Intelligent collaborative control method and system for firefighting drones

Abstract

The present invention discloses an intelligent collaborative control method and system for a firefighting UAV, and belongs to the field of automatic control technology. The intelligent collaborative control method calibrates the operation formation in the operation center according to the number of UAVs, calibrates the UAVs to each row of the operation formation according to the state parameters, and calibrates the column of each UAV according to the effective working radius. The present invention reconstructs the operation environment model and the effective working radius according to the actual operation situation. If the fire point coordinate data is outside the effective working radius of the UAV, the collaborative UAV of the current UAV is searched, and then the operation formation is updated to replace the operation point coordinates of the current UAV and the collaborative UAV. Intelligent collaborative control is realized, the efficiency and flexibility of firefighting operations are improved, and the ability to cope with emergencies is enhanced. At the same time, the best collaborative UAV is found according to the working margin of the alternative UAV and the current UAV, reducing the difficulty of modifying the operation formation.

Description

Technical Field

[0001] The present invention relates to the field of automation control technology, and in particular to an intelligent collaborative control method and system for a fire-fighting unmanned aerial vehicle. Background Art

[0002] UAV formations can be used in special fields such as firefighting and disaster relief. Chinese Patent Publication No. CN116301039A discloses a vehicle-mounted UAV collaborative operation control system and method, which calculates the fire point position coordinates and azimuth data corresponding to each UAV based on the fire point detection data and differential positioning data, and calculates the navigation terminal position coordinates and azimuth of the UAV. In the parallel operation mode, each UAV plans the corresponding route according to the navigation terminal position coordinates and azimuth of the UAV and performs firefighting operations at the same time. This method discloses a general method for cooperative control of UAVs in parallel operation mode. However, affected by its own state and external operating environment, the initial coordinates, attitude angles and formation strategies may not be able to meet the continuous operation of the UAV. It is necessary to adjust the attitude and formation strategy in combination with the actual environmental conditions to achieve intelligent collaborative control of UAVs. Summary of the invention

[0003] In response to the above problems, the present invention provides an intelligent collaborative control method and system for fire-fighting UAVs, which determines the operating point coordinates and attitude angle of each UAV according to fire-fighting needs, and realizes intelligent collaborative control of the operating formation based on the real-time operating environment model and effective working radius, thereby enhancing the ability to respond to emergencies.

[0004] The invention objectives of this application can be achieved through the following technical means:

[0005] An intelligent collaborative control method for a firefighting drone comprises the following steps:

[0006] Step 1: Multiple drones collect continuous first images, generate fire point coordinate data and an operation center based on the first images, and calibrate an operation formation at the operation center based on the number of drones. The operation formation has multiple sets of operation point coordinates;

[0007] Step 2: Generate the height data of each row of the operation formation, calibrate the drone to each row of the operation formation according to the state parameters, collect the medium input pressure of each drone, and predict the medium output parameters;

[0008] Step 3: Predict the effective working radius of the drone based on the medium output parameters and the operating environment model, calibrate the column of each drone based on the effective working radius, and determine the corresponding operating point coordinates based on the rows and columns of the drone;

[0009] Step 4: The UAV moves to the corresponding operation point coordinates, generates a first attitude angle according to the operation point coordinates, and then generates a second attitude angle according to the medium output parameters;

[0010] Step 5: The drone collects continuous second images, extracts medium features of the second images, and reconstructs the operating environment model based on the medium features;

[0011] Step 6: Calculate the effective working radius of each UAV according to the medium output parameters and the operating environment model. If the fire point coordinate data is outside the effective working radius of the UAV, generate an internal formation instruction and proceed to step 7, otherwise proceed to step 8;

[0012] Step 7: Find the coordinated UAV of the current UAV, replace the operation point coordinates of the current UAV and the coordinated UAV, and return to step 4;

[0013] Step 8: If an external formation instruction is received, update the operation formation according to the external formation instruction and return to step 1; otherwise, update the operation point coordinates according to the first image and return to step 4.

[0014] In the present invention, in step 1, the fire point features of the first image are extracted, the fire point coordinate data are generated according to the fire point features, the working surface is established along the fire point coordinate data according to the minimum working distance, the working center is extracted from the working surface, and the working formation is calibrated along the working center according to the maximum working density.

