A method and system for automatic road design for new energy wind farms

Abstract

The present application belongs to the technical field of road design for new energy wind farms, and discloses a method and system for automatic road design for new energy wind farms. By superimposing a model-predicted feature distribution image with a spatial planning image, all feature types planned for the wind farm are identified; the feature distribution image is divided into regions according to preset feature attributes, and the nodes on the cells are assigned values ​​by splitting cells to determine the feasibility of each node in road design; the segment path cost between each node and the adjacent nodes is calculated, and the position of the wind turbine in the road distribution network, all segment path costs corresponding to the node, and the assignment results are processed in combination with the path optimization algorithm to generate the optimal planned road for the wind turbine, which not only avoids errors caused by human subjective factors, but also effectively ensures the efficiency and accuracy of road design on the basis of reducing resource waste and energy consumption, thereby meeting road design requirements.

Description

Technical Field

[0001] The present application belongs to the technical field of road design for new energy wind farms, and in particular, relates to an automatic road design method and system for new energy wind farms. Background Art

[0002] As new energy gradually penetrates into various fields of people's work and life, new energy infrastructure represented by wind farms has ushered in a new wave of growth. With the rapid development of the wind farm construction industry, people have an increasingly strong demand for the precise design of wind farm operation and inspection roads.

[0003] Currently, wind farms are mainly located in remote areas or mountainous areas with complex terrain conditions. Traditional wind farm operation and maintenance roads are mainly designed manually, which are not only easily affected by design experience and subjective factors, but also prone to errors or unreasonable designs, and thus cannot meet road design requirements. Summary of the invention

[0004] In order to solve the above-mentioned technical problems that the traditional wind farm inspection roads are mainly designed manually, which are not only easily affected by design experience and subjective factors, but also prone to errors or unreasonable designs, and thus cannot meet the road design requirements, this application proposes a road automatic design method and system for new energy wind farms, and its technical solution is as follows:

[0005] In a first aspect, an embodiment of the present application provides a method for automatic road design for a new energy wind farm, comprising:

[0006] Acquire digital images of objects in the wind farm, and input the digital images into a deep learning model to obtain an object distribution image; wherein the deep learning model is trained by at least two sample images marked with object distribution labels;

[0007] Overlaying the object distribution image and the spatial planning image corresponding to the wind farm, and performing regional division processing on the overlayed object distribution image based on preset object attributes;

[0008] The object distribution image after the area division process is divided into at least two cells, and each node on the corresponding cell is assigned a value according to the area to which each cell belongs; wherein each cell corresponds to four nodes;

[0009] Calculate the segment path cost between each node and at least four adjacent nodes, and generate the planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node and the assignment results;

[0010] The planned roads of all wind turbines are added to the road distribution network, and the road distribution network including all wind turbines and all planned roads is displayed to the user.

[0011] In an optional solution of the first aspect, superimposing the ground feature distribution image and the spatial planning image corresponding to the wind farm includes:

[0012] Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour;

[0013] Based on the area ratio between any first edge contour and the corresponding second edge contour, the space planning image is scaled, and the object distribution image and the scaled space planning image are superimposed so that each first edge contour coincides with the corresponding second edge contour.

[0014] In another optional solution of the first aspect, based on preset ground object attributes, the ground object distribution image after the superposition processing is subjected to region division processing, including:

[0015] In the object distribution image after the overlay processing, the object type consistent with the preset object attribute and the object area corresponding to the object type are identified;

[0016] Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold;

[0017] In the overlay processed feature distribution image, the area excluding all buffer areas is divided into the utilization area.

[0018] In yet another optional solution of the first aspect, assigning a value to each node on the corresponding cell according to the region to which each cell belongs includes:

[0019] When the area to which any cell belongs is a buffer area, a first value assignment process is performed on each node on the corresponding cell;

[0020] When the area to which any cell belongs is a utilization area, the second value assignment process is performed on each node on the corresponding cell.

[0021] In another optional solution of the first aspect, calculating the segment path cost between each node and at least four adjacent nodes includes:

[0022] Determine at least four adjacent nodes corresponding to each node on all cells, and calculate the distance cost according to the path length between each node and each corresponding adjacent node;

[0023] Obtain the elevation of each node and each corresponding adjacent node on the object distribution image, and intercept the polyline segment between the elevation of each node and the elevation of each corresponding adjacent node in the preset elevation-mileage polyline;

[0024] The corresponding earthwork cost is calculated according to each broken line segment, and the distance cost and earthwork cost corresponding to each node and each adjacent node are weighted and summed to obtain the segment path cost.

[0025] In another optional solution of the first aspect, generating a planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node, and the assignment result, includes:

[0026] Identify the existing roads and the location of each wind turbine in the road distribution network of the wind farm, and map the existing roads and the location of each wind turbine on all cells;

[0027] According to the vertical path between each wind turbine and the existing road, the starting point and the end point are determined on all cells after mapping. Based on the path optimization algorithm, the starting point and the end point corresponding to each wind turbine, the path costs of all segments corresponding to each node, and the assignment results are processed to obtain the planned road for the corresponding wind turbine.

[0028] In another optional solution of the first aspect, before inputting the digital image of the ground object into the deep learning model to obtain the ground object distribution image, it also includes:

[0029] Preprocessing the digital image of the ground object and calculating the signal-to-noise ratio of the preprocessed digital image of the ground object;

[0030] Input the digital image of the ground object into the deep learning model to obtain the ground object distribution image, including:

[0031] When the signal-to-noise ratio exceeds the preset signal-to-noise ratio threshold, the preprocessed digital image of the ground object is input into the deep learning model to obtain the ground object distribution image.

[0032] In a second aspect, an embodiment of the present application provides an automatic road design system for a new energy wind farm, comprising:

[0033] A model output module is used to obtain digital images of ground objects in the wind farm, and input the digital images of the ground objects into a deep learning model to obtain a ground object distribution image; wherein the deep learning model is trained by at least two sample images marked with ground object distribution labels;

[0034] A region division module is used to perform superposition processing on the ground object distribution image and the spatial planning image corresponding to the wind farm, and perform region division processing on the superimposed ground object distribution image based on preset ground object attributes;

[0035] A node assignment module is used to divide the surface feature distribution image after the area division process into at least two cells, and assign a value to each node on the corresponding cell according to the area to which each cell belongs; wherein each cell corresponds to four nodes;

[0036] A path generation module is used to calculate the segment path cost between each node and at least four adjacent nodes, and generate a planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node and the assignment results;

[0037] The road display module is used to add the planned roads of all wind turbines to the road distribution network and display the road distribution network including all wind turbines and all planned roads to the user.

