Control method of inspection unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicle flight control, in particular to a control method of an inspection unmanned aerial vehicle, which comprises the following steps of: in response to a preset inspection instruction, obtaining a current target area wind speed, terrain information and a plurality of humidity values fed back by a plurality of humidity sensors arranged at different heights; flight parameters are constructed based on the wind speed and terrain information of the target area, heat distribution cloud charts at different heights are generated during the period, and the drip irrigation amount in the target area is calculated based on the heat distribution cloud charts. According to the control method of the inspection unmanned aerial vehicle, the flight path can be dynamically optimized through a wind speed self-adaptive flight height adjustment mechanism in combination with single effective width calculation of the thermal imager, and meanwhile, the ground heat distribution data of the same target area are collected at different heights in later data processing, so that the control precision of the inspection unmanned aerial vehicle is improved. And meanwhile, in combination with the environment humidity corresponding to different heights, the interference of the flight height and the environment humidity on the thermal imaging data is reduced, and the environment data acquisition precision is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) flight control, and in particular to a control method for an inspection UAV. Background Art

[0002] Currently, drone-based inspection technology has been widely used in agricultural monitoring, facility inspection and other fields, but there are still significant deficiencies in data collection accuracy and intelligent decision-making capabilities in complex environments. Traditional drone inspection methods mostly rely on preset flight paths and do not fully consider the impact of dynamic environmental factors (such as wind speed, terrain, humidity, etc.) on flight parameters, resulting in limited thermal imaging coverage, data redundancy or loss, etc. Secondly, existing technologies have weak capabilities for fusing and processing multi-source data, making it difficult to achieve high-precision environmental modeling. For example, thermal imaging data is easily affected by flight altitude and ambient humidity, and heat distribution maps collected at a single altitude cannot accurately reflect the surface temperature gradient. Therefore, there is an urgent need for a drone inspection control method that can adapt to environmental changes and deeply integrate multi-dimensional data. Summary of the Invention

[0003] The purpose of the present invention is to provide a control method for an inspection drone to improve the problem that the above-mentioned existing technology has weak fusion processing capabilities for multi-source data and is difficult to achieve high-precision environmental modeling.

[0004] In order to achieve the above objectives, the embodiments of the present application provide the following technical solutions:

[0005] On the one hand, an embodiment of the present application provides a control method for an inspection drone, the method comprising: in response to a preset inspection instruction, obtaining the current target area wind speed, terrain information and multiple humidity values fed back by multiple humidity sensors located at different heights, the terrain information including ground three-dimensional coordinate information and ultra-high-risk equipment three-dimensional coordinate information; obtaining the corresponding flight altitude threshold range in a preset operation and maintenance parameter comparison table based on the target area wind speed, selecting multiple different flight altitudes, and calculating a single effective thermal imaging width based on the flight altitude and the lens parameters of the thermal imager; constructing a flight path based on the single effective thermal imaging width, the ground three-dimensional coordinate information and the ultra-high-risk equipment three-dimensional coordinate information; performing drone flight operations based on the flight path and flight altitude, during which a heat distribution cloud map in the target area is constructed by a thermal imager located at the bottom of the drone, and then obtaining heat distribution cloud maps corresponding to different flight altitudes; constructing a temperature distribution map based on multiple heat distribution cloud maps and multiple humidity values, and calculating the drip irrigation amount in the target area based on the current temperature distribution map, wherein the temperature distribution map represents the high-precision low-altitude vegetation surface temperature distribution.

[0006] Optionally, constructing a temperature distribution map based on the plurality of heat distribution cloud maps and the plurality of humidity values includes:

[0007] Get the humidity values at four different heights and calculate the humidity at any height based on the four height-humidity values:

[0008] H(h)=a(h-h1) 3 +b(h-h1) 2 +c(h-h1)+H1;

[0009] Where h is any height, h1 is the minimum height, H1 is the humidity value corresponding to the minimum height h1; a, b, c are coefficients, and the solution process is:

[0010]

[0011] Where Δh i =h i -h1, i=2, 3, 4; H i is the height h i Corresponding humidity value;

[0012] A humidity distribution cloud map is constructed based on the humidity value corresponding to any height and the three-dimensional coordinate information of the ground, and a temperature distribution map is constructed based on the humidity cloud map, the first heat distribution cloud map and the second heat distribution cloud map.

