A method and system for optimizing drone inspection paths

Abstract

The present application relates to the field of drone inspection technology, and in particular to a method and system for optimizing drone inspection paths, the method comprising: obtaining multiple inspection areas in a global map of a material yard, and determining the scanning density of the inspection areas; calculating the scanning power required to scan each inspection area along a preset trajectory based on the scanning density; in response to the remaining power of the drone being no less than the power required for the global optimal path, using the global optimal path as the inspection path; otherwise, deleting any inspection area in the global optimal path until the remaining power is no less than the power required for the global optimal path, obtaining the inspection path; obtaining the inspection paths of the deleted inspection areas until all inspection paths cover all inspection areas. Through the technical solution of the present application, it is possible to ensure that the drone accurately obtains the optimization results of the inspection path under the premise of being able to return safely.

Description

Technical Field

[0001] The present application relates to the field of drone inspection technology, and in particular to a method and system for optimizing drone inspection paths. Background Art

[0002] As a storage place for sand, gravel, production raw materials and other materials, the material yard is responsible for storing and transferring materials. The amount of materials in the material yard has a significant impact on the company's production planning. Therefore, it is necessary to inspect the material yard regularly or irregularly to obtain the amount of materials in the material yard.

[0003] At present, the patent application document with application publication number CN111256703A discloses a multi-rotor UAV inspection path planning method, which includes: taking the minimum battery energy consumption and the cost of avoiding backlight as the objective function, completing path planning through an improved ant colony algorithm and an improved A* hybrid algorithm; the specific steps are as follows: Step 1, to ensure the safety and low energy consumption of the inspection, establish the minimum objective function of battery energy consumption and the cost of avoiding backlight; Step 2, analyze the factors affecting the battery energy consumption of the multi-rotor UAV inspection, and determine the hovering energy consumption and cruising energy consumption; Step 3, determine the relationship between sunlight and the multi-rotor UAV inspection track; Step 4, model the three-dimensional point cloud model of the scanned object to be inspected, and input the coordinates of all viewpoints during the cruise process; Step 5, use the improved ant colony algorithm to output the optimal cruise path, use the improved A* hybrid algorithm to output the optimal flight path between two adjacent viewpoints, and finally output the optimal track.

[0004] The above method takes the lowest battery energy consumption and backlight avoidance cost as the objective function, providing an optimal path with low energy consumption and backlight avoidance for drone inspections; however, the above method does not consider the impact of the remaining battery power of the drone on the inspection path during the inspection process, and cannot ensure that the drone can obtain an accurate inspection path under the premise of being able to return safely. Summary of the Invention

[0005] In order to solve the technical problem of inaccurate UAV inspection paths, the present application provides a method and system for optimizing UAV inspection paths, which can ensure that the UAV can accurately obtain the optimization results of the inspection path under the premise of being able to return safely.

[0006] In a first aspect, the present application provides a method for optimizing an inspection path of an unmanned aerial vehicle (UAV), the optimization method comprising: performing threshold segmentation on a global map of a material yard to obtain a plurality of inspection areas, and determining a scanning density of the inspection area based on the complexity of gradient information in each inspection area; determining a preset trajectory for each inspection area, and calculating a scanning power required to scan each inspection area along the preset trajectory based on the scanning density; in response to the remaining power of the UAV being not less than the required power of the global optimal path, using the global optimal path as the inspection path, and conversely, deleting any inspection area in the global optimal path until the remaining power is not less than the required power of the global optimal path, thereby obtaining an inspection path; and obtaining inspection paths for the deleted inspection areas until each inspection path covers all inspection areas. The method for obtaining the global optimal path is as follows: obtain the permutations and combinations of preset trajectories in each inspection area; take the charging position as the initial position, obtain the flight path from the initial position to the two endpoints of the first preset trajectory in any permutation and combination, and use the endpoint with the shortest flight path as the starting point and the other endpoint as the end point; take the end point of the first preset trajectory as the initial position, obtain the starting point and end point of each preset trajectory in turn, and obtain the flight path from the end point of the last preset trajectory to the charging position to obtain the global path of the permutation and combination; the global optimal path is the global path with the shortest total length.

