Railway slope inspection planning method and device

Abstract

The present application provides a kind of railway slope inspection planning method and device, it is related to measurement and control technical field, including the real scene three-dimensional model of the railway slope to be inspected is constructed;The inspection work data of unmanned aerial vehicle is calculated based on real scene three-dimensional model;The first endpoint of the railway slope to be inspected is divided based on inspection work data, and the first endpoint based on the division is constructed unmanned aerial vehicle nest node and the number of unmanned aerial vehicles is set;Based on unmanned aerial vehicle nest and unmanned aerial vehicle, inspection route planning is carried out, and the inspection scheme of the railway slope to be inspected is obtained.The present application calculates the inspection work data of unmanned aerial vehicle, arranges nest according to inspection work data, and then carries out inspection planning according to nest position, reduces the number of required nest construction, saves cost, and also improves the work efficiency of unmanned aerial vehicle inspection.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of measurement and control, in particular to a railway slope inspection planning method and device. BACKGROUND

[0002] When a railway is constructed in a mountainous area, a large number of tunnels and slopes are often built, and these engineering structures are challenged by geological activities and climate change. Landslides and other disasters on slopes seriously endanger the safety of train operation. In order to ensure the safety of high-speed rail, a slope monitoring system is usually established to regularly check the shape changes of slopes, analyze the deformation trend, and identify potential landslide hazards.

[0003] However, the existing slope monitoring methods mostly use fixed sensor equipment for continuous monitoring, which has high requirements for the surrounding environment. High-speed railway slopes are generally located in mountainous areas with complex terrain and landforms. The use of fixed cameras and radar monitoring methods is greatly limited by objective conditions such as terrain and network, making it difficult to erect and maintain fixed monitoring equipment.

[0004] The existing method of using unmanned aerial vehicles (UAVs) for slope inspection mostly requires the planning of inspection routes based on the setting of UAV nests. However, due to the inability to determine the location of well-constructed UAV nests, a large number of UAV nests need to be constructed, resulting in high consumption costs and low efficiency of UAV inspection. SUMMARY

[0005] The purpose of the present application is to provide a railway slope inspection planning method and device to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present application is as follows:

[0006] In a first aspect, the present application provides a railway slope inspection planning method, comprising:

[0007] Obtaining an oblique photogrammetry result of a railway slope to be inspected by a UAV, and constructing a real scene three-dimensional model of the railway slope to be inspected based on the oblique photogrammetry result;

[0008] Calculating inspection work data of the UAV based on the real scene three-dimensional model, the inspection work data including the inspection time of the UAV at each slope and the inspection time of the UAV between adjacent two slopes;

[0009] Dividing the railway slope to be inspected based on the inspection work data of the UAV, and constructing UAV nest nodes and setting the number of UAVs based on the first endpoints obtained by the division;

[0010] Planning an inspection route based on the UAV nest nodes and the number of UAVs to obtain an inspection scheme for the railway slope to be inspected.

[0011] In a second aspect, the application further provides a railway slope inspection planning device, comprising:

[0012] A first processing module is configured to acquire oblique photogrammetry results of a railway slope to be inspected by a UAV.

[0013] A second processing module is configured to construct a real scene three-dimensional model of the railway slope to be inspected based on the oblique photogrammetry results.

[0014] A third processing module is configured to calculate inspection work data of the UAV based on the real scene three-dimensional model, wherein the inspection work data comprises an inspection time of the UAV at each slope and an inspection time of the UAV between two adjacent slopes.

[0015] A fourth processing module is configured to divide the railway slope to be inspected into first end points based on the inspection work data of the UAV, and to construct UAV nest nodes and set a number of UAVs based on the first end points.

[0016] A fifth processing module is configured to plan an inspection route based on the UAV nest nodes and the number of UAVs, and to obtain an inspection scheme of the railway slope to be inspected.

[0017] The application has the following advantages:

[0018] The application calculates the inspection work data of the UAV, arranges the nest based on the inspection work data, and plans the inspection based on the nest position, thereby reducing the number of nests to be constructed, saving the cost, and improving the work efficiency of the UAV inspection.

[0019] Other features and advantages of the application will be illustrated in the following description, and some will become apparent from the description, or will be understood by those skilled in the art from the description, or will be understood by implementation of the embodiments of the application. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the application, and therefore should not be regarded as a limitation to the scope. For those skilled in the art, other related drawings can also be obtained from the drawings without creative labor.