[0013] Optionally, constructing a temperature distribution map based on the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map includes:

[0014] The first convolutional neural network branch extracts the deep features of the first heat distribution cloud map, the second convolutional neural network branch extracts the deep features of the second heat distribution cloud map; the third convolutional neural network branch extracts the shallow features of the humidity cloud map;

[0015] The vertical thermal gradient corresponding to each cell is constructed based on the height difference between the first heat distribution cloud map and the second heat distribution cloud map:

[0016] G final =αG local +(1-α)G global ;

[0017] Among them, α is the weight coefficient, G local is the local vertical thermal gradient of 3*3 pixels; G global is the global vertical thermal gradient of 8*8, G final To integrate the vertical thermal gradient;

[0018] G = (E1-E2) / Δz; E1 represents the heat value of the area corresponding to the first heat distribution cloud map, E2 represents the heat value of the area corresponding to the second heat distribution cloud map, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, G is the vertical thermal gradient, G is Glocal or G global ;

[0019] The deep features of the first heat distribution cloud map and the deep features of the second heat distribution cloud map are channel-joined, and a channel weight map is generated through the attention module. The calculation process of the weight map injects the shallow features of the humidity cloud map as a bias term to generate the fusion feature corresponding to the target height:

[0020]

[0021] Among them, F f is the fusion feature, W is the channel weight map, F1 is the depth feature of the first heat distribution cloud map, F2 is the depth feature of the second heat distribution cloud map, G fnnal To incorporate vertical thermal gradients, represents the channel weighting, represents concatenation, ⊙ represents element-wise multiplication;

[0022] The fused features are upsampled and reconstructed through the physical constraint decoder, and then the temperature distribution map corresponding to the target height is output.

[0023] Optionally, before extracting features from the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map, the method further includes:

[0024] The humidity cloud map, the first heat distribution cloud map and the second heat distribution cloud map are processed in a grid space, and the thermal value in each cell is normalized, and the humidity value in each cell is Z-Score standardized.

[0025] Optionally, the first convolutional neural network branch and the second convolutional neural network branch each include 3 convolutional layers, and the third convolutional neural network branch includes 2 convolutional layers.

[0026] Optionally, upsampling and reconstructing the fused features through a physical constraint decoder to output a temperature distribution map corresponding to the target height includes:

[0027] The physical constraint decoder contains 5 layers of transposed convolution and introduces a differentiable heat conduction equation constraint term in the loss function:

[0028]

[0029] in, is the physical constraint number loss term, which is used to constrain the prediction results of the model through thermodynamic equations; λ is the constraint strength coefficient; is the heat diffusion term, is the vertical convection term, βM is the humidity correction term, α and β are learnable parameters, M is the humidity distribution value, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, ΔT is the temperature difference between the first heat distribution cloud map and the second heat distribution cloud map in the same corresponding area, and H is the predicted heat value corresponding to the target height.

[0030] Optionally, calculating the drip irrigation amount in the target area based on the current temperature distribution map includes:

[0031] The temperature distribution map is divided according to the preset minimum drip irrigation control unit, and the total heat value corresponding to each minimum drip irrigation control unit is calculated. Based on the total heat value, the corresponding drip irrigation amount is found in the drip irrigation operation and maintenance parameter comparison table, and the drip irrigation operation is implemented based on the corresponding drip irrigation amount.