[0038] In a third aspect, an embodiment of the present application further provides a road automatic design system for a new energy wind farm, including a processor and a memory;

[0039] The processor is connected to the memory;

[0040] A memory for storing executable program codes;

[0041] The processor runs the program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the automatic road design method for a new energy wind farm provided by the first aspect of the embodiment of the present application or any one of the implementation methods of the first aspect.

[0042] In a fourth aspect, an embodiment of the present application provides a computer storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the automatic road design method for a new energy wind farm provided by the first aspect of the embodiment of the present application or any one of the implementation methods of the first aspect can be implemented.

[0043] In an embodiment of the present application, when designing an operation and inspection road for a new energy wind farm, all land feature types planned for the wind farm can be accurately identified by superimposing the land feature distribution image predicted by the model with the spatial planning image; then the processed land feature distribution image can be divided into regions according to preset land feature attributes, and the nodes on each cell can be assigned values ​​by splitting cells to determine the feasibility of each node during road design; then the segment path cost between each node and multiple adjacent nodes can be calculated, and the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node, and the assignment results can be processed in combination with the path optimization algorithm to generate the optimal planned road for the corresponding wind turbine, which can not only avoid errors caused by human subjective factors, but also effectively ensure the efficiency and accuracy of road design on the basis of reducing resource waste and energy consumption, thereby meeting road design requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 An overall flow chart of an automatic road design method for a new energy wind farm provided in an embodiment of the present application;

[0046] Figure 2 A schematic diagram of the effect of adjacent nodes provided in an embodiment of the present application;

[0047] Figure 3 A schematic diagram of the effect of a preset elevation-mileage polyline provided in an embodiment of the present application;

[0048] Figure 4 A schematic diagram of the structure of an automatic road design system for a new energy wind farm provided in an embodiment of the present application;

[0049] Figure 5 A schematic diagram of the structure of another automatic road design system for a new energy wind farm provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0051] In the following introduction, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The following introduction provides multiple embodiments of the present application, and different embodiments can be replaced or combined, so the present application can also be considered to include all possible combinations of the same and / or different embodiments recorded. Therefore, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then the present application should also be considered to include embodiments containing one or more of all other possible combinations of A, B, C, and D, although the embodiment may not be clearly recorded in the following text.

[0052] The following description provides examples and does not limit the scope, applicability or examples set forth in the claims. Changes may be made to the functions and arrangements of the elements described without departing from the scope of the present application. Various processes or components may be appropriately omitted, substituted or added to each example. For example, the described method may be performed in an order different from the order described, and various steps may be added, omitted or combined. In addition, features described in some examples may be combined in other examples.

[0053] See also Figure 1 , Figure 1 An overall flow chart of an automatic road design method for a new energy wind farm provided in an embodiment of the present application is shown.

[0054] like Figure 1 As shown, the automatic road design method for a new energy wind farm may include at least the following steps:

[0055] Step 102: Acquire digital images of objects in the wind farm, and input the digital images into a deep learning model to obtain an image of object distribution.

[0056] In an embodiment of the present application, the automatic road design method for a new energy wind farm can be but is not limited to being applied to a control terminal. The control terminal can pre-store data such as a spatial planning image corresponding to the new energy wind farm, a road distribution network, and the set positions of all wind turbines in the road distribution network, and can also establish a connection with an aerial surveying device to receive digital images of ground objects collected by the aerial surveying device above the new energy wind farm.

[0057] Among them, the spatial planning image can be understood as the spatial planning image of the area where the new energy wind farm is located after the land feature planning (it can also be understood as the three-adjustment map), which can be but not limited to downloading through a designated website or obtaining from a designated agency. The spatial planning image may include one or more built-up land feature areas, and one or more planned land feature areas to be built; the road distribution network can be understood as a road distribution network composed of all existing roads in the area where the new energy wind farm is located, which can be but not limited to existing road types such as main roads, secondary roads, branches, expressways and highways, and technical personnel can also mark the setting positions of all wind turbines in the road distribution network.

[0058] Of course, the road distribution network mentioned in the embodiment of the present application can also be but not limited to a convolutional neural network or other neural network to identify and process the road image (or the spatial planning image mentioned above) of the area where the new energy wind farm is located to obtain a road distribution image marked with all existing roads, and is not limited to this.

[0059] Here, the aerial surveying equipment may be but is not limited to a drone equipped with a laser radar, and the digital images of the ground objects collected by the aerial surveying equipment may include but are not limited to all the built ground objects in the new energy wind farm, and the digital elevation corresponding to each ground object (that is, elevation, which refers to the altitude of any point on the surface of the ground object compared to the horizontal plane).

[0060] It can be understood that the control terminal can accurately identify all land feature types planned in the wind farm by superimposing the land feature distribution image predicted by the model with the spatial planning image; then the processed land feature distribution image can be divided into regions according to the preset land feature attributes, and the nodes on each cell can be assigned values ​​by splitting the cells to determine the feasibility of each node in road design; then the segment path cost between each node and multiple adjacent nodes can be calculated, and the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node, and the assignment results can be processed in combination with the path optimization algorithm to generate the optimal planned road for the corresponding wind turbine, which can not only avoid errors caused by human subjective factors, but also effectively ensure the efficiency and accuracy of road design on the basis of reducing resource waste and energy consumption, thereby meeting road design requirements.

[0061] Specifically, when designing the inspection road for the new energy wind farm, the control terminal can, but is not limited to, obtain the digital image of the ground objects in the new energy wind farm through the above-mentioned aerial survey equipment. The digital image of the ground object can include all the ground objects in the area where the new energy wind farm is located and the digital elevation corresponding to each ground object. The ground object types here can be, but are not limited to, trees, waterways, roads, houses, and cultivated land, etc., and the digital image of the ground object is input into the trained deep learning model to obtain a more accurate ground object distribution image by using the model prediction method. Here, the structure of the deep learning model can be, but is not limited to, referring to the convolutional neural network structure known in the art, which can be trained by a plurality of sample images marked with ground object distribution labels. The sample images can be collected by aerial survey equipment over wind farms in different regions. The ground object distribution labels on the sample images can be marked by technicians, and different ground object types correspond to different marking methods, which are not limited to this.

