A mine operation repair method based on a unmanned aerial vehicle and related devices

Abstract

The application provides a mine operation repair method based on a unmanned aerial vehicle and related devices, and relates to the technical field of mine repair. Based on factors such as a terrain fluctuation factor, a wind field turbulence intensity factor, a unmanned aerial vehicle interference factor, a dropping angle, a wind direction angle and a dropping delay, a preliminary landing point of seed deposition is first given through a physical model, deviations caused by complex wind field turbulence, unmanned aerial vehicle interference and irregular terrain are corrected in combination with a machine learning algorithm, and nonlinear and interaction problems in an actual scene are considered, a quadratic term and a small amount of physical meaning interaction term are used for feature mapping, and the physical model is corrected in combination with the machine learning algorithm, so that accurate prediction of the seed deposition landing point can be better realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine restoration, in particular to a mine operation restoration method based on a UAV and related device. BACKGROUND

[0002] With the increasing awareness of global ecological environment protection, mine ecological restoration has become an important part of sustainable development. Traditional artificial vegetation restoration methods have low efficiency, high cost and high safety risks when facing complex mine topography, while UAV precision seeding technology provides a new solution for mine vegetation restoration. However, the complexity of the mine environment poses a serious challenge to the accuracy of UAV seeding.

[0003] Mine topography usually has complex features such as steep slopes, collapse pits and canyons, which can significantly change the local wind field distribution and form complex airflow phenomena such as turbulence and vortex. At the same time, environmental parameters such as temperature gradient, air pressure change and humidity distribution in the mine environment also affect the seed settling trajectory. Traditional UAV seeding systems mainly rely on simplified aerodynamic models, only considering the idealized balance relationship of gravity, buoyancy and air resistance, and are difficult to accurately predict the actual settling behavior of seeds in complex environments.

[0004] Traditional methods ignore key factors such as mine-specific topography, turbulence disturbance and UAV wake interference, and lack detailed modeling of different seed characteristics (density, shape, moisture content, etc.), resulting in large deviations between predicted results and actual conditions. Secondly, existing methods cannot effectively handle the disturbance of complex mine topography on the wind field, and the wind field modeling is too simplified, lacking real-time perception and adaptive adjustment capabilities for dynamic environmental changes. SUMMARY

[0005] In order to achieve accurate prediction of seed settling points during mine restoration, the present application provides a mine operation restoration method based on a UAV and related device.

[0006] In a first aspect, the present application provides a mine operation restoration method based on a UAV, which adopts the following technical solution:

[0007] A mine operation restoration method based on a UAV, comprising:

[0008] Determining the seed drop area based on the mine vegetation planting requirements, determining the take-off point, target drop point and flight trajectory of the UAV based on the received drop task;

[0009] Constructing a material aerodynamic model and correcting the settling velocity prediction based on a support vector regression algorithm with physical constraints;

[0010] Based on the terrain factor, the wind field turbulence intensity factor, the unmanned aerial vehicle interference factor, the release angle, the wind direction angle and the release delay factor, combined with the predicted settling velocity, the seed release settling point is predicted.

[0011] Based on the planning of the optimal flight path of the range of each region that can be released;

[0012] The step of predicting the seed release settling point based on the terrain factor, the wind field turbulence intensity factor, the unmanned aerial vehicle interference factor, the release angle, the wind direction angle and the release delay factor, combined with the predicted settling velocity, comprises:

[0013] The preliminary landing position is calculated using a physical model;

[0014] The deviation caused by the complex environment is intelligently corrected by using a Ridge regression algorithm;

[0015] Nonlinear and interaction effects are processed by introducing a quadratic term and a physical meaning interaction term feature mapping;

[0016] The final predicted landing point is obtained by combining the physical model prediction landing point and the machine learning correction amount.

[0017] Optionally, the step of constructing a material aerodynamic model and correcting the settling velocity prediction based on a support vector regression algorithm with physical constraints comprises:

[0018] A preliminary settling velocity estimate is obtained based on a traditional aerodynamic model;

[0019] A multi-dimensional feature vector containing seed characteristic parameters, environmental parameters and mine specific parameters is constructed;

[0020] The settling velocity prediction model is corrected using a support vector regression algorithm with physical constraints;

[0021] The seed characteristic parameters include seed density, equivalent diameter, sphericity and moisture content;

[0022] The environmental parameters include air density, aerodynamic viscosity, environmental temperature, atmospheric pressure, air specific humidity and environmental average wind speed;

[0023] The mine specific parameters include the terrain factor, the wind field turbulence intensity factor and the unmanned aerial vehicle interference factor.