[0032] In a second aspect, this embodiment provides a control system for an inspection drone, the system comprising:

[0033] an acquisition module, configured to respond to a preset inspection instruction and obtain the current wind speed and terrain information of the target area, and multiple humidity values fed back by multiple humidity sensors located at different heights, wherein the terrain information includes three-dimensional ground coordinate information and three-dimensional coordinate information of ultra-high-risk equipment;

[0034] a parameter calling module, configured to obtain a corresponding flight altitude threshold range from a preset operation and maintenance parameter comparison table based on the wind speed in the target area, select a plurality of different flight altitudes, and calculate a single effective thermal imaging width based on the flight altitude and the lens parameters of the thermal imager;

[0035] A path planning module is used to construct a flight path based on the single effective thermal imaging width, ground 3D coordinate information, and 3D coordinate information of ultra-high-risk equipment;

[0036] The thermal imaging acquisition module is used to construct a heat distribution cloud map in the target area through the thermal imager installed on the bottom of the drone during the drone flight operation, and then obtain the heat distribution cloud map corresponding to different flight altitudes;

[0037] A multi-source data fusion module is used to construct a temperature distribution map based on multiple heat distribution cloud maps and multiple humidity values, and calculate the drip irrigation amount in the target area based on the current temperature distribution map. The temperature distribution map represents the high-precision low-altitude vegetation surface temperature distribution.

[0038] In a third aspect, an embodiment of the present application provides a control device for an inspection drone, the device comprising a memory and a processor.

[0039] The memory is used to store computer programs; the processor is used to implement the steps of the control method of the above-mentioned inspection drone when executing the computer program.

[0040] In a fourth aspect, an embodiment of the present application provides a medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the control method of the above-mentioned inspection drone are implemented.

[0041] The beneficial effects of the present invention are:

[0042] The control method of the inspection drone described in the present invention can dynamically optimize the flight path through a wind speed adaptive flight altitude adjustment mechanism combined with the single effective width calculation of the thermal imager. At the same time, in the later data processing, by collecting ground heat distribution data of the same target area at different altitudes and combining the ambient humidity corresponding to different altitudes, the interference of the thermal imaging data by the flight altitude and ambient humidity is reduced, thereby significantly improving the accuracy of environmental data collection.

[0043] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the embodiments of the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a flow chart of a control method for an inspection drone according to an embodiment of the present invention;

[0046] Figure 2 is a schematic diagram of the principle of constructing a fusion feature corresponding to target height based on a neural network according to an embodiment of the present invention;

[0047] Figure 3 It is a schematic diagram of the control device structure of an inspection drone described in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] It should be noted that similar reference numerals or letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0050] Embodiment 1:

[0051] like Figure 1 As shown, this embodiment provides a control method for an inspection drone, which includes step S100, step S200, step S300, step S400 and step S500.

[0052] Due to its own structural factors, the drone is unable to accurately collect thermal imaging data at lower altitudes. That is, the drone's propellers will blow the target vegetation during the low-altitude heat collection process, causing heat disturbance around the target vegetation, which in turn leads to extremely poor progress in heat collection. Therefore, the conventional practice is to increase the cruising altitude of the drone to reduce its impact on the ground temperature. However, the imaging progress of the infrared thermal imager is greatly affected by the detection distance and the humidity of the environment. Especially for some hilly planting areas, the longitudinal height fluctuations are large, which leads to increased fluctuations in the environmental humidity. It is impossible to simply correct the heat distribution through a single environmental humidity. Therefore, this embodiment sets corresponding humidity sensors at different heights in the planting area.

[0053] Step S100: In response to a preset inspection instruction, the current target area wind speed, terrain information, and multiple humidity values fed back by multiple humidity sensors located at different heights are obtained. The terrain information includes three-dimensional ground coordinate information and three-dimensional coordinate information of ultra-high-risk equipment. The multiple humidity values represent humidity conditions at different heights and are used to later construct a humidity cloud map that matches the terrain information.