[0015] In the present invention, in step 2, the state parameter is the remaining power of the UAV, and each UAV is calibrated to the row with the largest altitude data in order from large to small remaining power. After the current row with the largest altitude data is filled, the row with the largest altitude data is searched again until all rows of the operating formation are calibrated.

[0016] In the present invention, in step 2, the medium pressure loss P of the current UAV is calculated. 1 =ρgh, ρ is the medium density, g is the gravity acceleration, h is the current UAV height data, medium output parameters , P 2 Enter the pressure for the medium.

[0017] In the present invention, in step 3, the operating environment model includes mdv xt / dt=ma x -bv xt , m is the outflow mass of the medium per unit time, v xt is the instantaneous horizontal velocity at time t, a x is the horizontal wind force coefficient, b is the drag coefficient, and the horizontal trajectory function S is generated according to the operating environment model. x =f 1 (t), S x For horizontal displacement, the medium output parameter and the lower limit of the medium parameter are substituted into the horizontal trajectory function to obtain the effective working radius.

[0018] In the present invention, in step 4, a working vector of the working point coordinates and the fire point coordinate data is constructed, the heading angle of the working vector is the first attitude angle, the horizontal working radius and the vertical working radius are calculated according to the working vector, and the vertical trajectory function S is generated according to the working environment model. y =f 2 (t), S y For vertical displacement, the horizontal working radius and the vertical working radius are input into the horizontal trajectory function and the vertical trajectory function respectively to calculate the second attitude angle.

[0019] In the present invention, in step 5, multiple sets of coordinate observation data are extracted according to the medium characteristics, a trajectory error function is constructed, and the horizontal wind coefficient and drag coefficient of the working environment model are iteratively updated until the trajectory error function is less than the benchmark error, and the working environment model is reconstructed based on the updated wind coefficient and drag coefficient.

[0020] In the present invention, in step 7, the first working margin of each UAV replacing the current UAV is calculated, the alternative UAVs whose first working margin is greater than 1 are extracted, the second working margin of the current UAV replacing each alternative UAV is calculated, the distance between each alternative UAV and the current UAV is calculated, the alternative UAVs are sorted from small to large along the distance, and the alternative UAVs are extracted in sequence until the second working margin of the extracted alternative UAV is greater than 1, and the alternative UAV is marked as a collaborative UAV.

[0021] In the present invention, the first replacement radius R of each drone replacing the current drone is calculated. 1 , the first working margin = R 1 / L 1 , L 1 is the modulus of the working vector of the current drone; calculate the second replacement radius R of the current drone replacing each candidate drone 2 , the second working margin = R 2 / L 2 , L 2 is the modulus of the working vector of the candidate drone.

[0022] A system for implementing the intelligent collaborative control method of the firefighting drone comprises: a plurality of drones and a control center, wherein:

[0023] Drones include:

[0024] A first camera unit, used for collecting continuous first images;

[0025] A second camera unit, used for collecting continuous second images;

[0026] An operation control unit, used for generating a first posture angle and a second posture angle;

[0027] A first communication unit, used for communicating with a control center;

[0028] The control center includes:

[0029] A first processing unit is used to generate fire point coordinate data, operation center and operation point coordinates;

[0030] A second processing unit, for extracting medium characteristics and reconstructing an operating environment model;

[0031] Equipment monitoring unit, used to calculate the effective working radius and generate internal formation instructions;

[0032] Equipment adjustment unit, used to update the operation formation;

[0033] A second communication unit, used for communicating with the drone;

[0034] The third communication unit is used to receive external formation instructions.

[0035] The beneficial effects of the intelligent collaborative control method and system for firefighting drones implemented in the present invention are as follows: the present invention calibrates the working formation according to the fire point coordinate data, the drone moves to the corresponding working point coordinates of the working formation, and then determines the attitude angle of the drone, so that the fire extinguishing medium of each drone can be delivered to the fire point. The working environment model and the effective working radius are adjusted in real time according to the actual working situation, and then the working formation is updated to realize intelligent collaborative control, improve the efficiency and flexibility of firefighting operations, and enhance the ability to respond to emergencies. At the same time, the best collaborative drone is found based on the working margin of the alternative drones and the current drones, reducing the difficulty of modifying the working formation. Furthermore, the setting of external formation instructions can ensure manual intervention in the firefighting process, and can be adjusted according to actual conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of the intelligent collaborative control method of the fire-fighting drone of the present invention;

[0037] Figure 2 is a schematic diagram of a first image of the present invention;