[0024] Optionally, the terrain factor is calculated by the following method:

[0025] Based on the unmanned aerial vehicle laser radar point cloud data, the local terrain slope angle and curvature are calculated;

[0026] The terrain factor is calculated according to the slope angle and the curvature, and the calculation formula is: wherein θ(x,y) is the slope angle, κ(x,y) is the curvature, , is the empirical weight coefficient.

[0027] Optionally, the wind field turbulence intensity factor is obtained by:

[0028] obtaining a local wind speed time series by a UAV carrying a wind speed sensor;

[0029] simulating the wind field under complex terrain based on the CFD simulation results;

[0030] combining the field meteorological monitoring station data to calculate the turbulence intensity factor.

[0031] Optionally, the UAV interference factor is calculated by:

[0032] constructing a UAV interference factor based on the flight height, flight speed and blade parameters;

[0033] The calculation formula is: ;

[0034] wherein is the downwash speed of the propeller, is the drop height, and ε is a small constant to avoid the denominator being zero.

[0035] Optionally, the step of planning the optimal flight path based on the range of each region where the UAV can be dropped includes:

[0036] based on the drop angle θ and the direction angle γ of the UAV, determining the range where the UAV can achieve accurate dropping in each region;

[0037] According to the drop range of multiple regions, a path optimization algorithm is used to plan the optimal flight path of the UAV for planting and repairing;

[0038] real-time sensing of the dynamic environmental changes of the mine, and adaptive adjustment of the flight path and seeding strategy.

[0039] The second aspect of the present application provides a mine operation repair system based on a UAV, comprising:

[0040] a region determination module for determining a seed drop region according to the mine vegetation planting demand, determining the take-off point, target drop point and flight trajectory of the UAV based on the received drop task;

[0041] a correction module for constructing a material aerodynamics model and performing settling velocity prediction correction based on a support vector regression algorithm with physical constraints;

[0042] The prediction module is configured to predict the seed dropping point based on a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle and a dropping delay factor, in combination with a predicted dropping speed.

[0043] The output module is configured to plan an optimal flight path based on the range of each region.

[0044] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the processor executes the method described above when running computer instructions stored in the memory.

[0045] In a fourth aspect, the present application provides a computer readable storage medium comprising instructions, which, when running on a computer, cause the computer to execute the method described above.

[0046] In summary, the present application is based on a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle and a dropping delay factor. Firstly, a preliminary dropping point of the seed is given by a physical model. Then, the deviation caused by complex wind field turbulence, UAV interference and irregular terrain is corrected by a machine learning algorithm. Furthermore, the non-linear and interactive problems in the actual scene are considered. Finally, the quadratic term and a small number of physical meaning interactive terms are used for feature mapping. Through the correction of the physical model combined with the machine learning algorithm, the accurate prediction of the seed dropping point can be better achieved. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 is a computer device structure schematic diagram of a hardware running environment related to an embodiment scheme of the present application.

[0048] Figure 2 is a flowchart of a first embodiment of a mine operation repair method based on a UAV of the present application.

[0049] Figure 3 is a structure block diagram of a first embodiment of a mine operation repair system based on a UAV of the present application. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0051] Referring to Figure 1 , Figure 1 is a computer device structure schematic diagram of a hardware running environment related to an embodiment scheme of the present application.

[0052] As Figure 1As shown, the computer device can include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection communication between these components. The user interface 1003 can include a display, an input unit such as a keyboard, and the optional user interface 1003 can further include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a wireless fidelity (Wi-Fi) interface). The memory 1005 can be a high-speed random access memory (RAM), and can also be a stable non-volatile memory (NVM), such as a disk memory. The memory 1005 can also be an independent storage device from the aforementioned processor 1001.