[0054] Step S200: Based on the wind speed in the target area, a corresponding flight altitude threshold range is obtained from a preset operation and maintenance parameter comparison table, and multiple different flight altitudes are selected. The single effective thermal imaging width is calculated based on the flight altitude and the lens parameters of the thermal imager. Specifically, the actual ground width corresponding to the width effective pixel is determined based on the width effective pixel points at different altitudes and the lens. This will not be described in detail here.

[0055] Step S300: constructing a flight path based on a single effective thermal imaging width, ground three-dimensional coordinate information, and ultra-high-risk equipment three-dimensional coordinate information, and constructing a flight restricted zone based on the ultra-high-risk equipment three-dimensional coordinate information;

[0056] Step S400: performing a UAV flight operation based on the flight path and flight altitude, during which a thermal imager provided at the bottom of the UAV is used to construct a heat distribution cloud map within the target area, thereby obtaining heat distribution cloud maps corresponding to different flight altitudes;

[0057] Step S500: construct a temperature distribution map based on the multiple heat distribution cloud maps and the multiple humidity values, and calculate the drip irrigation amount in the target area based on the current temperature distribution map, wherein the temperature distribution map represents the high-precision low-altitude vegetation surface temperature distribution.

[0058] The control method of the inspection drone described in this embodiment can dynamically optimize the flight path through a wind speed adaptive flight altitude adjustment mechanism combined with the single effective width calculation of the thermal imager. At the same time, in the later data processing, by collecting ground heat distribution data of the same target area at different altitudes and combining the ambient humidity corresponding to different altitudes, the interference of the thermal imaging data by the flight altitude and ambient humidity is reduced, thereby significantly improving the accuracy of environmental data collection.

[0059] Secondly, a specific implementation method of constructing a temperature distribution map based on the multiple heat distribution cloud maps and the multiple humidity values may be:

[0060] Before building the temperature distribution map, it is necessary to build a humidity cloud map that matches the terrain information based on the humidity values. The key here is to calculate the humidity at different heights.

[0061] However, the interpolation methods of humidity data in existing technologies are mostly based on linear assumptions, which makes it difficult to describe the nonlinear vertical humidity distribution law, affecting the accuracy of the subsequent humidity cloud map construction.

[0062] This embodiment provides a new vertical humidity construction method, specifically:

[0063] Step S511: Obtain humidity values at four different heights, and calculate the humidity at any height based on the four height-humidity values:

[0064] H(h)=a(h-h1) 3 +b(h-h1) 2 +c(h-h1)+H1;

[0065] Where h is any height, h1 is the minimum height, H1 is the humidity value corresponding to the minimum height h1; a, b, c are coefficients, and the solution process is:

[0066]

[0067] Where Δh i =h i -h1, i=2, 3, 4; H i is the height h i Corresponding humidity values, this embodiment is based on a humidity interpolation model solved by a cubic polynomial and a matrix to accurately depict the vertical humidity distribution;

[0068] Step S512: constructing a humidity distribution cloud map based on the humidity value corresponding to any height and the three-dimensional ground coordinate information, and constructing a temperature distribution map based on the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map;

[0069] The specific implementation method of constructing the temperature distribution map based on the humidity cloud map, the first heat distribution cloud map and the second heat distribution cloud map is as follows:

[0070] Step S520: Grid-based spatial alignment is performed on the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map to ensure that the features of the same geographic coordinate point can be accurately matched in the subsequent neural network. The thermal value in each cell is normalized. The data with unified dimension makes the neural network weight learning more stable and effectively reduces the subsequent gradient explosion risk. The humidity value in each cell is Z-Score normalized.