[0038] Figure 3 Generate a schematic diagram of the working area for the present invention;

[0039] Figure 4 It is a schematic diagram of the coordinates of multiple working points in the working formation of the present invention;

[0040] Figure 5 Schematic diagram of the posture of the drone of the present invention;

[0041] Figure 6 A schematic diagram of a first attitude angle of multiple UAVs of the present invention;

[0042] Figure 7 A schematic diagram of a second attitude angle of multiple UAVs of the present invention;

[0043] Figure 8 is a schematic diagram of a second image of the present invention;

[0044] Fig. 9 The present invention is a block diagram of a system for implementing the intelligent collaborative control method of the fire-fighting drone. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Embodiment 1

[0046] like Figures 1 to 8 The intelligent collaborative control method of a fire-fighting UAV of the present invention determines the operating point coordinates and attitude angle of each UAV according to the fire point coordinate data, updates the operating formation according to the real-time operating environment model and the effective working radius, and improves the efficiency and flexibility of firefighting operations. The present invention first defines a reference coordinate system OXYZ, and the collaborative control of all UAVs is completed under this reference coordinate system. The intelligent collaborative control method of a fire-fighting UAV of the present invention specifically includes the following steps.

[0047] Step 1: Multiple drones collect continuous first images, generate fire point coordinate data and an operation center based on the first images, calibrate the operation formation at the operation center based on the number of drones, and the operation formation has multiple sets of operation point coordinates. Preset the template features of the fire point, use machine vision to extract the fire point features in the first image, and determine the fire point contour based on the fire point features. Figure 2 , the first image is located in the vertical plane, and the coordinate system is, for example, O'X'Z'. Generate the fire point coordinate data at the center of the fire point contour, and establish the working surface along the fire point coordinate data according to the minimum working distance, and the minimum working distance is, for example, 1 m to 3 m. Extract the working center from the working surface, and the working center is, for example, the point on the working surface closest to the fire point coordinate data, or the intersection of the ideal parabola passing through the fire point coordinate data and the working surface, such as Figure 3 As shown. The operation formation is calibrated along the operation center according to the maximum operation density. The operation point coordinates are arranged in sequence around the operation center according to the maximum operation density. The operation point coordinates form an operation formation, as shown in Figure 4 In a preferred embodiment, the work formation is located in a circular work area, and the number of work point coordinates in each row and column of the work formation is not equal.

[0048] Step 2: Generate height data for each row of the operation formation, calibrate the drone to each row of the operation formation according to the state parameter, collect the medium input pressure of each drone, and predict the medium output parameter. The state parameter is the remaining power of the drone. Each drone is calibrated to the row of maximum height data in the order of remaining power from large to small. The operation formation consists of multiple rows and columns of operation point coordinates. After the current row of maximum height data is filled, re-search the row with the largest height data until all rows of the operation formation are calibrated. The medium is, for example, water, and the medium output parameter is, for example, the outlet flow rate. The method for predicting the medium output parameter is described in Example 2.

[0049] Step 3: Predict the effective working radius of the drone based on the medium output parameters and the operating environment model, calibrate the column of each drone based on the effective working radius, and determine the corresponding operating point coordinates based on the rows and columns of the drone. The effective working radius is the working radius at which the horizontal component of the medium output parameter reaches the lower limit of the medium parameter. The preferred calculation method of the effective working radius refers to Example 2.

[0050] Step 4: The drone moves to the corresponding work point coordinates, generates a first attitude angle according to the work point coordinates, and then generates a second attitude angle according to the medium output parameters. The first attitude angle and the second attitude angle of the drone determine the landing point of the medium. The present invention calibrates the work formation according to the fire point coordinate data, optimizes the attitude angle of the drone, and can deliver the fire extinguishing medium to the fire point. In order to ensure that the medium accurately falls into the fire point coordinate data, the present invention can use the method described in Example 3 to adjust the first attitude angle and the second attitude angle of the drone.

[0051] Step 5: The drone collects a continuous second image, extracts the medium characteristics of the second image, and reconstructs the operating environment model based on the medium characteristics. The second image is an image of the medium before it falls into the fire point coordinate data. Under ideal conditions, the medium characteristics are parabolic. Affected by the actual horizontal wind force, air resistance, etc., the medium characteristics form a non-parabolic trajectory. The operating environment model can be corrected based on the observation data of the non-parabolic trajectory. As described in Example 3, the present invention can preset the horizontal wind force and air resistance in order to construct an initial operating environment model.

