Power transmission line selection method and device based on large language model, equipment and medium

Abstract

The embodiment of the invention discloses a transmission line selection method and device based on a large language model, equipment and a medium. A specific embodiment of the method comprises the following steps: acquiring a remote sensing image and radar point cloud data; executing data fusion processing to generate four-dimensional fusion data; constructing an immature claim expert knowledge base based on a preset immature claim rule set; performing fine tuning processing on the initial large language model to obtain a fine tuning large model; performing path planning processing on the initial topographic image to generate at least one candidate power transmission line; inputting the candidate power transmission line set into the fine tuning model to output line evaluation data of each candidate power transmission line in the corresponding candidate power transmission line set, and obtaining a line evaluation data set; and based on the line evaluation data set, selecting a target power transmission line from the candidate power transmission line set, and printing a line drawing corresponding to the target power transmission line. According to the embodiment, the time for selecting the power transmission line is shortened, and the cost for constructing the power transmission line is reduced.

Description

TECHNICAL FIELD

[0001] Embodiments of the present disclosure relate to the technical field of computer technology, and in particular, to a power transmission line selection method and device based on a large language model, an electronic device, and a computer readable medium. BACKGROUND

[0002] As the core artery of the power system, the path selection of the power transmission line directly affects the economy, safety, and long-term operation efficiency of the power grid construction. Currently, when selecting a power transmission line, the commonly used method is to select by professional technical personnel through on-site measurement and visual assessment of the terrain and the distribution of ground objects, or to generate a route by combining a path planning algorithm with a GIS system.

[0003] However, when the above method is used to select a power transmission line, the following technical problems often exist:

[0004] When professional technical personnel select a power transmission line, a long time is required for selection, and the selection is highly subjective. The power transmission line generated by the path planning algorithm does not take into account problems such as green compensation, resulting in a high cost of constructing the power transmission line.

[0005] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present inventive concept, and therefore, it can include information that does not form the prior art known to those of ordinary skill in the art in the country. SUMMARY

[0006] The summary section is provided to introduce the concepts briefly in a simplified form, which will be described in detail in the specific embodiments section. The summary section is not intended to identify key or essential features of the claimed technical solution nor is it intended to be used to limit the scope of the claimed technical solution.

[0007] Some embodiments of the present disclosure propose a power transmission line selection method and device based on a large language model, an electronic device, and a computer readable medium to solve one or more of the technical problems mentioned in the background section.

[0008] In a first aspect, some embodiments of the present disclosure provide a power transmission line selection method based on a large language model, which comprises: acquiring remote sensing images and radar point cloud data, wherein the radar point cloud data comprises elevation data and ground object classification information; performing data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data; constructing a green compensation expert knowledge base based on a preset green compensation rule set; fine-tuning an initial large language model based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model; performing path planning processing on an initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set; inputting the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, thereby obtaining a line evaluation data set; selecting a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and printing a line drawing corresponding to the target power transmission line.

[0009] In a second aspect, some embodiments of the present disclosure provide a power transmission line selection device based on a large language model, which comprises: an acquisition unit configured to acquire remote sensing images and radar point cloud data, wherein the radar point cloud data comprises elevation data and ground object classification information; an execution unit configured to perform data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data; a construction unit configured to construct a green compensation expert knowledge base based on a preset green compensation rule set; a fine-tuning unit configured to fine-tune an initial large language model based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model; a path planning unit configured to perform path planning processing on an initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set; an input unit configured to input the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, thereby obtaining a line evaluation data set; a selection unit configured to select a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and print a line drawing corresponding to the target power transmission line.

[0010] In a third aspect, some embodiments of the present disclosure provide an electronic device, which comprises: one or more processors; a storage device having one or more programs stored thereon, when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any of the implementation manners of the first aspect.

[0011] In a fourth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein the program is executed by a processor to implement the method described in any of the implementation manners of the first aspect.