[0053] Those skilled in the art can understand that Figure 1 The structure shown in the figure does not constitute a limitation on the computer device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0054] As Figure 1 As shown, the memory 1005 as a storage medium can include an operating system, a network communication module, a user interface module, and a UAV-based mine operation repair program.

[0055] In Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the present application can be arranged in the computer device, and the computer device calls the UAV-based mine operation repair program stored in the memory 1005 through the processor 1001, and executes the UAV-based mine operation repair method provided in the present application.

[0056] The present application provides a UAV-based mine operation repair method, which is described in detail with reference to Figure 2 , Figure 2 The flowchart of the first embodiment of the UAV-based mine operation repair method of the present application is shown.

[0057] In this embodiment, the UAV-based mine operation repair method includes the following steps:

[0058] Step S10: Determine the seed drop area according to the mine vegetation planting requirements, determine the take-off point, target drop point and flight trajectory of the unmanned aerial vehicle based on the received drop task.

[0059] It should be noted that the core technology route in this embodiment is:

[0060] 1. Task planning: determine the drop area, take-off point, target point and flight trajectory;

[0061] 2. Settling velocity prediction: establish an aerodynamic physical model, and use a physically constrained support vector regression algorithm (PR-SVG) to correct the settling velocity prediction;

[0062] 3. Landing point prediction: based on the terrain undulation factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, drop angle, wind direction angle and drop delay factor, and the settling velocity predicted in the previous stage, the seed drop settling landing point is predicted. Specifically, the physical model is used to calculate the preliminary landing point, and the Ridge regression is used to correct the deviation caused by the complex environment;

[0063] 4. Path optimization: plan the optimal flight path based on the dropable range of each area.

[0064] Step S20: Build a material aerodynamics model, and correct the settling velocity prediction based on a physically constrained support vector regression algorithm.

[0065] It should be noted that the step of building a material aerodynamics model and correcting the settling velocity prediction based on a physically constrained support vector regression algorithm includes: obtaining a preliminary settling velocity estimate based on a traditional aerodynamics model; constructing a multi-dimensional feature vector containing seed characteristic parameters, environmental parameters and mine specific parameters; using a physically constrained support vector regression algorithm to correct the settling velocity prediction model; wherein the seed characteristic parameters include seed density, equivalent diameter, sphericity and moisture content; the environmental parameters include air density, aerodynamic viscosity, environmental temperature, atmospheric pressure, air specific humidity and environmental average wind speed; the mine specific parameters include the terrain undulation factor, the wind field turbulence intensity factor and the unmanned aerial vehicle interference factor.

[0066] It should be noted that the preliminary settling velocity estimate is obtained based on the traditional aerodynamics model, which improves the accuracy of subsequent prediction.

[0067] When the seed falls in the air, the settling velocity is mainly determined by three forces: gravity, buoyancy and air resistance. When the terminal velocity (i.e. when the settling velocity is stable) is reached, the three forces satisfy the force balance:

[0068]

[0069] wherein, 、 、 Gravity, buoyancy, air resistance, respectively.

[0070] The gravity formula is:

[0071]

[0072] The buoyancy formula is:

[0073]

[0074] wherein, is the seed mass; is the seed volume; is the seed density; is the air density; is the gravity acceleration.

[0075] The air resistance formula is:

[0076]

[0077] wherein, is the drag coefficient, related to the seed shape and Reynolds number; is the seed projection area in the airflow direction; is the settling velocity.

[0078] Therefore, when reaching the terminal settling velocity , the force balance is:

[0079]

[0080] After sorting, the settling velocity is:

[0081]

[0082] Due to the complex nonlinear environment of the mine and the different seed characteristics, it is difficult for the physical model to accurately predict the settling velocity. By constructing a physically constrained support vector regression algorithm (PR-SVG), the settling velocity prediction model is corrected to better adapt to the actual situation and improve the accuracy of prediction.

[0083] The input feature vector is:

[0084]

[0085] wherein, is the seed-related parameter, which is the seed density, seed equivalent diameter, seed sphericity, and seed moisture content, respectively; are environmental parameters, which are air density, air dynamic viscosity, ambient temperature, atmospheric pressure, air specific humidity, and ambient average wind speed, respectively; are mine-specific parameters, which are terrain relief factor, wind field turbulence intensity factor, and UAV interference factor, respectively.