[0071] Step S530: extracting deep features of the first heat distribution cloud map through the first convolutional neural network branch, extracting deep features of the second heat distribution cloud map through the second convolutional neural network branch; extracting shallow features of the humidity cloud map through the third convolutional neural network branch, wherein the first convolutional neural network branch and the second convolutional neural network branch each include three convolutional layers, and the third convolutional neural network branch includes two convolutional layers;

[0072] Step S540: Construct the vertical thermal gradient corresponding to each cell based on the height difference between the first heat distribution cloud image and the second heat distribution cloud image. In this process, it is necessary to overcome the problem of pixel alignment of images at different heights, and further introduce uncertainty modeling in the gradient calculation:

[0073] G final =αG local +(1-α)Gglobal ;

[0074] Among them, α is the adaptive weight coefficient, which is dynamically adjusted by the humidity value, G local is the local vertical thermal gradient of 3*3 pixels; G global is the global vertical thermal gradient of 8*8, G final To integrate the vertical thermal gradient;

[0075] G = (E1-E2) / Δz; E1 represents the heat value of the area corresponding to the first heat distribution cloud map, E2 represents the heat value of the area corresponding to the second heat distribution cloud map, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, G is the vertical thermal gradient, G is G local or G global ;

[0076] Step S550: Figure 2 As shown, the deep features of the first heat distribution cloud map and the deep features of the second heat distribution cloud map are channel-joined, and a channel weight map is generated through the attention module. The calculation process of the weight map injects the shallow features of the humidity cloud map as a bias term to generate the fusion feature corresponding to the target height:

[0077]

[0078] Among them, F f is the fusion feature, W is the channel weight map, F1 is the depth feature of the first heat distribution cloud map, F2 is the depth feature of the second heat distribution cloud map, G final To incorporate vertical thermal gradients, represents the channel weighting, represents concatenation, ⊙ represents element-wise multiplication;

[0079] Step S560: Upsample and reconstruct the fused features through a physical constraint decoder to output a temperature distribution map corresponding to the target height, wherein the physical constraint decoder includes 5 layers of transposed convolution and introduces a differentiable heat conduction equation constraint term into the loss function:

[0080]

[0081] in, is the physical constraint number loss term, which is used to constrain the prediction results of the model through thermodynamic equations; λ is the constraint strength coefficient; is the heat diffusion term, is the vertical convection term, βM is the humidity correction term, κ and β are learnable parameters, M is the humidity distribution value, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, ΔT is the temperature difference between the first heat distribution cloud map and the second heat distribution cloud map in the same corresponding area, and H is the predicted heat value corresponding to the target height;

[0082] The principle of the above steps is explained as follows: the thermal images taken by the drone during cruising at different altitudes will have imaging loss due to the shooting distance and ambient humidity. The main cause of this loss is the flight altitude of the drone, and the ideal amount distribution cloud map actually required for drip irrigation is generated by the drone when it is 1-1.5 meters higher than the vegetation. Therefore, this embodiment needs to use a neural network model to analyze the thermal imaging differences at different altitudes at the same ground coordinates to infer the heat distribution cloud map corresponding to the ideal height, that is, the temperature distribution map. In this inference process, it is necessary to take into account the impact of the humidity difference caused by the altitude of different planting positions on the temperature change gradient.

[0083] Step S570: Divide the temperature distribution map according to the preset minimum drip irrigation control unit, calculate the total heat value corresponding to each minimum drip irrigation control unit, find the corresponding drip irrigation amount in the drip irrigation operation and maintenance parameter comparison table based on the total heat value, and implement the drip irrigation operation based on the corresponding drip irrigation amount.

[0084] Example 2:

[0085] This embodiment provides a control system for an inspection drone, the system comprising:

[0086] an acquisition module, configured to respond to a preset inspection instruction and obtain the current wind speed and terrain information of the target area, and multiple humidity values fed back by multiple humidity sensors located at different heights, wherein the terrain information includes three-dimensional ground coordinate information and three-dimensional coordinate information of ultra-high-risk equipment;

[0087] a parameter calling module, configured to obtain a corresponding flight altitude threshold range from a preset operation and maintenance parameter comparison table based on the wind speed in the target area, select a plurality of different flight altitudes, and calculate a single effective thermal imaging width based on the flight altitude and the lens parameters of the thermal imager;