[0012] The above various embodiments of the present disclosure have the following beneficial effects: through the power transmission line selection method based on the large language model of some embodiments of the present disclosure, the time for selecting the power transmission line is reduced, and the cost of constructing the power transmission line is reduced. Specifically, the reason for consuming a long time to select the power transmission line and the high cost of constructing the power transmission line is that when the professional technicians select the power transmission line, a long time is consumed for selection, and the subjectivity of selection is strong, and the power transmission line generated by the path planning algorithm does not consider the green penalty and other problems, resulting in a high cost of constructing the power transmission line. Based on this, the power transmission line selection method based on the large language model of some embodiments of the present disclosure first acquires remote sensing images and radar point cloud data; based on the remote sensing images and the radar point cloud data, data fusion processing is performed to generate four-dimensional fusion data. Thus, the fused terrain data can be generated. Second, based on a preset green penalty rule set, a green penalty expert knowledge base is constructed; according to the four-dimensional fusion data and the green penalty expert knowledge base, an initial large language model is fine-tuned to obtain a fine-tuned large model. Thus, the large language model can be fine-tuned through the green penalty knowledge base, so as to include the power transmission line cost in the selection strategy. Then, an initial terrain image is subjected to path planning processing to generate at least one candidate power transmission line, obtaining a candidate power transmission line set. Thus, multiple candidate power transmission lines can be generated for selection. Then, the candidate power transmission line set is input into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, obtaining a line evaluation data set. Thus, each candidate line can be evaluated. Finally, based on the line evaluation data set, a target power transmission line is selected from the candidate power transmission line set, and a line drawing corresponding to the target power transmission line is printed. Thus, the image of the selected power transmission line can be printed. The time for selecting the power transmission line is reduced, and the cost of constructing the power transmission line is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0013] The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent upon consideration of the following detailed description, taken in conjunction with the accompanying drawings. Throughout the drawings, like or similar reference numerals designate identical or similar elements throughout the several views. It should be understood that the drawings are schematic and elements and features are not necessarily to scale.

[0014] Figure 1 is a flowchart of some embodiments of the power transmission line selection method based on the large language model according to the present disclosure;

[0015] Figure 2 is a structural schematic diagram of some embodiments of the power transmission line selection device based on the large language model according to the present disclosure;

[0016] Figure 3 is a structural schematic diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0017] Embodiments of the present disclosure will be described below in greater detail with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly and completely understood. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0018] It should also be noted that, for ease of description, only parts related to the present application are shown in the drawings. The embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0019] It should be noted that the terms "first", "second", and the like mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not intended to limit the order or interdependence of the functions performed by these devices, modules or units.

[0020] It should be noted that the terms "one", "multiple" mentioned in the present disclosure are illustrative and not restrictive, and those skilled in the art should understand that, unless otherwise explicitly stated in the context, it should be understood as "one or more".

[0021] The names of the messages or information exchanged between the devices in the embodiments of the present disclosure are only for illustrative purposes and are not intended to limit the scope of the messages or information.

[0022] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0023] Figure 1 Flow 100 of some embodiments of the power transmission line selection method based on a large language model according to the present disclosure is shown. The power transmission line selection method based on a large language model includes the following steps:

[0024] Step 101, acquiring remote sensing images and radar point cloud data.

[0025] In some embodiments, the execution subject (such as a server) of the power transmission line selection method based on a large language model can acquire remote sensing images and radar point cloud data. The radar point cloud data includes elevation data and ground object classification information. The elevation data can be an altitude. The ground object classification information can be pre-set classification information of ground objects. The remote sensing images can be high-resolution RGB images. The resolution of the remote sensing images is greater than or equal to 0.5 meters / pixel.

[0026] Optionally, after step 101, the following steps are further included:

[0027] Firstly, at least one image feature point corresponding to the remote sensing image is extracted to obtain an image feature point set.

[0028] In some embodiments, the execution subject can extract at least one image feature point corresponding to the remote sensing image to obtain an image feature point set. In practice, the image feature point set can be obtained by extracting the image feature point through the SIFT (Scale-Invariant Feature Transform) algorithm.