[0062] It can be understood that the feature distribution image can display the area where the wind farm is located, as well as the feature areas corresponding to each feature type in the area where the wind farm is located. Each feature type corresponds to one or more feature areas, and the feature areas corresponding to different feature types can be classified by texture features, edge features or color features. For example, but not limited to, all feature areas corresponding to trees can be filled with green, and all feature areas corresponding to houses can be filled with gray; for another example, but not limited to, all feature areas corresponding to trees use point texture, all feature areas corresponding to waterways use line texture, and all feature areas corresponding to houses use surface texture.

[0063] As an option of the embodiment of the present application, before inputting the digital image of the ground object into the deep learning model to obtain the ground object distribution image, it also includes:

[0064] Preprocessing the digital image of the ground object and calculating the signal-to-noise ratio of the preprocessed digital image of the ground object;

[0065] Input the digital image of the ground object into the deep learning model to obtain the ground object distribution image, including:

[0066] When the signal-to-noise ratio exceeds the preset signal-to-noise ratio threshold, the preprocessed digital image of the ground object is input into the deep learning model to obtain the ground object distribution image.

[0067] In order to ensure the quality of the digital image of the ground object and the accuracy of the ground object distribution image output by the deep learning model, the digital image of the ground object can also be preprocessed, and it can be determined whether the quality of the preprocessed digital image of the ground object meets the requirements.

[0068] Specifically, after acquiring the digital image of the ground object of the new energy wind farm, the control terminal can also pre-process the digital image of the ground object, and the pre-processing method can be but not limited to any one of filtering processing, equalization processing, color space conversion processing or normalization processing, and calculate the signal-to-noise ratio of the pre-processed digital image of the ground object based on a preset image processing algorithm. Here, the preset image processing algorithm can be a conventional technical means in the art, and will not be described in detail here.

[0069] It can be understood that when it is detected that the signal-to-noise ratio exceeds the preset signal-to-noise ratio threshold, it indicates that the quality of the preprocessed digital image of the ground object meets the requirements, and the preprocessed digital image of the ground object can be input into the deep learning model to obtain a more accurate ground object distribution image through model prediction.

[0070] When it is detected that the signal-to-noise ratio does not exceed the preset signal-to-noise ratio threshold, it indicates that the quality of the preprocessed digital image of the ground object is low, and then the digital image of the ground object of the new energy wind farm can be obtained again from other angles through but not limited to aerial measurement equipment, and the quality of the re-acquired digital image of the ground object can be judged by calculating the signal-to-noise ratio.

[0071] Of course, in the embodiment of the present application, other methods may be used to determine whether the quality of the pre-processed digital image of the ground object meets the requirements, which is not limited to the signal-to-noise ratio mentioned above.

[0072] Step 104: Overlay the object distribution image and the space planning image corresponding to the wind farm, and perform region division processing on the overlayed object distribution image based on preset object attributes.

[0073] Specifically, after obtaining the feature distribution image through the deep learning model, the control terminal can, but is not limited to, overlay the feature distribution image with the spatial planning image corresponding to the wind farm, so that the feature distribution image after the overlay processing includes all built feature areas and all feature areas planned to be built, thereby ensuring the effectiveness and reliability of subsequent planned roads, and can also perform regional division processing on the feature distribution image after the overlay processing in combination with preset feature attributes to divide the usable areas for road planning and the unusable areas for road planning.

[0074] Here, the features of landforms can be divided into those that need to be avoided during road planning and those that do not need to be avoided during road planning. The features types corresponding to the feature attribute that needs to be avoided during road planning may include but are not limited to houses and waterways, and the features types corresponding to the feature attribute that does not need to be avoided during road planning may include but are not limited to trees and cultivated land. The preset feature attribute mentioned above may be the feature that needs to be avoided during road planning, that is, on the feature distribution image after superposition processing, all feature types corresponding to the feature attribute that needs to be avoided during road planning can be divided into regions, and the areas divided according to the preset feature attributes are used as unusable areas, and the remaining areas not divided are used as usable areas.

[0075] As another option of the embodiment of the present application, superimposing the ground feature distribution image and the spatial planning image corresponding to the wind farm includes:

[0076] Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour;

[0077] Based on the area ratio between any first edge contour and the corresponding second edge contour, the space planning image is scaled, and the object distribution image and the scaled space planning image are superimposed so that each first edge contour coincides with the corresponding second edge contour.

[0078] In order to ensure the superposition effect between the feature distribution image and the space planning image, two edge contours corresponding to the same feature type can be identified in the feature distribution image and the space planning image respectively, so as to perform image superposition based on the two edge contours.

[0079] Specifically, when the feature distribution image and the spatial planning image are superimposed, the control terminal may also first determine at least one feature type with a higher degree of recognition in the feature distribution image (for example, but not limited to any one of houses, trees or cultivated land), and identify the first edge contour of each feature type in the feature distribution image. The first edge contour may be the edge contour corresponding to a partial feature area of ​​each feature type in the feature distribution image, or the edge contour corresponding to all feature areas of each feature type in the feature distribution image. Preferably, the edge contour corresponding to a regular-shaped feature area of ​​each feature type in the feature distribution image may be used as the first edge contour. For example, but not limited to, a rectangular house area in the feature distribution image may be used as the first edge contour.

[0080] Then, after identifying the first edge contour, the control terminal can also identify the second edge contour of each land feature type in the spatial planning image corresponding to the wind farm based on at least one land feature type with a higher degree of recognition, and the shape of the second edge contour of each land feature type can be consistent with the shape of the first edge contour of the corresponding land feature type. For example, but not limited to, when a house area with a rectangular shape is used as the first edge contour in the land feature distribution image, a house area with the same rectangular shape can be used as the second edge contour in the spatial planning image.

[0081] Then, after identifying the second edge contour, the control terminal can also calculate the area of ​​any first edge contour with a regular shape, and according to the type of land feature corresponding to the first edge contour with a regular shape, calculate the area of ​​the second edge contour of the land feature type, and calculate the ratio of the area of ​​the first edge contour to the area of ​​the second edge contour to obtain the area ratio.