[0007] An image acquisition device deployed in the material yard is used to obtain a global map of the material yard, and then the material stacking area in the material yard is determined, and the material stacking area is used as the inspection area; the scanning difficulty in the inspection area is measured according to the complexity of the gradient information in the inspection area, and a larger scanning density is assigned to the inspection area with greater complexity to accurately calculate the material quantity in each inspection area; further, the scanning power for scanning each inspection area along the preset trajectory is calculated based on the scanning density, and the required power for the global optimal path is obtained, and the remaining power of the drone is compared with the required power of the global optimal path. If the remaining power of the drone is not less than the required power of the global optimal path, the global optimal path is used as the inspection path. Otherwise, any inspection area in the global optimal path is deleted until the inspection path is obtained; the inspection path of the deleted inspection area is obtained again until each inspection path covers all inspection areas; automatic batching of inspection tasks is realized to ensure that the drone accurately obtains the optimization result of the inspection path under the premise of being able to return safely.

[0008] Preferably, obtaining multiple inspection areas includes: converting the global map of the material yard into the HSV space to obtain a hue channel map; using the maximum inter-class variance method to determine the segmentation threshold of the hue channel map, and performing threshold segmentation on the hue channel map to obtain a binary map; performing connected domain analysis on the binary map, obtaining multiple connected domains and using the circumscribed rectangle of each connected domain as the inspection area.

[0009] Accurately identify the material stacking area in the material yard and use the material stacking area as the inspection area. During the drone inspection process, you only need to use the laser radar on the drone to scan the inspection area to calculate the volume of materials in the inspection area.

[0010] Preferably, the inspection area Scan density for:

[0011] , is the density reference value, For inspection area The complexity of the internal gradient information, is the maximum allowed value of the gradient information, is the preset coefficient.

[0012] When the complexity of the gradient information is large, it means that the material has an irregular thickness and is accumulated in the inspection area. In order to accurately obtain the thickness of each point in the inspection area and then accurately calculate the volume of the material in the inspection area, the scanning density of the inspection area needs to be increased.

[0013] Preferably, the method for obtaining the preset coefficient includes: scanning the inspection area with an arbitrary scanning density to obtain the material thickness at each scanning position in the inspection area, interpolating the material thickness at each scanning position to obtain the material thickness at each position point in the inspection area; integrating the material thickness at each position point on the plane where the inspection area is located to obtain the material quantity in the inspection area; drawing a material quantity curve with the scanning density as the horizontal coordinate and the material quantity as the vertical coordinate, and using the scanning density corresponding to the inflection point in the material quantity curve as the density label of the inspection area; and calculating the preset coefficient based on the complexity of the inspection area and the density label.

[0014] Preferably, determining the preset trajectory of each inspection area includes: obtaining the size information of the inspection area, the size information including the length and width; The shortest sides of the length and width are divided into multiple scanning sub-areas, and the preset trajectory of the inspection area is constructed according to the midline of each scanning sub-area.

[0015] Preferably, calculating the scanning power of each inspection area along the preset track according to the scanning density includes: obtaining the scanning power corresponding to the scanning density of the inspection area; Scanning power for:

[0016] ; and are flight power and scanning power respectively, For the preset flight speed, is the length of the preset trajectory.

[0017] Since the scanning density of each inspection area is different, the power required to complete the scanning of each inspection area is different. Accurately calculating the scanning power of each inspection area provides a data basis for ensuring the safe return of the drone in the subsequent path optimization process.

[0018] Preferably, the power required for the global optimal path is the sum of the scanning power of each inspection area and the flight power consumption of the flight path between the inspection areas.

[0019] Accurately quantify the amount of power required for a drone to complete scanning of each inspection area along the globally optimal path.

[0020] Preferably, deleting any inspection area in the global optimal path includes: obtaining the global optimal path again after deleting any inspection area, and calculating the reduction in power required for the global optimal path after deleting the inspection area; and deleting the inspection area corresponding to the maximum reduction.

[0021] The greater the reduction, the greater the impact of the inspection area on the power required for the global optimal path. Deleting the inspection area can effectively reduce the power required for the global optimal path, thereby quickly obtaining the inspection path.