[0021] Figure 1 A flowchart of a railway slope inspection planning method according to an embodiment of the application;

[0022] Figure 2 A structural diagram of a railway slope inspection planning device according to an embodiment of the application;

[0023] Figure 3A schematic diagram of an initial three-dimensional model described in the embodiments of the present application;

[0024] Figure 4 A schematic diagram of the setting position of the UAV nest node described in the embodiments of the present application;

[0025] Figure 5 A schematic diagram of the relative position of the slope node, slope-free road section and UAV nest node described in the embodiments of the present application;

[0026] Figure 6 A schematic diagram of the transfer of the UAV between the nest described in the embodiments of the present application.

[0027] In the figure: 901, first processing module; 902, second processing module; 903, third processing module; 904, fourth processing module; 905, fifth processing module; 9011, first processing unit; 9012, second processing unit; 9013, third processing unit; 9014, fourth processing unit; 9015, fifth processing unit; 9021, sixth processing unit; 9022, seventh processing unit; 9031, eighth processing unit; 9032, ninth processing unit; 90121, third processing sub-module; 90122, fourth processing sub-module; 90123, fifth processing sub-module; 90124, sixth processing sub-module; 90141, first processing sub-module; 90142, second processing sub-module. DETAILED DESCRIPTION

[0028] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. The components of the embodiments of the present application 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 application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present application.

[0029] It should be noted that: similar labels and 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 and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms “first”, “second” and the like are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0030] Embodiment 1:

[0031] The embodiment provides a railway slope inspection planning method.

[0032] Referring to Figure 1 , the method comprises steps S1, S2, S3 and S4.

[0033] S1, acquiring oblique photogrammetry results of a railway slope to be inspected by a UAV, and constructing a real scene three-dimensional model of the railway slope to be inspected based on the oblique photogrammetry results.

[0034] Specifically, step S1 comprises:

[0035] S11, acquiring first information comprising satellite image data of the railway slope to be inspected, railway surrounding terrain data and railway planning vector data, constructing an initial three-dimensional model based on the first information, and obtaining an initial three-dimensional model of the railway slope to be inspected;

[0036] Specifically, step S11 comprises:

[0037] S111, constructing a scene three-dimensional model based on the railway surrounding terrain data, wherein the scene three-dimensional model is a three-dimensional model comprising two-dimensional plane data and vertical height data of the railway slope to be inspected, the railway surrounding terrain data is acquired by stereo image pairs, SAR or other satellite remote sensing, and the scene three-dimensional model is constructed by using the railway surrounding terrain data, wherein the scene three-dimensional model is 2.5D, and comprises two-dimensional plane data and vertical height data of the railway surrounding terrain;

[0038] S112, performing spatial registration on the satellite image data of the railway slope to be inspected and the railway planning vector data to obtain railway region information to be inspected, wherein the spatial coordinates of raster images and vector data are read by using GIS software, superimposed and displayed to achieve preliminary registration, and the same named points are marked manually or automatically, and the geographic registration tool of the GIS software is used to achieve accurate registration of the two kinds of data to obtain the railway region in the image;

[0039] S113, performing target recognition on the railway region information to be inspected based on a YOLOv8-seg model, and performing pixel classification on the railway region information to be inspected after target recognition by using the YOLOv8-seg model to obtain railway slope region information to be inspected, wherein a target recognition network based on the YOLOv8-seg model is used to extract regions with landslide triangular faces, slope attachments and bare rock and soil from the image, and then a semantic segmentation network based on the YOLOv8-seg is used to perform pixel classification on the recognition result of the previous step to obtain pixel-level regions on the image in the railway region in the image, i.e., the railway slope image region to be inspected;

[0040] S114, the scene 3D model and the railway slope area information to be inspected are superimposed to obtain the initial 3D model. The scene 3D model representing the terrain is superimposed with the railway slope image to be inspected to obtain the initial 3D model of the railway and surrounding scene. The scene 3D model and the railway slope image to be inspected are superimposed according to spatial position. Based on the pixel-level image area obtained in S23, the scene 3D model is divided into two types: inspection target area and non-target area. One method is to traverse the pixels of each inspection target in the satellite image and mark the triangular face containing the pixel in each horizontal space of the 3D model. Finally, all the marked triangular faces are the target inspection area. The significance of this step is to obtain the slope area required for the inspection area division.

[0041] S12, Calculate the inspection work data corresponding to the initial three-dimensional model based on the initial three-dimensional model of the railway slope to be inspected;

[0042] Specifically, step S12 includes:

[0043] S121, as Figure 3 As shown, after obtaining the area, shape, length, width, slope and cross-section of the slope from the initial 3D model of the railway and surrounding scene, the workload of UAV inspection is initially estimated. In this embodiment, the workload of UAV inspection is considered by calculating the inspection time of UAV on each slope and the inspection time of UAV between two adjacent slopes.