[0086] Terrain relief factor is calculated from local slope and curvature, which characterizes the influence of complex mine terrain on wind field disturbance. Mines usually have complex terrain (steep slopes, collapse pits, etc.), which will change the distribution of local wind field. Wind will accelerate on slopes and form vortexes in depressions, thus affecting the seed deposition trajectory. Therefore, first, based on the UAV LiDAR point cloud, the local terrain slope angle is calculated:

[0087]

[0088] wherein, is the terrain height function. From this, the terrain relief factor can be defined:

[0089]

[0090] wherein, is the slope angle, the steeper the slope, the greater the wind speed disturbance; is the curvature, a positive value indicates convex, a negative value indicates concave, reflecting the wind flow dispersion / aggregation situation; , is the empirical weight, which can be calibrated through wind tunnel experiments or numerical simulation.

[0091] In mine canyons or slope areas, air flow is often accompanied by turbulence, which will cause fluctuations in the actual seed deposition velocity. The turbulence intensity factor measures the proportion of local wind speed fluctuation relative to the average wind speed, reflecting the instability of air. The turbulence intensity is defined as:

[0092]

[0093] wherein, is the instantaneous wind speed at the i-th time point; is the average wind speed; is the number of sampling points. In practical applications, data sources include local wind speed time series obtained by UAV-mounted wind speed sensors, CFD simulation results (wind field simulation under complex terrain), and field meteorological monitoring stations.

[0094] The UAV's own propeller produces a wake effect, which will change the airflow environment when the seed is just released, affecting the initial conditions of deposition. The UAV interference factor is constructed considering flight height, flight speed, and propeller parameters:

[0095]

[0096] wherein is the propeller downwash speed, is the release height, and ε is a small constant to avoid denominator zero, which can be modeled empirically:

[0097]

[0098] wherein, is the total propeller thrust; is the propeller radius, is the rotor disc area; is the air density.

[0099] Based on the above feature vector, the correction factor is predicted using SVR:

[0100]

[0101] wherein, is the correction factor; , is the Lagrange multiplier; is the kernel function; is the bias term.

[0102] The final settling velocity is obtained:

[0103] .

[0104] Step S30: Based on the terrain factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, release angle, wind direction angle and release delay factor, combined with the predicted settling velocity, the seed release settling point is predicted.

[0105] In specific implementation, based on the terrain factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, release angle, wind direction angle and release delay factor, combined with the settling velocity predicted in the last stage, the seed release settling point is predicted. (This part gives the preliminary landing point through the physical model, and combines the machine learning algorithm to correct the deviation caused by complex wind field turbulence, unmanned aerial vehicle interference, irregular terrain, etc.

[0106] The seed settling trajectory is determined by the particle kinematics formula, and is jointly affected by the terrain factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, release angle and release delay factor. Therefore, after considering the wind speed field, terrain influence, unmanned aerial vehicle interference and other factors, the basic physical model prediction formula is:

[0107]

[0108] wherein, is the seed landing position predicted by the physical model, is its calculation formula; for predicting landing time; for three-dimensional wind speed field, are three wind speed components of the wind speed field; for UAV release angle; for the direction angle of the wind, i.e., the angle between the seed release direction and the relative incoming flow direction; for the release delay, i.e., the time delay between the UAV executing the release command and the actual release of the seed; for the remaining available features predicted by the physical model, which together calculate the preliminary landing position of the seed.

[0109] Based on the landing position predicted by the physical model , the terrain undulation factor, the wind field turbulence intensity factor, the UAV interference factor, the release angle and the release delay factor are incorporated into the machine learning correction prediction using a regularized linear model (Ridge regression).

[0110] The basic input for the Ridge regression model prediction is:

[0111]

[0112] wherein, is the release angle, which is positive relative to the horizontal downward direction; is the direction angle of the wind.

[0113] In order to take into account certain nonlinearity and interaction, quadratic terms + a small number of physically meaningful interaction terms are used for feature mapping:

[0114]

[0115] wherein, 1 is the bias; controls the factors of wind and direction, and through this form, the size and direction information of the wind is encoded into the features, so that the model can not only perceive how big the wind is, but also perceive which way the wind blows; the terrain undulation factor and the direction angle of the wind are combined to represent how the terrain enhances or weakens the effect of the wind along the wind direction; represents the coupling effect of the downdraft and the release angle, if the release angle is aligned with the downdraft direction, the downdraft effect is the largest, if the angle is different, the downdraft effect is weakened.