[0022] Preferably, after completing the scanning of each inspection area along each inspection path, the optimization method also includes: interpolating the material thickness of each scanning position in the inspection area to obtain the material thickness of each position point in the inspection area; integrating the material thickness of each position point on the plane where the inspection area is located to obtain the material quantity in the inspection area; and taking the sum of the material quantities in each inspection area as the total material amount in the material yard.

[0023] In the second aspect of the present application, a system for optimizing drone inspection paths is provided, comprising a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a method for optimizing drone inspection paths according to the first aspect of the present application is implemented.

[0024] The technical solution of this application has the following beneficial technical effects:

[0025] An image acquisition device deployed in the material yard is used to obtain a global map of the material yard, and then the material stacking area in the material yard is determined, and the material stacking area is used as the inspection area; the scanning difficulty in the inspection area is measured according to the complexity of the gradient information in the inspection area, and a larger scanning density is assigned to the inspection area with greater complexity to accurately calculate the material quantity in each inspection area; further, the scanning power for scanning each inspection area along the preset trajectory is calculated based on the scanning density, and the required power for the global optimal path is obtained, and the remaining power of the drone is compared with the required power of the global optimal path. If the remaining power of the drone is not less than the required power of the global optimal path, the global optimal path is used as the inspection path. Otherwise, any inspection area in the global optimal path is deleted until the inspection path is obtained; the inspection path of the deleted inspection area is obtained again until each inspection path covers all inspection areas; on the premise of ensuring that the drone can return safely, the inspection task is automatically batched, the optimization result of the inspection path is accurately obtained, and the material quantity in the material yard is accurately obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flowchart of a method for optimizing a drone inspection path according to an embodiment of the present application.

[0027] Figure 2 It is a schematic diagram of scanning an inspection area along a preset trajectory according to an embodiment of the present application.

[0028] Figure 3 This is a structural block diagram of a drone inspection path optimization system according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0030] According to the first aspect of the present application, the present application provides a method for optimizing a drone inspection path, which is used to inspect the material stacking area in the material yard and determine the inspection path to accurately obtain the material quantity in the material yard. Figure 1 This is a flow chart of a method for optimizing a drone inspection path according to an embodiment of the present application. Figure 1 As shown, the method for optimizing the UAV inspection path includes steps S101 to S103, which are described in detail below.

[0031] S101 , performing threshold segmentation on the global map of the material yard to obtain multiple inspection areas, and obtaining the complexity of the gradient information in each inspection area, and determining the scanning density of the inspection area according to the complexity.

[0032] In one embodiment, the material yard can be regarded as a warehouse for material storage, and an image acquisition device deployed in the material yard is used to collect a global map of the material yard. The global map of the material yard is an RGB image, and the global map of the material yard can cover all locations in the material yard; the global map of the material yard can reflect the material stacking area in the material yard. The purpose of using a drone to inspect the material yard is to inspect each material stacking area, and then obtain the volume of all materials in the material yard.

[0033] The material storage areas in the global map of the material yard can be determined based on the color difference between the materials and the material yard ground surface. These material storage areas serve as inspection areas. Specifically, obtaining multiple inspection areas includes: converting the global map of the material yard into HSV space to obtain a hue channel map; determining the segmentation threshold of the hue channel map using the maximum inter-class variance method, and performing threshold segmentation on the hue channel map to obtain a binary map; performing connected domain analysis on the binary map to obtain multiple connected domains, and using the bounding rectangle of each connected domain as the inspection area.

[0034] The multiple connected domains are multiple material storage areas.

[0035] In this way, the material stacking area in the material yard can be accurately identified and used as the inspection area. During the drone inspection process, it is only necessary to use the laser radar carried by the drone to scan the inspection area to calculate the volume of materials in the inspection area.

