L1 and d2< L2, L2 and L2 are both set thresholds, then the kth 3D grid and the ith 3D grid are clustered, that is, only the 3D grids with similar terrain conditions and vegetation conditions are clustered and fused, and the 3D grid area corresponding to each cluster after clustering represents the working area of each exploration point. Compared with the uniform division method, the working area of the exploration point obtained by clustering is obviously more reasonable.

[0079] Step 3.2: Obtain the real-time 3D coordinates of the unmanned aerial vehicle and map them into the prediction virtual environment of the forest fire area as the starting point.

[0080] Step 3.3: According to the starting point and the labeled interest point, generate several unmanned aerial vehicle non-navigation routes, so that each unmanned aerial vehicle non-navigation route passes through the starting point and the labeled interest point. The unmanned aerial vehicle non-navigation route refers to the path of the unmanned aerial vehicle in the flight process, which does not rely on the pre-set navigation point or navigation path planning, but relies on the real-time environmental information, task demand or human control for flight.

[0081] Step 4: For the planned multiple unmanned aerial vehicle non-navigation routes, select the optimal route according to the distribution weight algorithm.

[0082] First, a directed network topology graph is drawn according to the planned multiple unmanned aerial vehicle non-navigation routes, wherein the nodes represent the starting points and the marked points of interest, and the edges of the network topology graph represent the flight paths between the nodes.

[0083] Then, the theoretical passing time of each edge in the network topology graph is calculated by the following formula:

[0084]

[0085] In the formula, t s [j] and L[j] are the theoretical passing time and path distance of the jth edge in the unmanned aerial vehicle non-navigation route, v UAV is the preset flight speed of the unmanned aerial vehicle.

[0086] The initial weight value of each edge is assigned according to the theoretical passing time of each edge, and the theoretical passing time is directly proportional to the weight. Then, the current optimal route is obtained according to the weights of the edges in the unmanned aerial vehicle non-navigation route, so that the sum of the weights of the edges on the optimal route is the smallest.

[0087] However, in the flight process of the unmanned aerial vehicle, it is difficult to ensure that the flight is always carried out at the preset flight speed due to the influence of the environment and the state of the device itself. In actual work, there may also be path changes, U-turns, and the like. Therefore, for the flight paths that the unmanned aerial vehicle has experienced in the network topology graph, the weight of the edge is updated in the following manner:

[0088]

[0089] In the formula, w next [j] represents the updated weight of the jth edge, t UAV [j] represents the actual passing time of the jth edge experienced by the unmanned aerial vehicle, w now [j] represents the current weight of the jth edge, P fa represents the false alarm probability of the unmanned aerial vehicle flying on the jth edge at the preset flight speed next time, and the false alarm probability can be evaluated according to real-time environmental information and the state of the unmanned aerial vehicle device. The evaluation model is constructed by constructing historical experience data.

[0090] According to the updated weight of each edge, the current optimal route is reacquired, and the unmanned aerial vehicle flies according to the updated optimal route in the next time. In actual operation, the timing of weight updating can be selected, such as when the environment fluctuates greatly, when the flight state of the unmanned aerial vehicle changes greatly, and when other conditions obviously require the flight route of the unmanned aerial vehicle to be changed. The embodiment fully considers the historical flight data of the unmanned aerial vehicle and the false alarm probability of the normal flight of the unmanned aerial vehicle in the updating process, so that the selection of the optimal route takes into account the actual flight situation and the predicted flight situation of the unmanned aerial vehicle, which can greatly improve the exploration effect of the unmanned aerial vehicle.

[0091] Embodiment two

[0092] The embodiment proposes a non-navigation route planning system based on a fire rescue virtual reality scene, which is used to implement the non-navigation route planning method based on the fire rescue virtual reality scene of embodiment one, and the basic architecture is as shown in Figure 2 The unmanned aerial vehicle is used to explore the forest fire area, which includes:

[0093] The Beidou navigation signal transceiver carried on the unmanned aerial vehicle is used to receive the environmental data of the forest fire area in real time;

[0094] The computing node configured on the unmanned aerial vehicle is used to calculate the fire spread data according to the environmental data;

[0095] The VR imaging device carried on the unmanned aerial vehicle is used to construct a predicted virtual environment of the forest fire area based on the fire spread data, and to visualize the ground imaging;

[0096] The ground computing center divides the forest fire area into exploration point working areas based on the predicted virtual environment of the forest fire area, plans a non-navigation route for the unmanned aerial vehicle by calculating the distance between the real-time positioning coordinates of the unmanned aerial vehicle and the divided exploration point working areas, and selects the optimal route according to the link distribution weight algorithm.

[0097] The specific functions and work flows of each device in the system are the same as those of each step of the non-navigation route planning method based on the fire rescue virtual reality scene of embodiment one, and therefore will not be described here.

[0098] Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized by electronic hardware or a combination of electronic hardware and computer software. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0099] The above are only preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical scheme falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary skilled in the art, some improvements and refinements without departing from the principles of the present application shall be considered as the protection scope of the present application.