[0082] Next, when it is detected that the area ratio is inconsistent with the preset area ratio threshold, it indicates that there is a difference in the size of the land feature distribution image and the space planning image, and the control terminal can then scale the space planning image according to the area ratio. For example, when the area ratio is less than the preset area ratio threshold, the space planning image can be reduced, and the reduction ratio can be but not limited to the area ratio; for example, when the area ratio is greater than the preset area ratio threshold, the space planning image can be enlarged, and the enlargement ratio can be but not limited to the area ratio, and the land feature distribution image and the scaled space planning image are spatially superimposed, that is, the scaled space planning image is superimposed on the land feature distribution image, so that the first edge contour of each land feature type completely overlaps with the corresponding second edge contour.

[0083] As another option of the embodiment of the present application, based on preset ground object attributes, the ground object distribution image after the superposition processing is divided into regions, including:

[0084] In the object distribution image after the overlay processing, the object type consistent with the preset object attribute and the object area corresponding to the object type are identified;

[0085] Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold;

[0086] In the overlay processed feature distribution image, the area excluding all buffer areas is divided into the utilization area.

[0087] Specifically, taking the preset feature attribute as the above-mentioned road planning needing to be avoided as an example, when performing area division processing on the feature distribution image after superimposition processing, the control terminal can also identify all feature types such as houses and waterways that need to be avoided in road planning in the feature distribution image after superimposition processing, as well as all feature areas corresponding to all feature types, and mark feature points in each feature area. Here, the feature point in each feature area can be the center point corresponding to the feature area, but is not limited thereto.

[0088] Next, after marking the feature points in each feature area, the control terminal can also divide the corresponding buffer area in the superimposed feature distribution image with each feature point as the center and a preset distance threshold as the radius, that is, the unusable area mentioned above, and use the area in the superimposed feature distribution image minus all buffer areas as the usable area, that is, the area for road planning.

[0089] Of course, in the embodiment of the present application, it is also possible but not limited to combining with user planning needs, identifying the edge contour of the area where the wind farm is located in the overlaid terrain distribution image, and dividing the designated position within the edge contour of the area where the wind farm is located into a buffer area, that is, the unusable area mentioned above, and is not limited to this.

[0090] Step 106: Segment the object distribution image after the area division process into at least two cells, and assign a value to each node on the corresponding cell according to the area to which each cell belongs.

[0091] Specifically, after the area division processing is performed on the feature distribution image, the control terminal can determine the area of ​​each cell according to the area of ​​the feature distribution image, but is not limited to, so as to divide the feature distribution image after the area division processing into a plurality of cells distributed in an array, the area of ​​each cell is consistent, and the four nodes on the corresponding cell can be assigned values ​​according to the area to which each cell belongs on the feature distribution image. For example, when the area to which a certain cell belongs on the feature distribution image is a buffer area (i.e., an unusable area), the four nodes on the cell can be assigned values ​​respectively, and the result of the assignment of each node can be, but is not limited to, +∞; for another example, when the area to which a certain cell belongs on the feature distribution image is a usable area, the four nodes on the cell can be assigned values ​​respectively, and the result of the assignment of each node can be, but is not limited to, 0.

[0092] Here, the shape of each cell can be but is not limited to a rectangle, and its four nodes can correspond to the four vertices of the rectangle. Of course, in the embodiment of the present application, each cell can also be but is not limited to being assigned a value. For example, when the area to which a cell belongs on the terrain distribution image is a buffer area, the first assignment process can be performed on the cell, and the assignment result of the cell can be +∞; for example, when the area to which a cell belongs on the terrain distribution image is a utilization area, the second assignment process can be performed on the cell, and the assignment result of the cell can be 0, and is not limited to this.

[0093] It should be noted that in the embodiment of the present application, the terrain distribution image after area division processing can be divided into a larger number of cells as much as possible, that is, the area of ​​each cell is designed to be smaller, so that the starting point and end point on the subsequent planned road can be nodes on any cell, thereby improving the efficiency of ensuring the generation of subsequent planned roads.

[0094] Step 108, calculate the segment path cost between each node and at least four adjacent nodes, and generate the planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node and the assignment result.

[0095] Specifically, after assigning values ​​to the nodes on each cell, the control terminal can, but is not limited to, determine at least four adjacent nodes on the multiple adjacent cells corresponding to each node according to a preset direction angle, and calculate the segment path cost between each node and each corresponding adjacent node, so as to use the segment path cost between each node and each corresponding adjacent node as a generation indicator for the planned road, for example, but not limited to, the road with the lowest total segment path cost can be used as the planned road.

[0096] See here Figure 2 The schematic diagram of the effect of an adjacent node provided by an embodiment of the present application is shown as follows: Figure 2 2a shows the effect of each node corresponding to the adjacent nodes when the preset direction angle is 90 degrees. It can be seen that each node (P in 2a) ij ) corresponds to four adjacent cells and four adjacent nodes; 2b shows a schematic diagram of the effect of each node corresponding to the adjacent nodes when the preset direction angle is 45 degrees. It can be seen that each node (P in 2b) ij ) corresponds to four adjacent cells and eight adjacent nodes; 2c shows a schematic diagram of the effect of each node corresponding to the adjacent nodes when the preset direction angle is 26.5 degrees. It can be seen that each node (P in 2c) ij) corresponds to twelve adjacent cells and sixteen adjacent nodes, and the number of adjacent nodes determined by each node in the present application embodiment can be sixteen to ensure the efficiency of generating the planned road. Of course, the preset direction angle in the application embodiment can also be set to other angles, not limited to Figure 2 90 degrees, 45 degrees and 26.5 degrees are shown.

[0097] As another option of the embodiment of the present application, calculating the segment path cost between each node and at least four adjacent nodes includes:

[0098] Determine at least four adjacent nodes corresponding to each node on all cells, and calculate the distance cost according to the path length between each node and each corresponding adjacent node;

[0099] Obtain the elevation of each node and each corresponding adjacent node on the object distribution image, and intercept the polyline segment between the elevation of each node and the elevation of each corresponding adjacent node in the preset elevation-mileage polyline;

[0100] The corresponding earthwork cost is calculated according to each broken line segment, and the distance cost and earthwork cost corresponding to each node and each adjacent node are weighted and summed to obtain the segment path cost.