[0116] It should be noted that the wind field turbulence intensity factor is obtained by: obtaining the local wind speed time series by the UAV carrying a wind speed sensor; simulating the wind field under complex terrain based on the CFD simulation results; combining the data of the field meteorological monitoring station to calculate the turbulence intensity factor.

[0117] In the extended feature space, the Ridge regression predicts the correction as:

[0118]

[0119] where, is the extended feature vector (contains the linear term, quadratic term and part of the physically meaningful interaction term); is the regression weight matrix.

[0120] For the i-th experiment, the physically model predicted landing point is , and the actual observed landing point is .

[0121] The correction label is:

[0122]

[0123] Meanwhile, each sample corresponds to a set of input features , thus forming the matrix Collecting the true correction vectors of all N samples as the supervision signal for Ridge regression.

[0124] To avoid overfitting, introduce the L2 regularization term, the objective function is:

[0125]

[0126] where, > 0 is the regularization coefficient; is the Frobenius norm, which squares all elements and sums them up to get the total sum of L2 weight penalty.

[0127] Ridge regression has an analytical solution, the weight matrix can be written as:

[0128]

[0129] where, ; ; is the identity matrix, which is used to realize the regularization of Ridge regression.

[0130] Combined with the physical landing point, the final predicted landing point is:

[0131] .

[0132] It should be noted that the step of predicting the seed drop point based on the terrain factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, drop angle, wind direction angle and drop delay factor, combined with the predicted settling velocity, includes: using a physical model to calculate a preliminary drop point position; using a Ridge regression algorithm to intelligently correct the deviation caused by complex environment; processing nonlinearity and interaction effect by introducing a quadratic term and a physical meaning interaction term feature mapping; combining the physical model to predict the drop point and the machine learning correction amount to obtain the final predicted drop point.

[0133] Step S40: planning an optimal flight path based on the range of each region that can be planted.

[0134] It can be understood that the step of planning an optimal flight path based on the range of each region that can be planted includes: judging the range of each region that can be precisely planted by the unmanned aerial vehicle based on the drop angle θ and the direction angle γ of the unmanned aerial vehicle; planning an optimal flight path for the unmanned aerial vehicle to plant and repair according to the planting range of multiple regions by using a path optimization algorithm; and adaptively adjusting the flight path and the seeding strategy in real time according to the dynamic environmental changes of the mine.

[0135] It should be noted that the above-mentioned means are finally realized in the embodiment.

[0136] 1. The aerodynamic model output is used as the input and constraint of SVR to avoid the prediction result from deviating. Meanwhile, the specific features of the mine (terrain relief, turbulence intensity, and unmanned aerial vehicle wake interference) are introduced to improve the adaptability of the model to complex environment, so that the prediction model is upgraded from "pure machine learning" to a prediction method of "physical constraint + scene enhancement + adaptive weighting", which has obvious innovation.

[0137] 2. In the drop and landing prediction part, the present application is based on the terrain factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, drop angle, wind direction angle and drop delay, etc. Firstly, a preliminary drop point of the seed is given by a physical model, combined with a machine learning algorithm to correct the deviation caused by complex wind field turbulence, unmanned aerial vehicle interference, irregular terrain, etc. And considering the nonlinearity and interaction in the actual scene, a quadratic term + a small number of physical meaning interaction terms are used for feature mapping. Through the physical model combined with the machine learning algorithm correction, the accurate prediction of the seed drop point can be better realized.

[0138] The embodiment is based on factors such as a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle and a dropping delay, and first gives a preliminary landing point of seed deposition through a physical model, corrects deviations caused by complex wind field turbulence, UAV interference and irregular terrain by combining a machine learning algorithm, considers nonlinear and interaction problems in an actual scene, adopts a quadratic term and a small number of physical meaning interaction terms for feature mapping, and can better realize accurate prediction of a seed deposition landing point by correcting through a physical model combined with a machine learning algorithm.