[0088] A path planning module is used to construct a flight path based on the single effective thermal imaging width, ground 3D coordinate information, and 3D coordinate information of ultra-high-risk equipment;

[0089] The thermal imaging acquisition module is used to construct a heat distribution cloud map in the target area through the thermal imager installed on the bottom of the drone during the drone flight operation, and then obtain the heat distribution cloud map corresponding to different flight altitudes;

[0090] A multi-source data fusion module is used to construct a temperature distribution map based on multiple heat distribution cloud maps and multiple humidity values, and calculate the drip irrigation amount in the target area based on the current temperature distribution map. The temperature distribution map represents the high-precision low-altitude vegetation surface temperature distribution.

[0091] It should be noted that, regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0092] Example 3:

[0093] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides a control device for an inspection drone. The control device for an inspection drone described below and the control method for an inspection drone described above can refer to each other.

[0094] Figure 3 FIG. 1 is a block diagram of a control device for an inspection drone according to an exemplary embodiment. Figure 3 As shown, the electronic device 800 may include: a processor 801 , a memory 802 , and may further include one or more of a multimedia component 803 , an I / O interface 804 , and a communication component 805 .

[0095] The processor 801 is used to control the overall operation of the electronic device 800 to complete all or part of the steps in the control method of the inspection drone described above. The memory 802 is used to store various types of data to support the operation of the electronic device 800. Such data may include, for example, instructions for any application or method operating on the electronic device 800, as well as application-related data, such as contact information, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent through the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules. The above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the electronic device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G or 4G, or a combination of one or more thereof, so the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module, an NFC module.

[0096] In an exemplary embodiment, the electronic device 800 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the above-mentioned control method for the inspection drone.

[0097] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, the program instructions implement the steps of the above-mentioned inspection drone control method. For example, the computer-readable storage medium may be the above-mentioned memory 802 including the program instructions. The program instructions may be executed by the processor 801 of the electronic device 800 to implement the above-mentioned inspection drone control method.

[0098] Embodiment 4:

[0099] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides a readable storage medium. The readable storage medium described below and the control method of the inspection drone described above can refer to each other.

[0100] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the control method of the inspection drone of the above method embodiment.

[0101] The readable storage medium may specifically be any readable storage medium that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0102] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A control method for an inspection drone, characterized in that: The method comprises: In response to a preset inspection instruction, the system obtains the current wind speed and terrain information of the target area, as well as multiple humidity values fed back by multiple humidity sensors located at different heights. The terrain information includes three-dimensional ground coordinate information and three-dimensional coordinate information of ultra-high-risk equipment. Based on the wind speed in the target area, a corresponding flight altitude threshold range is obtained from a preset operation and maintenance parameter comparison table, and a plurality of different flight altitudes are selected, and a single effective thermal imaging width is calculated based on the flight altitude and the lens parameters of the thermal imager; Construct a flight path based on the single effective thermal imaging width, ground 3D coordinate information, and 3D coordinate information of ultra-high-risk equipment; The drone performs flight operations based on the flight path and flight altitude. During the flight, a thermal imager located at the bottom of the drone is used to construct a heat distribution cloud map within the target area, thereby obtaining heat distribution cloud maps corresponding to different flight altitudes. A temperature distribution map is constructed based on the multiple heat distribution cloud maps and the multiple humidity values, and the drip irrigation amount in the target area is calculated based on the current temperature distribution map. The temperature distribution map represents the surface temperature distribution of low-altitude vegetation with high precision.

2. The control method of the inspection drone according to claim 1, characterized in that: Constructing a temperature distribution map based on the plurality of heat distribution cloud maps and the plurality of humidity values includes: Get the humidity values at four different heights and calculate the humidity at any height based on the four height-humidity values: H(h)=a(h-h1) 3 +b(h-h1) 2 +c(h-h1)+H1; Where h is any height, h1 is the minimum height, H1 is the humidity value corresponding to the minimum height h1; a, b, c are coefficients, and the solution process is: Where Δh i =h i -h1, i=2, 3, 4; H i is the height h i Corresponding humidity value; A humidity distribution cloud map is constructed based on the humidity value corresponding to any height and the three-dimensional coordinate information of the ground, and a temperature distribution map is constructed based on the humidity cloud map, the first heat distribution cloud map and the second heat distribution cloud map.