[0101] In order to ensure that the segment path cost between each node and each corresponding adjacent node takes into account the actual road generation requirements, the distance cost and earthwork cost between each node and each corresponding adjacent node can be weightedly summed to obtain the corresponding segment path cost from the two cost perspectives of distance and earthwork.

[0102] Specifically, when calculating the segment path cost between each node on all cells and each corresponding adjacent node, the control terminal can also determine at least four adjacent nodes on multiple adjacent cells corresponding to each node according to the preset direction angle, so as to obtain the path length between each node and each corresponding adjacent node through the image measurement tool, and obtain the corresponding actual path distance by multiplying the actual scaling ratio corresponding to the object distribution image with each path length. Here, the actual scaling ratio corresponding to the object distribution image can be understood as the ratio of the actual straight-line distance between any two locations in the area where the wind farm is located to the straight-line distance between the corresponding two locations on the object distribution image, and in the embodiment of the present application, the actual scaling ratio corresponding to the object distribution image can be derived by a technician, and is not limited to this.

[0103] Then, after obtaining the actual path distance between each node and each corresponding adjacent node, the control terminal can also query the distance cost corresponding to each actual path distance in the preset distance-cost corresponding list. It is understandable that the preset distance-cost corresponding list may include at least two path distances, and the distance cost of each path distance obtained by the technician's analysis or the preset distance cost algorithm. Of course, in the embodiment of the present application, the distance cost corresponding to each actual path distance can also be calculated in combination with the preset distance-cost ratio, which is not limited to this.

[0104] Next, when calculating the earthwork cost between each node and each corresponding adjacent node, the control terminal can also obtain the elevation corresponding to each node in the above-mentioned digital image of the ground feature, so as to determine the fold line points corresponding to the two elevations in the preset elevation-mileage fold line according to the elevation corresponding to each node and the elevation corresponding to each corresponding adjacent node, and intercept the fold line segment between the two fold line points. Here, the preset elevation-mileage fold line can be obtained by technicians based on the terrain, altitude and road flatness analysis of the area where the wind farm is located, which can be composed of connecting lines between multiple fold line points, the horizontal coordinate of each fold line point can be but not limited to mileage, and the vertical coordinate of each fold line point can be but not limited to elevation.

[0105] Also see here Figure 3 The schematic diagram of the effect of a preset elevation-mileage polyline provided in an embodiment of the present application is shown as follows: Figure 3 The broken line shown is the preset elevation-mileage broken line, which may include but is not limited to nine broken line points, the horizontal coordinate of each broken line point corresponds to the mileage (in meters), the vertical coordinate of each broken line point corresponds to the elevation (in meters), and the preset elevation-mileage broken line is obtained by connecting two adjacent broken line points in sequence.

[0106] It can be understood that when determining the broken line points corresponding to two elevations respectively in the preset elevation-mileage broken line, taking the two elevations being the first elevation and the second elevation as an example, when the number of broken line points corresponding to the first elevation is greater than one, all the first miles corresponding to the first elevation can be first identified in the preset elevation-mileage broken line, and the difference between the second miles corresponding to the second elevation and each first miles can be calculated, so as to use the broken line point corresponding to the difference consistent with the corresponding actual path distance as the broken line point corresponding to the first elevation.

[0107] Next, after cutting out the broken line segment between the two broken line points, the control terminal can also sum up the lengths of all straight line segments corresponding to the broken line segment, and obtain the earthwork cost between each node and each corresponding adjacent node by multiplying the summation result with the preset earthwork unit cost. Here, the preset earthwork unit cost can be analyzed by the technician considering factors such as construction consumption cost. Of course, in the embodiment of the present application, the earthwork cost between each node and each corresponding adjacent node can also be obtained by other earthwork calculation methods, which is not limited to this.

[0108] Then, after obtaining the distance cost and earthwork cost corresponding to each node and each corresponding adjacent node, the control terminal can also perform weighted summation processing on the distance cost and earthwork cost according to the preset weight value to obtain the corresponding segment path cost. It can be understood that when the planning road requires the minimum path cost, the preset weight value can be adjusted, but not limited to, to increase the weight value corresponding to the distance cost and reduce the weight value corresponding to the earthwork cost; when the planning road requires the minimum earthwork cost, the preset weight value can be adjusted, but not limited to, to increase the weight value corresponding to the earthwork cost and reduce the weight value corresponding to the distance cost.

[0109] Furthermore, after calculating the segment path cost between each node and each corresponding adjacent node, in order to ensure the generation of the optimal planned road for each wind turbine, the control terminal may, but is not limited to, use a path optimization algorithm to process the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node, and the assignment results to generate the planned road for the corresponding wind turbine. The path optimization algorithm may, but is not limited to, be an A* algorithm, which comprehensively considers all nodes traversed between the end point and the starting point, all segment path costs corresponding to all nodes, and the assignment results according to the starting point and end point determined by each wind turbine and the existing road, and finally generates a planned road with the lowest total segment path cost and is feasible. The processing process may be a well-known technical means in the art, and will not be described in detail here.

[0110] It can be understood that when the control terminal generates the planned road for each wind turbine, it can first identify all existing roads in the road distribution network of the wind farm and all wind turbine locations marked by technicians in the road distribution network, and can map all existing roads and all wind turbine locations to all the cells mentioned above.

[0111] Here, when all existing roads and all wind turbine positions are mapped to all cells, the control terminal can, but is not limited to, determine the positional relationship between any existing road and all other existing roads in the road distribution network, and the positional relationship between the road and all wind turbines, which can be understood as azimuth and straight-line distance; then, the edge contour of the feature consistent with the edge contour shape of the existing road is identified on all cells, and the area ratio between the edge contour of the feature and the edge contour of the existing road is calculated, so as to adjust the straight-line distance in all positional relationships according to the area ratio; then, all other existing roads and the positions of all wind turbines can be identified on all cells according to all adjusted positional relationships and edge contours of the feature. Of course, in the embodiment of the present application, all existing roads and all wind turbine positions can also be mapped to all cells by an image mapping algorithm, but it will not be elaborated here.