[0036] In one embodiment, a drone equipped with a laser radar scans each inspection area to collect the material thickness at each point in any inspection area, and then calculates the volume of the material in the inspection area. Considering the different complexity of material stacking in different inspection areas, the difficulty of calculating the material volume in each inspection area will also vary. For example, if the material is laid out in the inspection area with the same thickness, scanning the inspection area with a lower scanning density can accurately collect the material thickness and then accurately calculate the material volume in the inspection area. If the material is piled up in the inspection area with irregular thickness, the lower scanning density will ignore the thickness variation of the material. Therefore, it is necessary to increase the scanning density to accurately obtain the thickness of each point in the inspection area and then accurately calculate the material volume in the inspection area.

[0037] After grayscale processing of the global map of the material yard, the gradient information of each position point in any inspection area is calculated. If the material is spread in the inspection area with the same thickness, the gradient information of each position point is 0; on the contrary, if the material is piled up in the inspection area with irregular thickness, for the position points with thickness changes, their gradient information is greater than 0, and the greater the thickness change, the greater the value of the gradient information; therefore, the scanning density of the inspection area can be determined according to the complexity of the gradient information in the inspection area.

[0038] Specifically, the inspection area Complexity of internal gradient information For inspection area The average gradient value within satisfies the relationship: ; is the maximum allowed value of the gradient information, For inspection area Intrinsic gradient information The frequency of occurrence.

[0039] Furthermore, according to the inspection area Complexity of internal gradient information Determine the scanning density of the inspection area, inspection area Scan density for: , is the density reference value, For inspection area The complexity of the internal gradient information, is the maximum allowed value of the gradient information, is the preset coefficient.

[0040] Among them, the density reference value The value of , The unit of scanning density is the number of laser points per square meter. The maximum value allowed for gradient information is 255, which is used to adjust the complexity. Normalized to a value between 0 and 1.

[0041] The greater the complexity, the higher the scanning density needs to be in order to accurately calculate the material volume. The preset coefficient can control the growth rate of the scanning density with the complexity, and the value of the preset coefficient can be determined based on the historical scanning process. Specifically, the method for obtaining the preset coefficient includes: scanning the inspection area with an arbitrary scanning density to obtain the material thickness at each scanning position in the inspection area, interpolating the material thickness at each scanning position to obtain the material thickness at each position point in the inspection area; integrating the material thickness at each position point on the plane where the inspection area is located to obtain the material quantity in the inspection area; drawing a material quantity curve with the scanning density as the horizontal coordinate and the material quantity as the vertical coordinate, and using the scanning density corresponding to the inflection point in the material quantity curve as the density label of the inspection area; and calculating the preset coefficient based on the complexity and density label of the inspection area.

[0042] It is understandable that when the scanning density is low, the material thickness of each position point in the inspection area cannot be accurately collected, and the material quantity obtained at this time has a large error; as the scanning density increases, the fewer the number of position points between adjacent scanning positions, the more accurately the material thickness of each position point in the inspection area can be collected, and the material quantity gradually approaches the true value; when the scanning density reaches the scanning density corresponding to the inflection point, the material quantity tends to stabilize. At this time, further increasing the scanning density will not only not significantly increase the accuracy of the material quantity but will also introduce a large amount of calculation. Therefore, using the scanning density corresponding to the inflection point as the density label can balance the material quantity accuracy and the calculation amount.

[0043] In this way, the stacking area of ​​the materials in the material yard is determined based on the overall map of the material yard, and the stacking area is used as the inspection area in the subsequent inspection process. The scanning density of each inspection area during the inspection process is determined according to the complexity of the gradient information to ensure that the amount of materials in the material yard can be accurately obtained after the inspection.

[0044] S102 , determining a preset track for each inspection area, and calculating a scanning power required to scan each inspection area along the preset track according to a scanning density.

[0045] In one embodiment, the inspection area is a rectangular area. When the drone scans the inspection area along a preset trajectory at a preset flight altitude, a preset flight speed, and a corresponding scanning density, the larger the size information of the inspection area, the greater the scanning density, and the longer the scanning time required, resulting in a greater amount of scanning power required to complete the scan.