3. The control method of the inspection drone according to claim 2, characterized in that: Constructing a temperature distribution map based on the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map includes: The first convolutional neural network branch extracts the deep features of the first heat distribution cloud map, the second convolutional neural network branch extracts the deep features of the second heat distribution cloud map; the third convolutional neural network branch extracts the shallow features of the humidity cloud map; The vertical thermal gradient corresponding to each cell is constructed based on the height difference between the first heat distribution cloud map and the second heat distribution cloud map: G final =αG local +(1-α)G global ; Among them, α is the adaptive weight coefficient, which is dynamically adjusted by the humidity value, G local is the local vertical thermal gradient of 3*3 pixels; G global is the global vertical thermal gradient of 8*8, G final To integrate the vertical thermal gradient; G = (E1-E2) / Δz; E1 represents the heat value of the area corresponding to the first heat distribution cloud map, E2 represents the heat value of the area corresponding to the second heat distribution cloud map, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, G is the vertical thermal gradient, G is G local or G global ; The deep features of the first heat distribution cloud map and the deep features of the second heat distribution cloud map are channel-joined, and a channel weight map is generated through the attention module. The calculation process of the weight map injects the shallow features of the humidity cloud map as a bias term to generate the fusion feature corresponding to the target height: Among them, F f is the fusion feature, W is the channel weight map, F1 is the depth feature of the first heat distribution cloud map, F2 is the depth feature of the second heat distribution cloud map, G final To incorporate vertical thermal gradients, represents the channel weighting, represents concatenation, ⊙ represents element-wise multiplication; The fused features are upsampled and reconstructed through the physical constraint decoder, and then the temperature distribution map corresponding to the target height is output.

4. The control method of the inspection drone according to claim 3, characterized in that: Before extracting features from the humidity cloud map, the first heat distribution cloud map, and the second heat distribution cloud map, the method further includes: The humidity cloud map, the first heat distribution cloud map and the second heat distribution cloud map are subjected to grid spatial alignment processing, the thermal value in each cell is normalized, and the humidity value in each cell is subjected to Z-Score standardization processing.

5. The control method of the inspection drone according to claim 3, characterized in that: The first convolutional neural network branch and the second convolutional neural network branch each include 3 convolutional layers, and the third convolutional neural network branch includes 2 convolutional layers.

6. The control method of the inspection drone according to any one of claims 4-5, characterized in that: The method of upsampling and reconstructing the fused features through the physical constraint decoder to output a temperature distribution map corresponding to the target height includes: The physical constraint decoder contains 5 layers of transposed convolution and introduces a differentiable heat conduction equation constraint term in the loss function: in, is the physical constraint number loss term, which is used to constrain the prediction results of the model through thermodynamic equations; λ is the constraint strength coefficient; is the heat diffusion term, is the vertical convection term, βM is the humidity correction term, α and β are learnable parameters, M is the humidity distribution value, Δz is the height difference between the first heat distribution cloud map and the second heat distribution cloud map, ΔT is the temperature difference between the first heat distribution cloud map and the second heat distribution cloud map in the same corresponding area, and H is the predicted heat value corresponding to the target height.

7. The control method of the inspection drone according to claim 1, characterized in that: The calculating of the drip irrigation amount in the target area based on the current temperature distribution map includes: The temperature distribution map is divided according to the preset minimum drip irrigation control unit, and the total heat value corresponding to each minimum drip irrigation control unit is calculated. Based on the total heat value, the corresponding drip irrigation amount is found in the drip irrigation operation and maintenance parameter comparison table, and the drip irrigation operation is implemented based on the corresponding drip irrigation amount.

Images