[0044] S122, a rectangle is used to fit the slope area, and a triangle represents the slope shape. The circumscribed rectangle has a length *l* parallel to the railway and a width *s* perpendicular to the railway. Since the shooting is done perpendicular to the slope surface, when the camera parameters, image resolution, and vertical distance from the ground are fixed, it can be assumed that the number of flights required in the horizontal direction is proportional to the slope length, that is, one flight is performed every *n* meters in the direction perpendicular to the railway. Therefore, the formula for calculating the workload of flights in the horizontal direction, i.e., the direction parallel to the railway, is:

[0045]

[0046] Where w1 is the amount of work required for the UAV to fly in the horizontal direction, l is the length of the circumscribed rectangle corresponding to the slope parallel to the railway, s is the width of the circumscribed rectangle corresponding to the slope perpendicular to the railway, α is the reciprocal of the maximum flight speed of the UAV in the horizontal direction, and p is the slope of the slope. The idea of ​​this formula is time = distance / speed, and the calculated w1 is the time.

[0047] S123, it is also necessary to calculate the workload required for the drone to fly in the vertical direction, that is, the direction perpendicular to the railway. The calculation formula is:

[0048] w2=β·p·2s(p 2 +1) -2

[0049] Where w2 is the workload required for the drone to fly in the vertical direction, β is the reciprocal of the maximum flight speed of the drone in the vertical direction, p is the slope of the slope, and s is the width of the circumscribed rectangle of the slope perpendicular to the railway. Since there is a height difference on the slope, the height difference is calculated separately using this formula to obtain the workload w2 of the drone in the vertical direction.

[0050] S124, Through the calculations in steps S311 and S312, the total workload of the UAV on a single slope is finally obtained, and the calculation formula is as follows:

[0051]

[0052] Among them, W n Let w1 be the total workload of the UAV on a single slope, w2 be the workload of the UAV in the horizontal direction, w2 be the workload of the UAV in the vertical direction, n be the distance of each flight in the direction perpendicular to the railway, α be the reciprocal of the maximum flight speed of the UAV in the horizontal direction, β be the reciprocal of the maximum flight speed of the UAV in the vertical direction, p be the slope of the slope, and s be the width of the circumscribed rectangle of the slope perpendicular to the railway.

[0053] S125, Calculate the workload w of the UAV flying over the slopeless road section between adjacent slopes. k The calculation formula is:

[0054] w k =α·k

[0055] Where k is the distance between two adjacent slopes, i.e. the length of the road section without slopes.

[0056] S13, based on the inspection work data corresponding to the initial three-dimensional model, divide the railway slope to be inspected into second endpoints, and construct UAV nest nodes and set the number of UAVs based on the divided second endpoints;

[0057] Specifically, step S13 includes:

[0058] S131, as Figure 4 As shown, for the workload obtained in step S3, a railway is represented as an alternation between slope-free railway sections and sloped sections, where w ki w represents the workload required for the UAV to fly on the i-th slope-free road segment. niLet P1, P2, and P3 represent the total workload required for the UAV to fly at the i-th slope node, where P1, P2, and P3 represent three adjacent UAV nest nodes. All required workloads can be represented by a workload list [w]. k1 ,w n1 ,w k2 ,w n2 ,w k3 ,w n3 ,……]express;

[0059] S132, for a specific type of drone, the drone's working capacity is considered based on its maximum flight time, with a 20% allowance for unexpected events, and the remaining portion is considered the drone's effective working capacity. Taking the DJI Matrice 300RTK as an example, in P mode, the maximum horizontal flight speed is 17m / s, the maximum altitude is 4m / s, and the maximum flight time is 55min. Its total working capacity is its maximum flight time of 3300 seconds, and its effective working capacity is M = 3300 * 0.8 = 2640s. For the work calculation in inspection, let α = 1 / 17 (horizontal speed) and β = 1 / 4 (vertical speed). Each task can be converted into a corresponding time, and the greater the distance and the greater the slope, the more time is required.

[0060] S133, such as Figure 4 As shown, for a UAV nest node P2, UAVs starting operations from nest node P2 can operate in two ways: to the left towards UAV nest node P1 and to the right towards UAV nest node P3. Therefore, it is necessary to ensure that the workload required for operations in both directions is less than the effective working capacity M. Based on the workload list in step S41 and the effective working capacity of the UAV in step S42, the workload list [w] is... k1 ,w n1 ,w k2 ,w n2 ,w k3 ,w n3 The task list is divided into multiple segments where the total workload is less than the effective workload M. For example, if the workload list is [12, 43, 53, 24, 73, 21, 4, 56...], and the effective workload is M = 80, then the division result is: [|12, 43|53, 24|73|21, 4|56...]. The goal is to ensure that the workload in each segment is less than M. A greedy algorithm can be used for this division.