Claims

1. A non-navigation route planning method based on a virtual reality scenario for fire rescue, wherein unmanned aerial vehicles (UAVs) are used to explore forest fire areas during fire rescue, characterized in that... Includes the following steps: Step 1: Receive real-time environmental data of the forest fire area through the Beidou navigation signal transceiver mounted on the drone; the computing nodes configured on the drone calculate the fire spread data based on the environmental data. Step 2: Based on fire spread data, construct a predicted virtual environment of the forest fire area using VR imaging equipment mounted on a drone, and visualize it on the ground. Step 3: Based on the predicted virtual environment of the forest fire area, the exploration point working area of ​​the forest fire area is divided. According to the real-time coordinates of the UAV and the divided exploration point working area, the non-navigation route of the UAV is planned. The specific process of planning the non-navigation route of the UAV in Step 3 is as follows: Step 3.1: Based on the predicted virtual environment of the forest fire area, divide the forest fire area into exploration point working areas, and mark points of interest on the obtained exploration point working areas after division. Step 3.2: Obtain the real-time 3D coordinates of the drone and map them onto the predicted virtual environment of the forest fire area as the starting point; Step 3.3: Based on the starting point and the marked points of interest, generate several unnavigated routes for the UAV, such that each unnavigated route passes through the starting point and the marked points of interest; Step 4: For the planned multiple unnavigation routes for drones, select the optimal route based on the distribution weight algorithm.

2. The non-navigation route planning method based on a virtual reality fire rescue scenario as described in claim 1, characterized in that: In step 1, the environmental data includes wind speed, air temperature, humidity, moisture content of combustibles, and digital elevation model data of the ground; the fire spread data includes fire spread rate, fire spread slope and aspect of the forest surface.

3. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 2, characterized in that: The fire spread rate was calculated using the following multiple linear regression model: v=β0+β1v wind +β2P+ε; In the formula, v represents the fire spread rate, β0, β1, and β2 are model coefficients, and β2 increases with the fire spread rate. wind Let f represent wind speed, and let P represent the comprehensive influencing factors composed of air temperature T, humidity H, and combustible material moisture content M. P = f(T,H,M), where f represents an empirical function and ε represents empirical error. The slope and aspect of the forest surface where fire spreads are calculated based on digital elevation model data.

4. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 1, characterized in that: In step 2, the process of constructing a predictive virtual environment for forest fire areas specifically involves: Based on digital elevation model data of forest fire areas, 3D modeling software is used to create the original virtual environment of forest fire areas when no fire occurred. By dividing the time period and rendering the original virtual environment based on the current fire spread data, a predicted virtual environment of the forest fire area for the next time period is obtained.

5. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 1, characterized in that: In the process of constructing a predictive virtual environment for forest fire areas, the current fire spread data is updated at the end of each time period, and the original virtual environment is re-rendered to obtain the predictive virtual environment for forest fire areas in the next time period.

6. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 1, characterized in that: In step 3.1, the division of the exploration point working area in the forest fire area specifically involves: Step 3.1.1: Divide the predicted virtual environment of the forest fire area into uniform 3D meshes; Step 3.1.2: The segmented 3D mesh is clustered using a feature clustering algorithm as follows: Extract terrain feature vectors from each 3D mesh. and vegetation feature vector k represents the 3D mesh number. Represent m terrain features, Represent n vegetation features; Adjacent 3D grids whose similarity to both terrain and vegetation features is within a set threshold are clustered. The similarity calculation formula for terrain and vegetation features is as follows: In the formula, and Let d1 and d2 represent the Minkowski distances of terrain features and vegetation features between the k-th and l-th 3D grids, respectively, to represent the similarity between terrain features and vegetation features. p represents the set similarity metric parameter. If d1 < L1 and d2 < L2, and L2 and L2 are both set thresholds, then the k-th and l-th 3D grids are clustered. Step 3.1.3: The 3D grid regions corresponding to each category after clustering represent the working areas of each exploration point.

7. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 1, characterized in that: In step 4, the process of selecting the optimal route based on the distribution weight algorithm is specifically as follows: A directed network topology graph is drawn based on the planned multiple unmanned aerial vehicle (UAV) non-navigation routes. In the network topology graph, nodes represent starting points and marked points of interest, and edges represent flight paths between nodes. The theoretical travel time for each edge in the network topology graph is calculated using the following formula: In the formula, t s [j] and L[j] represent the theoretical travel time and path distance of the j-th edge in the unnavigated route of the UAV, respectively. UAV The preset flight speed for the drone; Each edge is initially weighted based on its theoretical travel time, with the theoretical travel time being directly proportional to the weight. Based on the weights of each edge on the unnavigated route of the drone, the current optimal route is obtained, which minimizes the sum of the weights of each edge on the optimal route.

8. The non-navigation route planning method based on a virtual reality scenario for fire rescue as described in claim 7, characterized in that: For the flight paths already traversed by the drone in the network topology graph, the edge weights are updated in the following way: In the formula, w next [j] represents the updated weight of the j-th edge, t UAV [j] represents the actual travel time of the drone traversing the j-th edge, w now [j] represents the current weight of the j-th edge, P fa This represents the false alarm probability that the drone will fly at the preset speed on the j-th edge. Based on the updated weights of each edge, the current optimal route is obtained again.

9. A non-navigation route planning system based on a virtual reality scenario for fire rescue, used to implement the non-navigation route planning method as described in any one of claims 1-8, employing unmanned aerial vehicles (UAVs) to explore forest fire areas, characterized in that, include: The Beidou navigation signal transceiver mounted on the drone is used to receive environmental data from forest fire areas in real time; The computing nodes configured on the drone are used to calculate fire spread data based on environmental data; VR imaging equipment mounted on drones is used to construct a predictive virtual environment of forest fire areas based on fire spread data and visualize it on the ground. The ground computing center, based on the predicted virtual environment of forest fire areas, divides the forest fire areas into exploration point working areas. By calculating the distance between the real-time positioning coordinates of UAVs and the divided exploration point working areas, it plans the non-navigation routes of UAVs and selects the optimal route according to the link distribution weight algorithm.

Images