[0029] Secondly, based on a preset algorithm, the false matching points in the image feature point set included in the remote sensing image are eliminated to generate an eliminated remote sensing image.

[0030] In some embodiments, the execution subject can eliminate the false matching points in the image feature point set included in the remote sensing image based on a preset algorithm to generate an eliminated remote sensing image. The preset algorithm can be the RANSAC (Random Sample Consensus) algorithm.

[0031] Thirdly, the radar point cloud data is denoised through bilateral filtering to generate denoised radar point cloud data.

[0032] In some embodiments, the execution subject can denoise the radar point cloud data through bilateral filtering to generate denoised radar point cloud data.

[0033] Fourthly, the geometric features of the denoised radar point cloud data are determined.

[0034] In some embodiments, the execution subject can determine the geometric features of the denoised radar point cloud data. The geometric features include the normal vector and the curvature. In practice, the local surface features can be solved by searching the 0.5m neighborhood points based on the KD tree and through the PCA (Principal Component Analysis).

[0035] Step 102, based on the remote sensing image and the radar point cloud data, data fusion processing is performed to generate four-dimensional fusion data.

[0036] In some embodiments, the execution subject can perform data fusion processing based on the remote sensing image and the radar point cloud data to generate four-dimensional fusion data.

[0037] In practice, the data fusion processing can be performed based on the remote sensing image and the radar point cloud data to generate four-dimensional fusion data through the following steps:

[0038] In a first step, the radar point cloud data is converted to generate converted point cloud data. Here, the radar point cloud data can be converted from a local coordinate system to a WGS84 geographic coordinate system.

[0039] In a second step, an image resolution corresponding to the remote sensing image is determined based on the remote sensing image.

[0040] In a third step, the converted point cloud data is resampled based on bilinear interpolation and the image resolution to generate resampled point cloud data. In practice, the converted point cloud data can be resampled by bilinear interpolation.

[0041] In a fourth step, the remote sensing image and the resampled point cloud data are channel spliced to generate a four-dimensional tensor as four-dimensional fusion data.

[0042] In the process of using the technical solutions to solve the technical problems mentioned in the background, the following technical problems often occur: when performing terrain modeling, due to the incompleteness of the obtained terrain data, a digital model corresponding to the actual terrain cannot be generated, resulting in the inability to generate a corresponding power transmission line. In view of the existing technical status, terrain modeling can be performed by the following steps:

[0043] Optionally, after the fourth step, the following steps are further included:

[0044] In a fifth step, a digital model of a target region is generated based on the four-dimensional fusion data. In practice, the digital model can be constructed by a moving least squares method.

[0045] In a sixth step, terrain repair processing is performed on the digital model to generate a repaired digital model. Here, the execution subject can perform continuity repair on the digital model by a Poisson equation.

[0046] In a seventh step, a residual distribution map corresponding to the repaired digital model is generated. In practice, the residual distribution map can be constructed by a thin-plate spline interpolation.

[0047] In an eighth step, a spatial error value corresponding to the repaired digital model is determined according to the residual distribution map. In practice, the ratio of the area of a region with a residual greater than a preset value to the area of the residual distribution map can be determined as the spatial error value.

[0048] In a ninth step, in response to the spatial error value being greater than or equal to a preset error value, terrain repair processing is performed on the digital model to generate a repaired digital model.