[0061] S134, setting the UAV and the nest at the end of each paragraph, and adding two nests and a redundant UAV at the beginning and end of the line, dividing the area according to the workload, and making preliminary arrangement of the nest and UAV nodes is a simplified method, which aims to leave a margin for subsequent node adjustment after fine modeling, and in fact, the nest nodes can be arranged between the nodes, i.e. W k which can be split, and can be solved by a dynamic programming method.

[0062] S14, based on the UAV nest nodes and the number of UAVs, the flight path of the oblique photogrammetry route is planned, and the oblique photogrammetry route is obtained. The railway slope to be inspected is subjected to oblique photogrammetry according to the oblique photogrammetry route, and the real three-dimensional model is constructed based on the results of oblique photogrammetry.

[0063] Specifically, step S14 includes:

[0064] As shown in Figure 4 , Figure 4 only shows part of the route of the UAV nest nodes P1, P2 and P3, after the establishment, the oblique photogrammetry route of the UAV is generated based on the UAV nest nodes P1, P2 and P3, and the UAV performs oblique photogrammetry on the railway slope to be inspected according to the oblique photogrammetry route, to obtain the oblique photogrammetry result. The flight path is planned by the nest and the UAV arranged in step S13, to obtain the oblique photogrammetry route, wherein the entire inspection scene is subjected to oblique photogrammetry according to the oblique photogrammetry route, and a fine real three-dimensional model of the entire railway slope inspection scene is constructed. The existing method of using a UAV to inspect a slope mostly needs to first manually perform oblique photography along the railway, and then construct a real three-dimensional model based on the results of oblique photography, and plan a flight path and arrange a nest based on the real three-dimensional model, so as to complete automatic inspection of the UAV on the slope. However, manual oblique photography along the railway has a large amount of field work and consumes a high cost. The present embodiment first establishes an initial three-dimensional model to initially arrange the nest, and can realize oblique photography of the UAV along the railway without manual intervention, so as to construct a real three-dimensional model, and then adjust the position of the nest based on the real three-dimensional model, and further realize automatic inspection of the UAV. The arrangement of the UAV nest and the flight path planning of the UAV are realized by two-stage modeling, which greatly reduces the field work and reduces the overall workload, thereby saving the cost.

[0065] S2, calculate the inspection work data of the unmanned aerial vehicle based on the real three-dimensional model, the inspection work data including the inspection time of the unmanned aerial vehicle at each slope and the inspection time of the unmanned aerial vehicle between two adjacent slopes, in the embodiment, the calculation of the inspection work data of the unmanned aerial vehicle corresponding to the initial three-dimensional model has been introduced in detail in step S12, and the method for calculating the inspection work data of the unmanned aerial vehicle based on the real three-dimensional model is the same as that in step S12;

[0066] S3, divide the railway slope to be inspected at a first end point based on the inspection work data of the unmanned aerial vehicle, and construct the unmanned aerial vehicle nest node and set the number of unmanned aerial vehicles based on the first end point, in the embodiment, the position of the unmanned aerial vehicle nest node has been preliminarily planned in step S13, so after the real three-dimensional model is established, only the position of the unmanned aerial vehicle nest node needs to be adjusted, the first end point is re-divided based on the real three-dimensional model, and the number and position of the unmanned aerial vehicle nest node and the preset number of unmanned aerial vehicles are adjusted based on the second end point obtained after the re-division;

[0067] Specifically, step S3 includes:

[0068] S31, path planning based on the refined real three-dimensional model, re-calculation of the work amount corresponding to the flight paragraph obtained in step S5, and judgment of whether the re-calculated work amount exceeds the work capacity of the unmanned aerial vehicle; the length of the flight path changes only affects the calculation of the work amount, and does not affect the calculation of the work capacity M of the unmanned aerial vehicle;

[0069] S32, if there is a flight path that exceeds the work capacity of the unmanned aerial vehicle, i.e. an unqualified flight path, the flight paragraph needs to be adjusted, i.e. the position of the unmanned aerial vehicle nest node needs to be adjusted again;

[0070] Specifically, step S32 includes:

[0071] S321, first, the transfer path between the slope nodes or the flight mode of the unmanned aerial vehicle can be adjusted, for example: when the path is automatically generated, in order to simplify the algorithm logic, a rectangle is taken as the inspection object, and in actual cases, it is usually a semi-elliptical shape, so the redundant route can be adjusted to reduce the work amount required by the node;