[0046] See Figure 2 , is a schematic diagram of scanning the inspection area along a preset trajectory according to an embodiment of the present application. The preset trajectory of the inspection area is related to the inspection area size information, which includes the length and width of the inspection area. In order to reduce the number of times the drone changes its flight direction in the preset trajectory, the scanning width of the laser radar is used. Divide the shortest sides of the length and width into multiple scanning sub-areas, and construct a preset trajectory of the inspection area based on the midpoint of each scanning sub-area; Figure 2 The inspection area shown has 4 scanning sub-areas in total. When the UAV flies along the center line of any scanning sub-area, it can scan all the position points in the scanning sub-area. Figure 2 The preset trajectory shown includes the center lines of four scanning sub-areas and three changes in flight direction.

[0047] In one embodiment, due to the different scanning densities of each inspection area, the scanning power required to complete each inspection area is different. To ensure that the drone can return safely after the inspection, it is necessary to accurately predict the scanning power of each inspection area based on the size information and scanning density of each inspection area. Specifically, calculating the scanning power of each inspection area along the preset trajectory based on the scanning density includes: obtaining the scanning power corresponding to the scanning density of the inspection area; Scanning power for: ; and are flight power and scanning power respectively, For the preset flight speed, is the length of the preset trajectory. The flight power is related to the model of the drone.

[0048] In this way, the scanning power of each inspection area is determined.

[0049] S103, in response to the remaining power of the drone being no less than the power required by the global optimal path, the global optimal path is used as the inspection path. Conversely, any inspection area in the global optimal path is deleted until the remaining power is no less than the power required by the global optimal path, and the inspection path is obtained.

[0050] In one embodiment, obtaining the global optimal path includes: obtaining the permutation and combination of preset trajectories of each inspection area; taking the charging position as the initial position, obtaining the flight path from the initial position to the two endpoints of the first preset trajectory in any permutation and combination, taking the endpoint with the shortest flight path as the starting point, and the other endpoint as the ending point; taking the ending point of the first preset trajectory as the initial position, obtaining the starting point and ending point of each preset trajectory in turn, and obtaining the flight path from the ending point of the last preset trajectory to the charging position, to obtain the global path of the permutation and combination; and taking the global path with the shortest total length as the global optimal path.

[0051] For example, for the convenience of description, the five inspection areas are marked as a, b, c, d, and e, and the charging position is marked as 0; each inspection area corresponds to a preset track, and each preset track corresponds to two endpoints. The two endpoints of the preset track corresponding to inspection area a are marked as a1 and a2. For the permutation combination cdeab, first, the charging position 0 is used as the initial position, and the path planning algorithm is used to obtain the flight path from the initial position to the two endpoints of the first preset track in the permutation combination (that is, the preset track corresponding to inspection area c), which are respectively marked as and ; and the shortest flight path is ,Will As the starting point of the first preset trajectory, As the end point of the first preset trajectory; further, the end point of the first preset trajectory For the initial position, obtain the flight path from the initial position to the two endpoints of the next preset trajectory (i.e., the preset trajectory corresponding to inspection area d), and then determine the starting and ending points of the next preset trajectory. In this way, the starting and ending points of all preset trajectories in the permutation combination cdeab are obtained. See Table 1 for the starting and ending points of each preset trajectory in the permutation combination cdeab.

[0052] Table 1 Starting and ending points of each preset trajectory

[0053]

[0054] In this way, the global path of the permutation and combination cdeab is: ;in, is the preset trajectory corresponding to the inspection area c, The path planning algorithm is used to obtain to The flight path is a path planning algorithm that can use the A-star algorithm or the artificial potential field method, which is not limited in this application. After obtaining all permutations and combinations of global paths, the global path with the smallest total length is determined as the global optimal path. In one embodiment, the power required for the global optimal path is the sum of the scanning power consumption of each inspection area and the flight power consumption of the flight path between each inspection area.

[0055] After obtaining the flight path, the flight power consumption of each flight path is calculated. In the scenario of the embodiment of the present application, the drone performs inspections at a set flight altitude and a set flight speed, and the drone does not need to scan when flying along the flight path to the next inspection area. Therefore, the ratio between the length of any flight path and the preset flight speed is used as the predicted flight time, and the product of the predicted flight time and the flight power is used as the flight power consumption of the flight path.