[0139] In addition, the embodiment of the application further provides a computer readable storage medium, and the storage medium stores a program for mine operation repair based on a UAV. The program for mine operation repair based on a UAV is executed by a processor to realize the steps of the method for mine operation repair based on a UAV as described above.

[0140] Reference Figure 3 , Figure 3 is a structural block diagram of a first embodiment of a mine operation repair system based on a UAV of the application.

[0141] As Figure 3 shown, the mine operation repair system based on a UAV provided by the embodiment of the application comprises:

[0142] A region determination module 10 is configured to determine a seed dropping region according to mine vegetation planting requirements, and determine a takeoff point, a target dropping point and a flight trajectory of a UAV based on a received dropping task.

[0143] A correction module 20 is configured to construct a material aerodynamics model, and correct a deposition speed prediction based on a support vector regression algorithm of physical constraints.

[0144] A prediction module 30 is configured to realize prediction of a seed dropping deposition landing point based on a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle and a dropping delay factor, and in combination with a predicted deposition speed.

[0145] An output module 40 is configured to plan an optimal flight path based on a droppable range of each region.

[0146] It should be understood that the above is only for illustration, and does not constitute any limitation on the technical solutions of the application. In specific applications, a person skilled in the art can set it up as needed, and the application does not limit this.

[0147] The embodiment is based on factors such as a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle, and a dropping delay. A preliminary landing point of the seeds is given through a physical model, deviations caused by complex wind field turbulence, UAV interference, and irregular terrain are corrected by combining a machine learning algorithm, nonlinear and interaction problems in an actual scene are considered, quadratic terms and a small number of physical interaction terms are used for feature mapping, and the physical model is corrected by combining the machine learning algorithm, so that the precision of the seed landing point is better predicted.

[0148] It should be noted that the above-described workflow is merely illustrative and does not limit the protection scope of the present application. In actual applications, a person skilled in the art can select part or all of the above-described workflow to achieve the purpose of the embodiment according to actual needs, and the present application is not limited in this regard.

[0149] In addition, technical details not described in detail in the embodiment can be found in the method for repairing mine operations based on UAVs provided by any embodiment of the present application, and will not be described here.

[0150] In addition, it should be noted that in this document, the terms "comprise", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or system. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of another identical element in the process, method, article or system that includes the element.

[0151] The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0152] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, also can be through hardware, but in many cases the former is the better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part of the contribution to the prior art can be embodied in the form of software products, the computer software product is stored in a storage medium (such as read-only memory (ReadOnly Memory, ROM) / RAM, disk, optical disc), including a number of instructions to make a terminal device (may be a mobile phone, computer, server, or network equipment, etc.) executes the method of each embodiment of the present application. The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, any equivalent structure or equivalent process transformation made by using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for repairing a mine operation based on a drone, characterized by, The method comprises the following steps: determining the seed drop area according to the mine vegetation planting requirements, determining the take-off point, target drop point and flight trajectory of the unmanned aerial vehicle based on the received drop task; constructing a material aerodynamic model, and correcting the settling velocity prediction based on a support vector regression algorithm with physical constraints; based on the terrain undulation factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, drop angle, wind direction angle and drop delay factor, combined with the predicted settling velocity, the settling drop point of the seed is predicted, and the physical model prediction formula is: in, The seed landing location predicted by the physical model. Its calculation formula; This refers to the estimated landing time. For a three-dimensional wind speed field, These are the three wind speed components of the wind speed field; The angle from which the drone is deployed; The wind direction angle is the angle between the direction of seed release and the direction of the relative incoming flow. The release delay is the time delay between the drone executing the release command and the actual release of the seeds. The remaining available features predicted by the physical model work together to calculate the initial landing position of the seed. These are specific parameters for the mine, namely, topographic relief factor, wind field turbulence intensity factor, and UAV interference factor; On the basis of the physical model predicted drop point , the terrain relief factor, wind field turbulence intensity factor, UAV interference factor, drop angle and drop delay factor are included in the machine learning correction prediction by using linear model with regularization ; the basic input of the Ridge regression model prediction is: wherein, is the angle of release, positive relative to horizontal downward; is the direction angle of the wind; planning the optimal flight path based on the range of each region where the seed can be dropped; wherein the step of predicting the settling drop point of the seed based on the terrain undulation factor, wind field turbulence intensity factor, unmanned aerial vehicle interference factor, drop angle, wind direction angle and drop delay factor, combined with the predicted settling velocity, comprises: calculating the preliminary drop point position using a physical model; using a Ridge regression algorithm to intelligently correct the deviation caused by complex environments; processing nonlinearity and interaction effects by introducing quadratic terms and physical meaning interaction terms; combining the physical model prediction drop point and the machine learning correction amount to obtain the final prediction drop point.