[0052] Step 6: Calculate the effective working radius of each drone based on the medium output parameters and the operating environment model. If the fire point coordinate data is outside the effective working radius of the drone, generate an internal formation command and go to step 7, otherwise go to step 8. If the fire point coordinate data is outside the effective working radius of the drone, adjusting the first attitude angle or the second attitude angle of the current drone cannot ensure that the medium falls into the fire point coordinate data. The current drone cannot operate effectively and needs to start collaborative control. The present invention adjusts the operating environment model and the effective working radius in real time according to the actual operating conditions, and then updates the operating formation.

[0053] Step 7: Find the cooperative drone of the current drone, replace the operation point coordinates of the current drone and the cooperative drone, and return to step 4. If the cooperative drone cannot be found, end the task. Since the current drone cannot maintain effective operation, the present invention selects a cooperative drone from other drones to replace the current drone. The present invention calculates the working margins of the alternative drones and the current drone, and finds the best cooperative drone according to the sorting algorithm, thereby reducing the difficulty of modifying the operation formation and avoiding the cooperative drones being too far away, resulting in the replacement of the operation point coordinates causing the medium pipeline to become tangled. The preferred sorting algorithm of the present invention is described in reference to Example 4.

[0054] Step 8: If an external formation instruction is received, update the operation formation according to the external formation instruction and return to step 1; otherwise, update the operation point coordinates according to the first image and return to step 4. Specifically, extract multiple sets of adjacent first images, use the image difference algorithm to update the fire point coordinate data and the operation center, and then calculate the coordinate adjustment of the current operation center relative to the operation center of the previous first image, and calculate the updated coordinates of each operation point according to the coordinate adjustment. The external formation instruction usually contains the target trajectory of any UAV, and the UAV moves to the corresponding coordinates according to the target trajectory. The setting of the external formation instruction can ensure manual intervention in the firefighting process and can be adjusted according to the actual situation. Embodiment 2

[0055] This embodiment further discloses a method for calibrating a drone to the coordinates of an operation point of an operation formation.

[0056] Step 201: calibrate the drone to each row of the operation formation according to the state parameter. The state parameter is the remaining power of the drone. Each drone is calibrated to the row with the largest altitude data in descending order of the remaining power. After the row with the largest altitude data is filled, the row with the largest altitude data is searched again until all rows of the operation formation are calibrated. The larger the state parameter, the higher the altitude and the greater the mounting capacity that the drone can maintain. The drone with a larger state parameter is calibrated to the row with the maximum altitude data, and the filling of the operation formation is completed in sequence.

[0057] Step 202: Predict the medium output parameter. The medium output parameter is the outlet flow rate of the medium pipeline. Collect the medium input pressure P of each drone. 1 , the current medium pressure loss P of the UAV 1 =ρgh, ρ is the medium density, g is the gravity acceleration, h is the current UAV height data, medium output parameters .

[0058] Step 301: Initialize the operating environment model. The operating environment model includes a horizontal resistance model: mdv xt / dt=ma x -bvxt and vertical resistance model: mdv yt / dt=mg-bv yt m is the mass of medium flowing out per unit time, for example, the mass of medium flowing out per second. xt is the horizontal instantaneous velocity at time t, v yt is the vertical instantaneous velocity at time t, a x is the horizontal wind force coefficient, b is the drag coefficient, and g is the gravity acceleration. When the angle between the horizontal wind force and the working vector is less than π / 2, a x When the horizontal wind force and the working vector are greater than π / 2, a x When the horizontal wind force and the working vector are equal to π / 2, a x is 0.

[0059] Step 302: Generate a horizontal trajectory function S according to the working environment model x =f 1 (t), S x is the horizontal displacement. Horizontal resistance model mdv xt / dt=ma x -bv xt The general solution is v xt =C×exp(-bt / m)+ma x / b, t = 0, v xt The horizontal component v of the medium output parameter x0 . Therefore the constant C = v 0 -ma x / b. Substituting into v xt =(v x0 -ma x / b)exp(-bt / m)+ma x / b, inverse solution to obtain the velocity attenuation function .

[0060] Output parameter v for medium xt Integrate to obtain the horizontal trajectory function . Further, along v x0 to v xt Definite integral on interval, .