[0056] In one embodiment, in response to the drone's remaining battery power being no less than the required battery power of the global optimal path, indicating that the drone's remaining battery power is sufficient to complete scanning of each inspection area and safely return to a charging location, the global optimal path is used as the inspection path, and an optimization result for the drone's inspection path is obtained. Conversely, in response to the drone's remaining battery power being less than the required battery power of the global optimal path, indicating that the drone's remaining battery power is insufficient to support scanning of all inspection areas, any inspection area in the global optimal path is deleted, and the remaining battery power is compared with the required battery power of the updated global optimal path until the remaining battery power is no less than the required battery power of the updated global optimal path, at which point the updated global optimal path is used as the inspection path.

[0057] Wherein, deleting any inspection area in the global optimal path includes: the said any inspection area is the last inspection area in the global optimal path. For example, if the global optimal path is: ; The last inspection area is b. After deleting the last inspection area in the global optimal path, the updated global optimal path is: .

[0058] In another embodiment, deleting any inspection area in the global optimal path includes: obtaining the global optimal path again after deleting any inspection area, and calculating the reduction in power required for the global optimal path after deleting the inspection area; and deleting the inspection area corresponding to the maximum reduction.

[0059] The greater the reduction, the greater the impact of the inspection area on the power required for the global optimal path. Deleting the inspection area can effectively reduce the power required for the global optimal path, thereby quickly obtaining the inspection path.

[0060] In this way, a globally optimal path covering all inspection areas is obtained. The globally optimal path includes pre-set trajectories within each inspection area and flight paths between them. The pre-set trajectories ensure scanning accuracy in each inspection area and accurately calculate the amount of material in each scanned area; the flight path ensures the shortest path between inspection areas, thus ensuring inspection efficiency. Furthermore, by comparing the remaining battery power of the drone with the power required for the globally optimal path, inspection tasks are automatically batched, and inspection areas further back in the globally optimal path are deleted to ensure the drone's safe return.

[0061] S104: Obtain inspection paths of the deleted inspection areas until all inspection paths cover all inspection areas.

[0062] In one embodiment, when a patrol path cannot obtain coverage of all patrol areas, it is necessary to obtain the global optimal path of each deleted patrol area again as the patrol path of each deleted patrol area until each patrol path covers all patrol areas, and the scanning of all patrol areas can be completed.

[0063] It should be noted that in order to complete the scanning of all inspection areas, at least one inspection path will be obtained; multiple drones can be deployed, and one drone can conduct inspections along one inspection path at the same time to improve inspection efficiency; or only one drone can be deployed, and when the drone finishes the inspection of one inspection path, it returns to the charging position, and starts the inspection of the next inspection path after charging is completed.

[0064] In one embodiment, after completing the scanning of each inspection area along each inspection path, the optimization method also includes: interpolating the material thickness of each scanning position in the inspection area to obtain the material thickness of each position point in the inspection area; integrating the material thickness of each position point on the plane where the inspection area is located to obtain the material quantity in the inspection area; and taking the sum of the material quantities in each inspection area as the total material amount in the material yard.

[0065] Among them, the inspection area Amount of material for: ; and Inspection areas The width and length, The location point within the inspection area Material thickness.

[0066] In this way, the inspection task of the material yard is completed and the total amount of materials in the material yard is accurately calculated.

[0067] According to the second aspect of the present application, the present application also provides a system for optimizing drone inspection paths. Figure 3 This is a structural block diagram of a drone inspection path optimization system according to an embodiment of the present application. Figure 3 As shown, the system 50 includes a processor and a memory. The memory stores computer program instructions. When executed by the processor, the computer program instructions implement the method for optimizing a drone inspection path according to the first aspect of the present application. The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are well known in the art and are therefore not described in detail here.

[0068] It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all fall within the scope of protection of the present application.