[0072] The automatic transfer between the slope nodes is generated in a two-dimensional sequence, and the actual scene is three-dimensional, there is a large height difference between the slope nodes and the adjacent slope nodes, since the vertical flight of the unmanned aerial vehicle consumes much more than the horizontal flight, changing the sequence can reduce the total work amount required;

[0073] The unmanned aerial vehicle has multiple flight models, and the use of different sensors and algorithms, such as whether to use IMU and whether to use RTK, will also affect the flight speed;

[0074] S322, if the adjacent routes have higher redundancy, adjust the position of the nest, and redistribute the work content of the unqualified route and its adjacent routes according to the adjusted position of the nest, so that each route is qualified. For example: if the workload between the nest node A and the nest node B adjacent to the nest node A exceeds the effective working capacity of the unmanned aerial vehicle, and the workload between the nest node B and the nest node C adjacent to the nest node B has redundancy, that is, the workload is more than the effective working capacity, at this time, if the redundant workload is greater than the excess workload, the position of the nest node B can be moved, and the route is adjusted correspondingly, so that the workload between the nest nodes A and B, and the workload between the nest nodes B and C are all less than the effective working capacity of the unmanned aerial vehicle.

[0075] S323, if the redundant workload is less than the excess workload in step S622, then a new nest node D is added between the nest node A and the nest node B, and the positions of the nest node B and the nest node D are moved, so that the workload between the nest nodes is less than the effective working capacity of the unmanned aerial vehicle, and then the route is re-planned according to the adjusted nest node.

[0076] S4, based on the unmanned aerial vehicle nest node and the unmanned aerial vehicle, a route planning for inspection is performed to obtain an inspection scheme for the railway slope to be inspected;

[0077] Specifically, step S4 includes:

[0078] After adjusting the position of the nest, the unmanned aerial vehicle nest node, the slope node and the no-slope road segment between the slope nodes are as shown in Figure 5 , wherein there are m slope nodes and m no-slope road segments corresponding to the m slope nodes, w ki is the workload required for the unmanned aerial vehicle to fly on the i-th no-slope road segment, w ni is the total workload required for the unmanned aerial vehicle to fly on the i-th slope node, 1≤i≤m; there are q unmanned aerial vehicle nest nodes, P i is the i-th unmanned aerial vehicle nest node, 1≤i≤q, and one of the first and last two nests is an empty nest without unmanned aerial vehicles.

[0079] Only the transfer of the unmanned aerial vehicle between the nests is considered, as shown in Figure 6 Figure 6 ​In the figure, "has" and "no" respectively represent that the nest contains or does not contain a UAV; "no>has" represents that there is no UAV in the nest at the beginning of this inspection task, and there is a UAV after completion, and the UAV inspection process alternates according to the two transfer modes in the figure; according to the transfer of the UAV between the nests, the UAV inspection task is planned, and the embodiment adopts an alternating operation of multi-nest linkage to avoid the consumption of returning to the starting point after the UAV operation, compared with the mode that one nest is responsible for a region, effectively improving the inspection operation capability of the UAV, and also reducing the number of nests required and the consumption of cost.

[0080] Embodiment 2:

[0081] As shown in Figure 2 The embodiment provides a railway slope inspection planning device, which comprises a first processing module 901, a second processing module 902, a third processing module 903, a fourth processing module 904 and a fifth processing module 905, and specifically comprises:

[0082] The first processing module is used for acquiring oblique photogrammetry results of a railway slope to be inspected by a UAV.

[0083] The second processing module is used for constructing a real scene three-dimensional model of the railway slope to be inspected based on the oblique photogrammetry results.

[0084] The third processing module is used for calculating inspection work data of the UAV based on the real scene three-dimensional model, wherein the inspection work data comprises inspection time of the UAV at each slope and inspection time of the UAV between adjacent two slopes.

[0085] The fourth processing module is used for performing first endpoint division on the railway slope to be inspected based on the inspection work data of the UAV, and constructing UAV nest nodes based on the first endpoints obtained by the division and setting a preset number of UAVs.

[0086] The fifth processing module is used for performing inspection route planning based on the UAV nest nodes and the UAVs to obtain an inspection scheme of the railway slope to be inspected.

[0087] The first processing module comprises a first processing unit 9011, a second processing unit 9012, a third processing unit 9013, a fourth processing unit 9014 and a fifth processing unit 9015.

[0088] The first processing unit is used for acquiring first information, wherein the first information comprises satellite image data, railway surrounding terrain data and railway planning vector data of the railway slope to be inspected.

[0089] The second processing unit is configured to construct an initial three-dimensional model based on the first information, so as to obtain an initial three-dimensional model of the railway slope to be inspected.