[0112] Next, after mapping all existing roads and all wind turbine locations to all cells, the control terminal can also filter out the nearest existing road corresponding to each wind turbine according to the vertical path length between each wind turbine and each existing road, and sort all wind turbines in ascending order of vertical path length according to the vertical path length between each wind turbine and the corresponding nearest existing road, so as to generate the planned road for each wind turbine in turn according to the sorting results of all wind turbines, thereby avoiding problems such as omissions in road design.

[0113] Then, after sorting all the wind turbines, the control terminal can use the vertical path between each wind turbine and the corresponding nearest existing road as the starting point, and the location of the corresponding wind turbine as the end point, and use the A* algorithm mentioned above to process the starting point, end point, all segment path costs corresponding to each node, and the assignment results corresponding to each wind turbine to obtain the planned road of the corresponding wind turbine. It should be noted that the starting point corresponding to each wind turbine in the embodiment of the present application can be but not limited to any node on all cells, and the end point corresponding to each wind turbine in the embodiment of the present application can also be but not limited to any node on all cells, so as to ensure the efficiency of generating the planned road.

[0114] Step 110: add the planned roads of all wind turbines to the road distribution network, and display the road distribution network including all wind turbines and all planned roads to the user.

[0115] Specifically, after obtaining the planned roads for all wind turbines, the control terminal can, but is not limited to, add the planned roads for all wind turbines to the road distribution network, and all added planned roads can be marked and distinguished from all existing roads in the road distribution network, and upon receiving the user's road display operation, the road distribution network including all wind turbines and all planned roads is displayed to the user, so that the user can perform road construction according to all the planned roads, or adjust any planned road, but is not limited to this.

[0116] See also Figure 4 , Figure 4 A schematic diagram of the structure of an automatic road design system for a new energy wind farm provided in an embodiment of the present application is shown.

[0117] like Figure 4 As shown, the automatic road design system for a new energy wind farm may include at least a model output module 401, a region division module 402, a node assignment module 403, a path generation module 404 and a road display module 405, wherein:

[0118] The model output module 401 is used to obtain digital images of ground objects in the wind farm, and input the digital images of the ground objects into the deep learning model to obtain a ground object distribution image; wherein the deep learning model is trained by at least two sample images marked with ground object distribution labels;

[0119] The region division module 402 is used to perform a superposition process on the ground object distribution image and the space planning image corresponding to the wind farm, and perform a region division process on the superimposed ground object distribution image based on preset ground object attributes;

[0120] The node assignment module 403 is used to divide the surface feature distribution image after the area division process into at least two cells, and assign a value to each node on the corresponding cell according to the area to which each cell belongs; wherein each cell corresponds to four nodes;

[0121] The path generation module 404 is used to calculate the segment path cost between each node and at least four adjacent nodes, and generate a planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node and the assignment result;

[0122] The road display module 405 is used to add the planned roads of all wind turbines to the road distribution network, and display the road distribution network including all wind turbines and all planned roads to the user.

[0123] In some possible embodiments, the overlay processing of the ground object distribution image and the space planning image corresponding to the wind farm includes:

[0124] Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour;

[0125] Based on the area ratio between any first edge contour and the corresponding second edge contour, the space planning image is scaled, and the object distribution image and the scaled space planning image are superimposed so that each first edge contour coincides with the corresponding second edge contour.

[0126] In some possible embodiments, based on preset ground object attributes, the ground object distribution image after the overlay processing is divided into regions, including:

[0127] In the object distribution image after the overlay processing, the object type consistent with the preset object attribute and the object area corresponding to the object type are identified;

[0128] Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold;

[0129] In the overlay processed feature distribution image, the area excluding all buffer areas is divided into the utilization area.

[0130] In some possible embodiments, according to the region to which each cell belongs, each node on the corresponding cell is assigned a value, including:

[0131] When the area to which any cell belongs is a buffer area, a first value assignment process is performed on each node on the corresponding cell;

[0132] When the area to which any cell belongs is a utilization area, the second value assignment process is performed on each node on the corresponding cell.

[0133] In some possible embodiments, calculating the segment path cost between each node and at least four adjacent nodes includes:

[0134] Determine at least four adjacent nodes corresponding to each node on all cells, and calculate the distance cost according to the path length between each node and each corresponding adjacent node;

[0135] Obtain the elevation of each node and each corresponding adjacent node on the object distribution image, and intercept the polyline segment between the elevation of each node and the elevation of each corresponding adjacent node in the preset elevation-mileage polyline;

[0136] The corresponding earthwork cost is calculated according to each broken line segment, and the distance cost and earthwork cost corresponding to each node and each adjacent node are weighted and summed to obtain the segment path cost.

[0137] In some possible embodiments, according to the position of each wind turbine in the road distribution network of the wind farm, the path costs of all segments corresponding to each node, and the assignment results, a planned road for the corresponding wind turbine is generated, including:

[0138] Identify the existing roads and the location of each wind turbine in the road distribution network of the wind farm, and map the existing roads and the location of each wind turbine on all cells;

[0139] According to the vertical path between each wind turbine and the existing road, the starting point and the end point are determined on all cells after mapping. Based on the path optimization algorithm, the starting point and the end point corresponding to each wind turbine, the path costs of all segments corresponding to each node, and the assignment results are processed to obtain the planned road for the corresponding wind turbine.

[0140] In some possible embodiments, before inputting the digital image of the ground object into the deep learning model to obtain the ground object distribution image, the method further includes:

[0141] Preprocessing the digital image of the ground object and calculating the signal-to-noise ratio of the preprocessed digital image of the ground object;

[0142] Input the digital image of the ground object into the deep learning model to obtain the ground object distribution image, including:

[0143] When the signal-to-noise ratio exceeds the preset signal-to-noise ratio threshold, the preprocessed digital image of the ground object is input into the deep learning model to obtain the ground object distribution image.

[0144] Those skilled in the art can clearly understand that the technical solutions of the embodiments of the present application can be implemented with the help of software and / or hardware. The "unit" and "module" in this specification refer to software and / or hardware that can independently complete or cooperate with other components to complete specific functions, where the hardware can be, for example, a field programmable gate array (FPGA), an integrated circuit (IC), etc.

[0145] See also Figure 5 , Figure 5 A structural schematic diagram of another automatic road design system for a new energy wind farm provided in an embodiment of the present application is shown.