Claims

1. A method for optimizing an unmanned aerial vehicle inspection path, characterized in that: The optimization method includes: Perform threshold segmentation on the global map of the material yard to obtain multiple inspection areas; Determine the inspection area based on the complexity of the gradient information in each inspection area Scan density , the formula is: , is the density reference value, For inspection area The complexity of the internal gradient information, is the maximum allowed value of the gradient information, is the preset coefficient; determines the preset trajectory of each inspection area; Calculate the scanning power of each inspection area along the preset trajectory according to the scanning density, including: obtaining the scanning power corresponding to the scanning density of the inspection area; Scanning power for: ; and are flight power and scanning power respectively, For the preset flight speed, is the length of the preset trajectory; In response to the remaining power of the UAV being no less than the power required by the global optimal path, it indicates that the remaining power of the UAV is sufficient to complete the scanning of each inspection area and return safely to the charging location, and the global optimal path is used as the inspection path; otherwise, any inspection area in the global optimal path is deleted until the remaining power is no less than the power required by the global optimal path, and the inspection path is obtained; Obtain the inspection paths of the deleted inspection areas until all inspection paths cover all inspection areas; The method for obtaining the global optimal path is as follows: obtain the permutation and combination of the preset trajectories of each inspection area; take the charging position as the initial position, obtain the flight path from the initial position to the two endpoints of the first preset trajectory in any permutation and combination, and take the endpoint with the shortest flight path as the starting point and the other endpoint as the ending point; take the ending point of the first preset trajectory as the initial position, obtain the starting point and ending point of each preset trajectory in turn, and obtain the flight path from the ending point of the last preset trajectory to the charging position to obtain the global path of the permutation and combination; the global optimal path is the global path with the shortest total length; the global optimal path includes the preset trajectories within each inspection area and the flight paths between the inspection areas. The preset trajectories can ensure the scanning accuracy of each inspection area, and the flight paths can ensure that the path between the inspection areas is the shortest; The power required for the global optimal path is the sum of the scanning power of each inspection area and the flight power consumption of the flight path between inspection areas; Deleting any inspection area in the global optimal path includes: obtaining the global optimal path again after deleting any inspection area, calculating the reduction in power required for the global optimal path after deleting the inspection area; and deleting the inspection area corresponding to the maximum reduction.

2. The method for optimizing an unmanned aerial vehicle inspection path according to claim 1, characterized in that: The method of obtaining multiple inspection areas includes: converting the global map of the material yard into the HSV space to obtain a hue channel map; determining the segmentation threshold of the hue channel map by using the maximum inter-class variance method, and performing threshold segmentation on the hue channel map to obtain a binary map; performing connected domain analysis on the binary map to obtain multiple connected domains and using the circumscribed rectangle of each connected domain as the inspection area.

3. The method for optimizing an unmanned aerial vehicle inspection path according to claim 1, characterized in that: The method for obtaining the preset coefficient includes: Scan the inspection area with any scanning density to obtain the material thickness at each scanning position in the inspection area, and interpolate the material thickness at each scanning position to obtain the material thickness at each position in the inspection area; The material thickness at each position point on the plane where the inspection area is located is integrated to obtain the material quantity in the inspection area; a material quantity curve is drawn with the scanning density as the horizontal coordinate and the material quantity as the vertical coordinate, and the scanning density corresponding to the inflection point in the material quantity curve is used as the density label of the inspection area; the preset coefficient is calculated based on the complexity of the inspection area and the density label.

4. The method for optimizing an unmanned aerial vehicle inspection path according to claim 1, characterized in that: Determine the preset trajectory for each inspection area including: Obtaining the size information of the inspection area, including the length and width; The shortest sides of the length and width are divided according to the scanning width of the laser radar to obtain multiple scanning sub-areas, and the preset trajectory of the inspection area is constructed according to the midline of each scanning sub-area.

5. The method for optimizing an unmanned aerial vehicle inspection path according to claim 1, characterized in that: After completing the scanning of each inspection area along each inspection path, the optimization method further includes: Interpolate the material thickness at each scanning position within the inspection area to obtain the material thickness at each position point within the inspection area; The material thickness at each position point on the plane where the inspection area is located is integrated to obtain the material quantity in the inspection area; the sum of the material quantities in each inspection area is taken as the total material quantity in the material yard.

6. A drone inspection path optimization system, characterized by: The method comprises a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for optimizing a drone inspection path according to any one of claims 1 to 5 is implemented.

Images