2. The method of claim 1, wherein, The step of constructing a material aerodynamic model and correcting the settling velocity prediction based on a support vector regression algorithm with physical constraints comprises: obtaining a preliminary settling velocity estimate based on a traditional aerodynamic model; constructing a multi-dimensional feature vector containing seed characteristic parameters, environmental parameters and mine specific parameters; correcting the settling velocity prediction model using a support vector regression algorithm with physical constraints; wherein the seed characteristic parameters include seed density, equivalent diameter, sphericity and moisture content; the environmental parameters include air density, aerodynamic viscosity, environmental temperature, atmospheric pressure, air relative humidity and environmental average wind speed; the mine specific parameters include the terrain undulation factor, the wind field turbulence intensity factor and the unmanned aerial vehicle interference factor.

3. The method of claim 2, wherein, The terrain undulation factor is calculated by: calculating the local terrain slope angle and curvature based on the unmanned aerial vehicle laser radar point cloud data; The terrain fluctuation factor is calculated according to the slope angle and the curvature, and the calculation formula is: where θ(x, y) is the slope angle, κ(x, y) is the curvature, , is an empirical weight coefficient.

4. The method of claim 2, wherein, The wind field turbulence intensity factor is obtained by: obtaining the local wind speed time series by the unmanned aerial vehicle carrying a wind speed sensor; simulating the wind field under complex terrain based on CFD simulation results; combining the on-site meteorological monitoring station data to calculate the turbulence intensity factor.

5. The method of claim 1, wherein, The unmanned aerial vehicle interference factor is calculated by: constructing the unmanned aerial vehicle interference factor based on the flight height, flight speed and blade parameters; The calculation formula is: ; wherein is the propeller wash speed, is the drop height, and ε is a small constant to avoid division by zero.

6. The method of claim 1, wherein, The step of planning the optimal flight path based on the range of each region where the seed can be dropped comprises: judging the range of each region where the unmanned aerial vehicle can achieve accurate drop based on the drop angle θ and the direction angle γ of the unmanned aerial vehicle; planning the optimal flight path of the unmanned aerial vehicle planting and repairing according to the drop range of multiple regions by using a path optimization algorithm; real-time sensing of mine dynamic environment changes, and self-adaptive adjustment of flight path and seeding strategy.

7. An unmanned aerial vehicle based mine operation remediation system, characterized by, The method of claim 1 comprises: a region determination module for determining the seed drop area according to the mine vegetation planting requirements, and determining the take-off point, target drop point and flight trajectory of the unmanned aerial vehicle based on the received drop task; The correction module is configured to build a material aerodynamic model and correct the settling velocity prediction based on a support vector regression algorithm with physical constraints; The prediction module is configured to predict the seed dropping settling point based on a terrain undulation factor, a wind field turbulence intensity factor, a UAV interference factor, a dropping angle, a wind direction angle and a dropping delay factor, and in combination with the predicted settling velocity. The prediction module is further configured to calculate a preliminary dropping point position using a physical model. The ridge regression algorithm is adopted to intelligently correct the deviation caused by complex environment. Nonlinear and interaction effects are processed through the introduction of quadratic terms and physical meaning interaction terms for feature mapping. The final predicted dropping point is obtained by combining the physical model prediction dropping point and the machine learning correction amount. The output module is configured to plan an optimal flight path based on the range of each region where the seeds can be dropped.

8. A computer device, comprising: The device comprises a memory and a processor, and the processor executes the method according to any one of claims 1 to 6 when running computer instructions stored in the memory.

9. A computer-readable storage medium, characterized in that, The instructions, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 6.

Images