[0090] The third processing unit is configured to calculate inspection work data corresponding to the initial three-dimensional model of the railway slope to be inspected based on the initial three-dimensional model.

[0091] The fourth processing unit is configured to perform second endpoint division on the railway slope to be inspected based on the inspection work data corresponding to the initial three-dimensional model, and construct a UAV nest node and set a preset number of UAVs based on the second endpoints obtained by the division.

[0092] The fifth processing unit is configured to perform flight path planning of a tilt photogrammetry route based on the UAV nest node and the number of UAVs, obtain the tilt photogrammetry route, perform tilt photogrammetry on the railway slope to be inspected according to the tilt photogrammetry route, and construct the real three-dimensional model based on a result of the tilt photogrammetry.

[0093] The second processing module includes a sixth processing unit 9021 and a seventh processing unit 9022.

[0094] The sixth processing unit is configured to calculate inspection time of the UAV on each slope based on a maximum flight speed of the UAV and second information of the real three-dimensional model, the maximum flight speed of the UAV including a maximum flight speed of the UAV in a horizontal direction and a maximum flight speed of the UAV in a vertical direction, and the second information including a shape, an area, a length, a width, a slope, and a cross section of each slope in the real three-dimensional model.

[0095] The seventh processing unit is configured to calculate inspection time of the UAV between two adjacent slopes based on the maximum flight speed of the UAV in the horizontal direction.

[0096] The fourth processing unit includes a first processing submodule 90141 and a second processing submodule 90142.

[0097] The first processing submodule is configured to obtain inspection work data of the UAV corresponding to the initial three-dimensional model and a maximum single flight time of the UAV.

[0098] The second processing submodule is configured to divide the railway slope to be inspected into a plurality of first paragraphs based on the maximum single flight time of the unmanned aerial vehicle and the inspection work data of the unmanned aerial vehicle corresponding to the initial three-dimensional model, to obtain the second endpoint, wherein the first paragraph is composed of a slope to be inspected and a non-slope section between adjacent slopes, a first time corresponding to the first paragraph is not more than the maximum single flight time of the unmanned aerial vehicle, the first time is a sum of an inspection time of a slope corresponding to the first paragraph and an inspection time of the unmanned aerial vehicle between two adjacent slopes, and an endpoint of each first paragraph is the second endpoint.

[0099] The third processing module includes an eighth processing unit 9031 and a ninth processing unit 9032.

[0100] The eighth processing unit is configured to calculate the inspection work data corresponding to the real-scene three-dimensional model based on third information of the real-scene three-dimensional model, to obtain a second time corresponding to the first paragraph, the second time is a sum of an inspection time of a slope corresponding to the first paragraph and an inspection time of the unmanned aerial vehicle between two adjacent slopes, calculated according to the real-scene three-dimensional model, and the third information includes a shape, an area, a length, a width, a slope, a cross section of each slope in the real-scene three-dimensional model, and a distance between adjacent slopes.

[0101] The ninth processing unit is configured to redivide the second endpoint based on the second time, wherein if the second time corresponding to the first paragraph exceeds the maximum single flight time of the unmanned aerial vehicle, the second endpoint is moved or added until the second time corresponding to each first paragraph is lower than the maximum single flight time of the unmanned aerial vehicle, to obtain the first endpoint.

[0102] The second processing unit includes a third processing submodule 90121, a fourth processing submodule 90122, a fifth processing submodule 90123, and a sixth processing submodule 90124.

[0103] The third processing submodule is configured to construct a scene three-dimensional model based on railway surrounding terrain data, the scene three-dimensional model being a three-dimensional model including two-dimensional plane data and vertical height data of a railway slope to be inspected.

[0104] The fourth processing submodule is configured to spatially register satellite image data and railway planning vector data of the railway slope to be inspected, to obtain railway region information to be inspected.

[0105] The fifth processing submodule is used to perform target recognition on the railway area information to be inspected based on the YOLOv8-seg model, and to perform pixel classification on the railway area information to be inspected after target recognition using the YOLOv8-seg model to obtain the railway slope area information to be inspected.

[0106] The sixth processing submodule is used to overlay the scene 3D model and the railway slope area information to be inspected to obtain the initial 3D model.