[0061] Step 304: Substitute the medium output parameter and the medium parameter lower limit into the horizontal trajectory function to obtain the effective working radius. In order to complete the task, the minimum medium output parameter, i.e., the medium parameter lower limit v min , for example, 2.15 m / s. Calculate the horizontal instantaneous velocity v according to the velocity decay function xt Attenuation to the lower limit of medium parameter v min Time t 0 , . 0 Substituting into the horizontal trajectory function, we can get the effective working radius R 0 =f 1 (t 0 ).

[0062] Step 305: calibrate the column of each drone according to the effective working radius, and determine the coordinates of the corresponding operation point of the drone according to the row and column. In step 201, after each row of the operation formation is filled, the order of the drones in the same row is not determined. The smaller the effective working radius, the smaller the working range of the drone. Calibrate the drone with the smallest effective working radius to the column closest to the operation center. After the column closest to the current operation center is calibrated, select the next column closest to the operation center until all columns of the current row are filled, and then fill the next row. Embodiment 3

[0063] This embodiment further discloses a method for generating a first attitude angle and a second attitude angle of a drone according to an operating environment model and updating the operating environment model.

[0064] Step 401: construct a working vector of the working point coordinates and the fire point coordinate data, the heading angle of the working vector is the first attitude angle β, such as Figure 5 and Figure 6 The heading angle is the rotation angle of the working vector around the Z axis. The direction of the working vector is the same as the direction of the medium outlet parameter. The starting point of the working vector is the operating point coordinates, and the focus is on the fire point coordinate data.

[0065] Step 402: Calculate the horizontal working radius and vertical working radius according to the working vector. The projection of the working vector on the horizontal plane is the horizontal working radius L 3 The vertical height of the working vector is the vertical working radius L 4 .

[0066] Step 403: Generate a vertical trajectory function S according to the operating environment model y =f 2 (t) ,S y is the vertical displacement. yt Usually vertically downward, the resistance is vertically upward. The vertical resistance model is mdv yt / dt=mg-bv yt At t=0, the vertical instantaneous velocity v yt is the vertical component v of the medium output parameter y0 . Combined with the second embodiment, along v y0 to v yt The integrated vertical trajectory function m is the medium mass per unit time, v y0 is the vertical instantaneous velocity at time t, b is the drag coefficient, and g is the acceleration due to gravity.

[0067] Step 404: Input the horizontal working radius into the horizontal trajectory function, and input the vertical working radius into the vertical trajectory function to obtain Formula 1: ; Formula 2: , the vertical component v y0 = S inγv 0 With the horizontal component v x0 =cosγv 0 Substitute the above two equations to calculate the second posture angle γ, as Figure 7 .

[0068] Step 501: The drone collects continuous second images and extracts medium features of the second images. Figure 8 , extract the centerline trajectory of the medium in the second image, and transform the centerline trajectory into the coordinate system OXYZ by combining the coordinate conversion algorithm such as the depth map or stereo vision technology. The medium feature is the coordinate (x, y, z) of any point on the centerline trajectory.

[0069] Step 502: Extract multiple groups of coordinate observation data according to the medium characteristics and construct a trajectory error function. For example, extract N groups of point coordinates (x n ,y n ,z n ) as the coordinate observation data, 1≤n≤N. Calculate the coordinate observation data (x n ,y n ,z n ) horizontal displacement , substitute the horizontal displacement into the horizontal trajectory function S x =f 1 (t), calculate (x n ,y n ,z n ) corresponds to the time t n Then t n Substitute into the vertical trajectory function and calculate the height S yn . Trajectory error function .

[0070] Step 503: Iteratively update the horizontal wind force coefficient and drag coefficient of the working environment model until the trajectory error function is less than the reference error, which is, for example, 10 cm. 2 . Rebuild the operating environment model based on the updated wind force coefficient and drag coefficient. Embodiment 4

[0071] This embodiment further discloses a method for replacing the operating point coordinates of the current UAV and the cooperative UAV.

[0072] Step 701: Calculate the first replacement radius of the current drone replacing each drone. Use the method of Embodiment 1 to generate a speed attenuation function based on the operating environment model of the current drone. Combine the medium input pressure of other drones with the altitude data of the current drone to calculate the medium output parameters of other drones when they are at the altitude data of the current drone.

[0073] The medium output parameter and the medium parameter lower limit v min Substitute the speed attenuation function of the current drone Get the first decay time t', then substitute the first decay time t' into the horizontal trajectory function and vertical trajectory function of the current drone to get the horizontal distance and vertical distance, and then calculate the square sum to get the first replacement radius R 1 .