[0146] like Figure 5As shown, the automatic road design system 500 for a new energy wind farm may include at least one processor 501 , at least one network interface 504 , a user interface 503 , a memory 505 , and at least one communication bus 502 .

[0147] The communication bus 502 may be used to realize the connection and communication among the above-mentioned components.

[0148] The user interface 503 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.

[0149] The network interface 504 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.

[0150] Among them, the processor 501 may include one or more processing cores. The processor 501 uses various interfaces and lines to connect various parts in the road automatic design system 500 for new energy wind farms, and executes various functions and processes data of the road automatic design system 500 for new energy wind farms by running or executing instructions, programs, code sets or instruction sets stored in the memory 505, and calling data stored in the memory 505. Optionally, the processor 501 can be implemented in at least one hardware form of DSP, FPGA, and PLA. The processor 501 can integrate one or a combination of CPU, GPU, modem, etc. Among them, the CPU mainly processes the operating system, user interface, and application program; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 501, and it can be implemented by a single chip.

[0151] The memory 505 may include a RAM or a ROM. Optionally, the memory 505 includes a non-transitory computer-readable medium. The memory 505 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 505 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 505 may optionally be at least one storage device located away from the aforementioned processor 501. As Figure 5 As shown, the memory 505 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an automatic road design application for a new energy wind farm.

[0152] Specifically, the processor 501 may be used to call the automatic road design application for the new energy wind farm stored in the memory 505, and specifically perform the following operations:

[0153] Acquire digital images of objects in the wind farm, and input the digital images into a deep learning model to obtain an object distribution image; wherein the deep learning model is trained by at least two sample images marked with object distribution labels;

[0154] Overlaying the object distribution image and the spatial planning image corresponding to the wind farm, and performing regional division processing on the overlayed object distribution image based on preset object attributes;

[0155] The object distribution image after the area division process is divided into at least two cells, and each node on the corresponding cell is assigned a value according to the area to which each cell belongs; wherein each cell corresponds to four nodes;

[0156] Calculate the segment path cost between each node and at least four adjacent nodes, and generate the planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all segment path costs corresponding to each node and the assignment results;

[0157] The planned roads of all wind turbines are added to the road distribution network, and the road distribution network including all wind turbines and all planned roads is displayed to the user.

[0158] In some possible embodiments, the overlay processing of the ground object distribution image and the space planning image corresponding to the wind farm includes:

[0159] Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour;

[0160] Based on the area ratio between any first edge contour and the corresponding second edge contour, the space planning image is scaled, and the object distribution image and the scaled space planning image are superimposed so that each first edge contour coincides with the corresponding second edge contour.

[0161] In some possible embodiments, based on preset ground object attributes, the ground object distribution image after the overlay processing is divided into regions, including:

[0162] In the object distribution image after the overlay processing, the object type consistent with the preset object attribute and the object area corresponding to the object type are identified;

[0163] Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold;

[0164] In the overlay processed feature distribution image, the area excluding all buffer areas is divided into the utilization area.

[0165] In some possible embodiments, according to the region to which each cell belongs, each node on the corresponding cell is assigned a value, including:

[0166] When the area to which any cell belongs is a buffer area, a first value assignment process is performed on each node on the corresponding cell;

[0167] When the area to which any cell belongs is a utilization area, the second value assignment process is performed on each node on the corresponding cell.

[0168] In some possible embodiments, calculating the segment path cost between each node and at least four adjacent nodes includes:

[0169] Determine at least four adjacent nodes corresponding to each node on all cells, and calculate the distance cost according to the path length between each node and each corresponding adjacent node;

[0170] Obtain the elevation of each node and each corresponding adjacent node on the object distribution image, and intercept the polyline segment between the elevation of each node and the elevation of each corresponding adjacent node in the preset elevation-mileage polyline;

[0171] The corresponding earthwork cost is calculated according to each broken line segment, and the distance cost and earthwork cost corresponding to each node and each adjacent node are weighted and summed to obtain the segment path cost.

[0172] In some possible embodiments, according to the position of each wind turbine in the road distribution network of the wind farm, the path costs of all segments corresponding to each node, and the assignment results, a planned road for the corresponding wind turbine is generated, including:

[0173] Identify the existing roads and the location of each wind turbine in the road distribution network of the wind farm, and map the existing roads and the location of each wind turbine on all cells;

[0174] According to the vertical path between each wind turbine and the existing road, the starting point and the end point are determined on all cells after mapping. Based on the path optimization algorithm, the starting point and the end point corresponding to each wind turbine, the path costs of all segments corresponding to each node, and the assignment results are processed to obtain the planned road for the corresponding wind turbine.

[0175] In some possible embodiments, before inputting the digital image of the ground object into the deep learning model to obtain the ground object distribution image, the method further includes:

[0176] Preprocessing the digital image of the ground object and calculating the signal-to-noise ratio of the preprocessed digital image of the ground object;

[0177] Input the digital image of the ground object into the deep learning model to obtain the ground object distribution image, including:

[0178] When the signal-to-noise ratio exceeds the preset signal-to-noise ratio threshold, the preprocessed digital image of the ground object is input into the deep learning model to obtain the ground object distribution image.

[0179] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0180] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0181] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0182] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0183] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0184] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