[0107] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0108] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0109] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A railway slope inspection planning method, characterized in that, include: Obtain the oblique photogrammetry results of the railway slope to be inspected by the UAV, and construct a real-scene 3D model of the railway slope to be inspected based on the oblique photogrammetry results; Based on the aforementioned real-scene 3D model, the inspection work data of the UAV is calculated. This inspection work data includes the inspection time of the UAV on each slope and the inspection time of the UAV between two adjacent slopes, including: The inspection time of the drone on each slope is calculated based on the maximum flight speed of the drone and the second information of the real-scene 3D model. The maximum flight speed of the drone includes the maximum flight speed of the drone in the horizontal direction and the maximum flight speed in the vertical direction. The second information includes the shape, area, length, width, slope and cross-section of each slope in the real-scene 3D model. The inspection time of the UAV between two adjacent slopes is calculated based on the maximum flight speed of the UAV in the horizontal direction. The inspection time is obtained by calculating the workload required for the UAV to fly. The workload required for the flight is calculated by fitting a rectangle to the slope area, a triangle to represent the slope shape, and the length l of the circumscribed rectangle parallel to the railway and the width s perpendicular to the railway. p is the slope of the slope. The total workload of a drone on a single slope is calculated using the following formula: Among them, W n Let be the total workload of the UAV on a single slope, n be the distance of each flight in the direction perpendicular to the railway, α be the reciprocal of the maximum flight speed of the UAV in the horizontal direction, β be the reciprocal of the maximum flight speed of the UAV in the vertical direction, p be the slope of the slope, and s be the width of the circumscribed rectangle of the slope perpendicular to the railway. Based on the inspection work data of the UAV, the first endpoint is divided into the railway slope to be inspected, and UAV nest nodes are constructed and the number of UAVs is set based on the first endpoint. Each inspection time in the inspection work data is used as the workload of the UAV. The workload between any two adjacent UAV nest nodes is less than the effective working capacity of the UAV. The effective working capacity refers to the remaining part of the UAV's working capacity reserved to deal with unexpected events. Based on the drone nest nodes and the number of drones, an inspection route is planned to obtain an inspection plan for the railway slope to be inspected.

2. The railway slope inspection planning method according to claim 1, characterized in that... The construction of a realistic 3D model of the railway slope to be inspected based on the oblique photogrammetry results includes: Obtain first information, which includes satellite imagery data of the railway slope to be inspected, terrain data of the railway surrounding area, and railway planning vector data; Based on the first information, an initial three-dimensional model is constructed to obtain the initial three-dimensional model of the railway slope to be inspected. Based on the initial three-dimensional model of the railway slope to be inspected, calculate the inspection work data corresponding to the initial three-dimensional model; Based on the inspection data of the UAV corresponding to the initial three-dimensional model, the railway slope to be inspected is divided into second endpoints, and UAV nest nodes are constructed and the number of UAVs is set based on the divided second endpoints. Based on the UAV nest node and the number of UAVs, the trajectory planning of the oblique photogrammetry route is carried out to obtain the oblique photogrammetry route. Oblique photogrammetry is carried out on the railway slope to be inspected according to the oblique photogrammetry route, and the real scene 3D model is constructed based on the oblique photogrammetry results.

3. The railway slope inspection planning method according to claim 2, characterized in that... The second endpoint division of the railway slope to be inspected based on the inspection data of the UAV corresponding to the initial three-dimensional model includes: Obtain the inspection work data of the UAV corresponding to the initial 3D model and the maximum single flight time of the UAV; Based on the maximum single flight time of the UAV and the inspection data of the UAV corresponding to the initial three-dimensional model, the railway slope to be inspected is divided into multiple first segments to obtain the second endpoint. The first segment consists of the slope of the railway slope to be inspected and the slopeless section between adjacent slopes. The first time corresponding to the first segment does not exceed the maximum single flight time of the UAV. The first time is the sum of the inspection time of the slope corresponding to the first segment and the inspection time of the UAV between two adjacent slopes. The endpoint of each first segment is the second endpoint.

4. The railway slope inspection planning method according to claim 3, characterized in that... The inspection data based on drones is used to divide the railway slope to be inspected into its first endpoints, including: Based on the third information of the real-scene 3D model, the inspection work data corresponding to the real-scene 3D model is calculated to obtain the second time corresponding to the first segment. The second time is the sum of the inspection time of the slope corresponding to the first segment calculated according to the real-scene 3D model and the inspection time of the UAV between two adjacent slopes. The third information includes the shape, area, length, width, slope, cross-section and distance between adjacent slopes of each slope in the real-scene 3D model. The second endpoint is re-divided based on the second time. If the second time corresponding to the first segment exceeds the maximum single flight time of the drone, the second endpoint is moved or added until the second time corresponding to each first segment is lower than the maximum single flight time of the drone, thus obtaining the first endpoint.