[0074] Step 702: Generate a first working margin = R 1 / L 1 , L 1 is the modulus of the working vector of the current UAV, and the candidate UAVs whose first working margin is greater than 1 are extracted. The first working margin being greater than 1 indicates that the candidate UAV can enter the working point coordinates of the current UAV and operate normally.

[0075] Step 703: Calculate the second replacement radius of the current drone replacing each candidate drone. Referring to step 701, substitute the medium input pressure of the current drone into the operating environment model of the candidate drone to generate the second replacement radius R 2 .

[0076] Step 704: Generate a second working margin = R 2 / L 2 , L 2 is the modulus of the working vector of the candidate drone. Referring to step 702, if the second working margin is greater than 1, it indicates that the current drone can enter the working point coordinates of the candidate drone and operate normally, thereby completing the selection of the drone. In a more preferred embodiment, in order to ensure the safety of the operation, the candidate drone with the first working margin or the first working margin greater than 1.8 to 2 is extracted.

[0077] Step 704: Search for cooperative drones in the candidate drones according to the sorting algorithm, and replace the coordinates of the operation points of the current drone and the cooperative drone. Preferably, the candidate drone with the smallest distance from the current drone is selected to minimize the impact of reordering on the overall cooperative work. The sorting algorithm is: calculate the distance between each candidate drone and the current drone, sort the candidate drones in order from small to large along the distance, and extract the candidate drones in order until the second working margin of the extracted candidate drone is greater than 1, and the candidate drone is marked as a cooperative drone. Embodiment 5

[0078] like Fig. 9 A system for implementing the intelligent cooperative control method of the fire-fighting UAV of the present invention comprises: a plurality of UAVs and a control center. A medium pipeline is fixed at the bottom of the UAV. The medium pipeline outputs the fire-fighting medium in a predetermined direction according to the posture of the UAV. The control center is located on the ground and sends a cooperative control signal to the control center.

[0079] The drone includes: a first camera unit, a second camera unit, an operation control unit, a first communication unit, and a first main control unit. The first camera unit is used to collect continuous first images, and the first camera unit is, for example, a thermal imager. The second camera unit is used to collect continuous second images, and the second camera unit is, for example, a binocular vision device that can collect a depth map of a firefighting medium. The operation control unit is used to generate a first attitude angle and a second attitude angle. The first communication unit is used to communicate with a control center. The first main control unit is used to coordinate the operation of devices such as a battery, an operation control unit, and a motor.

[0080] The control center includes: a first processing unit, a second processing unit, an equipment monitoring unit, an equipment adjustment unit, a second communication unit, a third communication unit and a second main control unit. The first processing unit is used to generate fire point coordinate data, operation center and operation point coordinates. The second processing unit is used to extract medium characteristics and reconstruct the operation environment model. The equipment monitoring unit is used to calculate the effective working radius and generate internal formation instructions. The equipment adjustment unit is used to update the operation formation. The second communication unit is used to communicate with the drone. The third communication unit is used to receive external formation instructions. The third communication unit can be connected to the external central control system to facilitate the user to actively intervene in the posture and trajectory of the drone.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent collaborative control method for a firefighting drone, characterized in that: The following steps are involved: Step 1: Multiple drones collect continuous first images, generate fire point coordinate data and an operation center based on the first images, and calibrate an operation formation at the operation center based on the number of drones. The operation formation has multiple sets of operation point coordinates; Step 2: Generate the height data of each row of the operation formation, calibrate the drone to each row of the operation formation according to the state parameters, collect the medium input pressure of each drone, and predict the medium output parameters; Step 3: Predict the effective working radius of the drone based on the medium output parameters and the operating environment model, calibrate the column of each drone based on the effective working radius, and determine the corresponding operating point coordinates based on the rows and columns of the drone. The operating environment model includes mdv xt / dt=ma x -bv xt , m is the outflow mass of the medium per unit time, v xt is the instantaneous horizontal velocity at time t, a x is the horizontal wind force coefficient, b is the drag coefficient, and the horizontal trajectory function S is generated according to the operating environment model. x =f1(t),S x For horizontal displacement, substitute the medium output parameter and the lower limit of the medium parameter into the horizontal trajectory function to obtain the effective working radius; Step 4: The UAV moves to the corresponding operation point coordinates, generates a first attitude angle according to the operation point coordinates, and then generates a second attitude angle according to the medium output parameters; Step 5: The drone collects continuous second images, extracts medium features of the second images, and reconstructs the operating environment model based on the medium features, wherein the centerline trajectory of the medium in the second image is extracted, and the medium features are the coordinates (x, y, z) of any point on the centerline trajectory. Multiple sets of coordinate observation data are extracted based on the medium features, and a trajectory error function is constructed. The horizontal wind coefficient and drag coefficient of the operating environment model are iteratively updated until the trajectory error function is less than the reference error, and the operating environment model is reconstructed based on the updated wind coefficient and drag coefficient. Step 6: Calculate the effective working radius of each UAV according to the medium output parameters and the operating environment model. If the fire point coordinate data is outside the effective working radius of the UAV, generate an internal formation instruction and proceed to step 7, otherwise proceed to step 8; Step 7: Find the coordinated UAV of the current UAV, replace the operation point coordinates of the current UAV and the coordinated UAV, and return to step 4; Step 8: If an external formation instruction is received, update the operation formation according to the external formation instruction and return to step 1; otherwise, update the operation point coordinates according to the first image and return to step 4.