Claims

1. A method for automatic road design for a new energy wind farm, characterized in that: include: Acquire digital images of objects in the wind farm, and input the digital images into a deep learning model to obtain an object distribution image; wherein the deep learning model is trained by at least two sample images marked with object distribution labels; Overlaying the object distribution image and the space planning image corresponding to the wind farm, and performing region division processing on the object distribution image after the overlay processing based on preset object attributes; The surface feature distribution image after the area division process is divided into at least two cells, and each node on the corresponding cell is assigned a value according to the area to which each cell belongs; wherein each cell corresponds to four nodes; Calculate the segment path cost between each node and at least four adjacent nodes, and generate a planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all the segment path costs corresponding to each node, and the assignment result; Adding the planned roads of all the wind turbines to the road distribution network, and displaying the road distribution network including all the wind turbines and all the planned roads to the user; The superimposing process of the ground feature distribution image and the space planning image corresponding to the wind farm includes: Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour; Based on the area ratio between any one of the first edge contours and the corresponding second edge contour, scaling the space planning image, and superimposing the object distribution image and the scaled space planning image, so that each of the first edge contours overlaps with the corresponding second edge contour; The step of performing region division processing on the superimposed surface feature distribution image based on preset surface feature attributes includes: In the object distribution image after the overlay processing, identifying the object type consistent with the preset object attribute and the object area corresponding to the object type; Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold; Dividing the area of ​​the surface feature distribution image after the superposition processing, excluding all the buffer areas, into a utilization area; The step of assigning a value to each node on the corresponding cell according to the region to which each cell belongs includes: When the area to which any one of the cells belongs is the buffer area, performing a first value assignment process on each node on the corresponding cell; When the area to which any one of the cells belongs is the utilization area, performing a second value assignment process on each node on the corresponding cell; The step of calculating the segment path cost between each node and at least four adjacent nodes includes: Determine at least four adjacent nodes corresponding to each of the nodes on all the cells, and convert the distance cost according to the path length between each of the nodes and each of the corresponding adjacent nodes; Obtaining the elevation of each of the nodes and each of the corresponding adjacent nodes on the feature distribution image, and intercepting a polyline segment between the elevation of each of the nodes and the elevation of each of the corresponding adjacent nodes in a preset elevation-mileage polyline; The corresponding earthwork cost is calculated according to each of the broken line segments, and the distance cost and the earthwork cost corresponding to each of the nodes and each of the adjacent nodes are weighted summed to obtain the segment path cost; The generating of the planned road corresponding to the wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all the segment path costs corresponding to each node and the assignment result includes: Identifying existing roads and the location of each wind turbine in the road distribution network of the wind farm, and mapping the existing roads and the location of each wind turbine on all the cells; According to the vertical path between each wind turbine and the existing road, the starting point and the end point are determined on all the cells after mapping, and based on the path optimization algorithm, the starting point and the end point corresponding to each wind turbine, all the segment path costs corresponding to each node and the assignment results are processed to obtain the planned road for the corresponding wind turbine.

2. The method according to claim 1, characterized in that Before inputting the digital image of the ground object into the deep learning model to obtain the ground object distribution image, it also includes: Preprocessing the digital image of the ground object and calculating the signal-to-noise ratio of the preprocessed digital image of the ground object; The step of inputting the digital image of the ground object into the deep learning model to obtain a ground object distribution image comprises: When the signal-to-noise ratio exceeds a preset signal-to-noise ratio threshold, the preprocessed digital image of the ground object is input into a deep learning model to obtain a ground object distribution image.

3. An automatic road design system for a new energy wind farm, characterized in that: include: A model output module is used to obtain digital images of ground objects in the wind farm, and input the digital images of ground objects into a deep learning model to obtain a ground object distribution image; wherein the deep learning model is trained by at least two sample images marked with ground object distribution labels; A region division module, used for superimposing the object distribution image and the space planning image corresponding to the wind farm, and performing region division processing on the object distribution image after superimposition processing based on preset object attributes; A node assignment module, used for dividing the surface feature distribution image after the area division process into at least two cells, and assigning a value to each node on the corresponding cell according to the area to which each cell belongs; wherein each cell corresponds to four nodes; A path generation module, used to calculate the segment path cost between each node and at least four adjacent nodes, and generate a planned road for the corresponding wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all the segment path costs corresponding to each node and the assignment result; A road display module, used to add the planned roads of all the wind turbines to the road distribution network, and display the road distribution network including all the wind turbines and all the planned roads to the user; The superimposing process of the ground feature distribution image and the space planning image corresponding to the wind farm includes: Identify a first edge contour of at least one type of land feature in the land feature distribution image, and determine a second edge contour of each type of land feature in the spatial planning image corresponding to the wind farm; wherein the shape of the first edge contour of each type of land feature is consistent with the shape of the corresponding second edge contour; Based on the area ratio between any one of the first edge contours and the corresponding second edge contour, scaling the space planning image, and superimposing the object distribution image and the scaled space planning image, so that each of the first edge contours overlaps with the corresponding second edge contour; The step of performing region division processing on the superimposed surface feature distribution image based on preset surface feature attributes includes: In the object distribution image after the overlay processing, identifying the object type consistent with the preset object attribute and the object area corresponding to the object type; Marking feature points in the feature area, and dividing buffer areas corresponding to the feature points in the feature distribution image after superposition processing; wherein the distance between any area edge point on each buffer area and the corresponding feature point is a preset distance threshold; Dividing the area of ​​the surface feature distribution image after the superposition processing, excluding all the buffer areas, into a utilization area; The step of assigning a value to each node on the corresponding cell according to the region to which each cell belongs includes: When the area to which any one of the cells belongs is the buffer area, performing a first value assignment process on each node on the corresponding cell; When the area to which any one of the cells belongs is the utilization area, performing a second value assignment process on each node on the corresponding cell; The step of calculating the segment path cost between each node and at least four adjacent nodes includes: Determine at least four adjacent nodes corresponding to each of the nodes on all the cells, and convert the distance cost according to the path length between each of the nodes and each of the corresponding adjacent nodes; Obtaining the elevation of each of the nodes and each of the corresponding adjacent nodes on the feature distribution image, and intercepting a polyline segment between the elevation of each of the nodes and the elevation of each of the corresponding adjacent nodes in a preset elevation-mileage polyline; The corresponding earthwork cost is calculated according to each of the broken line segments, and the distance cost and the earthwork cost corresponding to each of the nodes and each of the adjacent nodes are weighted summed to obtain the segment path cost; The generating of the planned road corresponding to the wind turbine according to the position of each wind turbine in the road distribution network of the wind farm, all the segment path costs corresponding to each node and the assignment result includes: Identifying existing roads and the location of each wind turbine in the road distribution network of the wind farm, and mapping the existing roads and the location of each wind turbine on all the cells; According to the vertical path between each wind turbine and the existing road, the starting point and the end point are determined on all the cells after mapping, and based on the path optimization algorithm, the starting point and the end point corresponding to each wind turbine, all the segment path costs corresponding to each node and the assignment results are processed to obtain the planned road for the corresponding wind turbine.

4. An automatic road design system for a new energy wind farm, characterized in that: including a processor and a memory; The processor is connected to the memory; The memory is used to store executable program code; The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the steps of the method according to any one of claims 1 to 2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a computer or a processor, the computer or the processor executes the steps of the method according to any one of claims 1 to 2.

Images