5. A railway slope inspection and planning device, characterized in that, include: The first processing module is used to acquire the oblique photogrammetry results of the railway slope to be inspected by the UAV; The second processing module is used to construct a real-scene 3D model of the railway slope to be inspected based on the oblique photogrammetry results. The third processing module is used to calculate the inspection work data of the UAV based on the real-scene 3D model. The inspection work data includes the inspection time of the UAV on each slope and the inspection time of the UAV between two adjacent slopes, including: The sixth processing unit is used to calculate the inspection time of the UAV on each slope based on the maximum flight speed of the UAV and the second information of the real-scene 3D model. The maximum flight speed of the UAV includes the maximum flight speed of the UAV in the horizontal direction and the maximum flight speed in the vertical direction. The second information includes the shape, area, length, width, slope and cross-section of each slope in the real-scene 3D model. The seventh processing unit is used to calculate the inspection time of the UAV between two adjacent slopes based on the maximum flight speed of the UAV in the horizontal direction. The inspection time is obtained by calculating the workload required for the UAV to fly. The workload required for the flight is calculated by fitting a rectangle to the slope area, a triangle to the slope shape, and the length l of the circumscribed rectangle parallel to the railway and the width s perpendicular to the railway, where p is the slope of the slope. The total workload of a drone on a single slope is calculated using the following formula: Among them, W n Let be the total workload of the UAV on a single slope, n be the distance of each flight in the direction perpendicular to the railway, α be the reciprocal of the maximum flight speed of the UAV in the horizontal direction, β be the reciprocal of the maximum flight speed of the UAV in the vertical direction, p be the slope of the slope, and s be the width of the circumscribed rectangle of the slope perpendicular to the railway. The fourth processing module is used to divide the railway slope to be inspected into first endpoints based on the inspection work data of the UAV, and to construct UAV nest nodes and set the number of UAVs based on the first endpoints. Each inspection time in the inspection work data is used as the workload of the UAV. The workload between any two adjacent UAV nest nodes is less than the effective working capacity of the UAV. The effective working capacity refers to the portion of the UAV's working capacity reserved to cope with the remaining portion except for unexpected events. The fifth processing module is used to plan the inspection route based on the UAV nest node and the number of UAVs, and obtain the inspection plan for the railway slope to be inspected.

6. A railway slope inspection and planning device according to claim 5, characterized in that, The first processing module includes: The first processing unit is used to acquire first information, which includes satellite image data of the railway slope to be inspected, terrain data around the railway, and railway planning vector data. The second processing unit is used to construct an initial three-dimensional model based on the first information to obtain the initial three-dimensional model of the railway slope to be inspected. The third processing unit is used to calculate the inspection work data corresponding to the initial three-dimensional model based on the initial three-dimensional model of the railway slope to be inspected. The fourth processing unit is used to divide the railway slope to be inspected into second endpoints based on the inspection work data corresponding to the initial three-dimensional model, and to construct UAV nest nodes and set the number of UAVs based on the divided second endpoints. The fifth processing unit is used to plan the trajectory of the oblique photogrammetry route based on the UAV nest node and the number of UAVs, obtain the oblique photogrammetry route, perform oblique photogrammetry on the railway slope to be inspected according to the oblique photogrammetry route, and construct the real-scene 3D model based on the oblique photogrammetry results.

7. A railway slope inspection and planning device according to claim 6, characterized in that, The fourth processing unit includes: The first processing submodule is used to obtain the inspection work data of the UAV corresponding to the initial three-dimensional model and the maximum single flight time of the UAV. The second processing submodule is used to divide the railway slope to be inspected into multiple first segments based on the maximum single flight time of the UAV and the inspection work data of the UAV corresponding to the initial three-dimensional model, and obtain the second endpoint. The first segment consists of the slope of the railway slope to be inspected and the slopeless section between adjacent slopes. The first time corresponding to the first segment does not exceed the maximum single flight time of the UAV. The first time is the sum of the inspection time of the slope corresponding to the first segment and the inspection time of the UAV between two adjacent slopes. The endpoint of each first segment is the second endpoint.

8. A railway slope inspection and planning device according to claim 7, characterized in that, The third processing module includes: The eighth processing unit is used to calculate the inspection work data corresponding to the real scene 3D model based on the third information of the real scene 3D model, and obtain the second time corresponding to the first segment. The second time is the sum of the inspection time of the slope corresponding to the first segment calculated according to the real scene 3D model and the inspection time of the UAV between two adjacent slopes. The third information includes the shape, area, length, width, slope, cross-section and distance between adjacent slopes of each slope in the real scene 3D model. The ninth processing unit is used to re-divide the second endpoint based on the second time, wherein if the second time corresponding to the first segment exceeds the maximum single flight time of the UAV, the second endpoint is moved or added until the second time corresponding to each first segment is lower than the maximum single flight time of the UAV, thus obtaining the first endpoint.

Images