2. The intelligent collaborative control method for firefighting drone according to claim 1 is characterized in that: In step 1, the fire point features of the first image are extracted, the fire point coordinate data are generated according to the fire point features, the working surface is established along the fire point coordinate data according to the minimum working distance, the working center is extracted from the working surface, and the working formation is calibrated along the working center according to the maximum working density.

3. The intelligent collaborative control method for firefighting drone according to claim 1, characterized in that: In step 2, the state parameter is the remaining power of the UAV. Each UAV is calibrated to the row with the largest altitude data in order from large to small remaining power. After the current row with the largest altitude data is filled, the row with the largest altitude data is searched again until all rows of the operation formation are calibrated.

4. The intelligent collaborative control method for firefighting drone according to claim 1, characterized in that: In step 2, the medium pressure loss P1=ρgh of the current drone is calculated, where ρ is the medium density, g is the gravity acceleration, h is the height data of the current drone, and the medium output parameter , P2 is the medium input pressure.

5. The intelligent collaborative control method for firefighting drone according to claim 1, characterized in that: In step 4, a working vector of the working point coordinates and the fire point coordinate data is constructed. The heading angle of the working vector is the first attitude angle. The horizontal working radius and the vertical working radius are calculated according to the working vector. The vertical trajectory function S is generated according to the working environment model. y =f2(t),S y For vertical displacement, the horizontal working radius and the vertical working radius are input into the horizontal trajectory function and the vertical trajectory function respectively to calculate the second attitude angle.

6. The intelligent collaborative control method for firefighting drone according to claim 1, characterized in that: In step 7, the first working margin of each UAV replacing the current UAV is calculated, and alternative UAVs whose first working margin is greater than 1 are extracted. The second working margin of the current UAV replacing each alternative UAV is calculated, and the distance between each alternative UAV and the current UAV is calculated. The alternative UAVs are sorted from small to large along the distance, and the alternative UAVs are extracted in sequence until the second working margin of the extracted alternative UAV is greater than 1. The alternative UAV is marked as a cooperative UAV.

7. The intelligent collaborative control method for firefighting drone according to claim 6, characterized in that: Calculate the first replacement radius R1 of each drone replacing the current drone, the first working margin = R1 / L1, L1 is the modulus length of the working vector of the current drone; Calculate the second replacement radius R2 of the current UAV replacing each candidate UAV, the second working margin = R2 / L2, L2 is the modulus length of the working vector of the candidate UAV.

8. A system for implementing the intelligent collaborative control method of the firefighting drone according to claim 1, characterized in that: include: Multiple drones and a control center, where Drones include: A first camera unit, used for collecting continuous first images; A second camera unit, used for collecting continuous second images; An operation control unit, used for generating a first posture angle and a second posture angle; A first communication unit, used for communicating with a control center; The control center includes: A first processing unit is used to generate fire point coordinate data, operation center and operation point coordinates; A second processing unit, for extracting medium characteristics and reconstructing an operating environment model; Equipment monitoring unit, used to calculate the effective working radius and generate internal formation instructions; Equipment adjustment unit, used to update the operation formation; A second communication unit, used for communicating with the drone; The third communication unit is used to receive external formation instructions.

Images