[0049] The first step to the optional step above, as one of the invention points of the embodiments of the present disclosure, solves the technical problem that "when performing terrain modeling, due to the defects in the obtained terrain data, a digital model corresponding to the actual terrain cannot be generated, resulting in the inability to generate a corresponding power transmission line". The reason why the corresponding power transmission line cannot be generated is as follows: when performing terrain modeling, due to the defects in the obtained terrain data, a digital model corresponding to the actual terrain cannot be generated, resulting in the inability to generate a corresponding power transmission line. If the above factors are solved, the effect of avoiding the situation that the corresponding power transmission line cannot be generated can be achieved. In order to achieve this effect, the present disclosure first generates a digital model corresponding to the target area based on the four-dimensional fusion data. Thus, an initial digital model can be generated. Second, the digital model is subjected to terrain repair processing to generate a repaired digital model. Thus, the digital model with missing information can be repaired. Third, a residual distribution map corresponding to the repaired digital model is generated. In practice, the residual distribution map can be constructed by thin plate spline interpolation. Thus, the residual of the repaired digital model and the actual terrain can be determined, and thus the residual distribution is constructed. Fourth, according to the residual distribution map, the spatial error value corresponding to the repaired digital model is determined; in response to the spatial error value being greater than or equal to a preset error value, the digital model is subjected to terrain repair processing to generate a repaired digital model. Thus, when the residual is large, the digital model can be repaired again, so as to obtain an accurate digital model, and thus the situation that the corresponding power transmission line cannot be generated due to the missing terrain data can be avoided.

[0050] In step 103, a green compensation expert knowledge base is constructed based on a preset green compensation rule set.

[0051] In some embodiments, the execution subject can construct a green compensation expert knowledge base based on a preset green compensation rule set. In practice, the preset green compensation rule set of the target area can be encoded to generate a green compensation expert knowledge base.

[0052] In step 104, an initial large language model is fine-tuned based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model.

[0053] In some embodiments, the execution subject can fine-tune an initial large language model based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model. In practice, multi-modal data of fused remote sensing images and laser point clouds can be input to train the model to extract spatial attributes of ground objects.

[0054] In step 105, an initial terrain image is subjected to path planning processing to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set.

[0055] In some embodiments, the execution subject can perform path planning processing on the initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set.

[0056] In practice, the initial terrain image can be processed by the following steps to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set:

[0057] Firstly, the initial terrain image is processed by three-dimensional modeling to generate a three-dimensional initial terrain model.

[0058] Secondly, the three-dimensional initial terrain model is processed by discretization to generate a discretized three-dimensional terrain model. The discretized three-dimensional terrain model includes at least one terrain grid. The discretization processing can be performed by dividing the terrain cells through Voronoi diagram, establishing adjacency relationship through Delaunay triangulation, and generating the discretized three-dimensional terrain model.

[0059] Thirdly, based on the preset path planning algorithm and the at least one terrain grid included in the discretized three-dimensional terrain model, at least one candidate power transmission line is generated, thereby obtaining a candidate power transmission line set. In practice, the multi-constraint Dijkstra optimization algorithm can be used to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set. Here, the multi-constraint Dijkstra optimization algorithm can be configured with a dynamic weight adjustment strategy: in the initial stage, the technical and economic weight is dominant (60%), and in the iteration stage, the weight distribution is adjusted according to the reinforcement learning feedback.

[0060] In the process of using the technical solutions to solve the technical problems mentioned in the background, the following technical problems often occur: when selecting a power transmission line, due to the large number of nodes between the starting point and the ending point, it takes a long time to select the candidate power transmission line. In view of the above technical problems, the conventional solution is generally to use depth-first search or breadth-first search algorithm to search for the power transmission line. However, the above conventional solution still has the following problems: when using the depth-first search or breadth-first search algorithm to select the power transmission line, all nodes need to be traversed, which results in a long time to select the candidate power transmission line.

[0061] In some optional implementations of some embodiments, the execution subject can perform path planning processing on the initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set, by the following steps:

[0062] Firstly, the starting point and the ending point corresponding to the power transmission line are obtained as the initial information of the power transmission line.

[0063] Secondly, based on the initial information of the power transmission line, a corresponding set of route node information is determined. The route node information in the set of route node information can be coordinate information of a path node between the start point and the end point.

[0064] Thirdly, a set of path nodes and a set of exploration paths are generated. The set of path nodes and the set of exploration paths are initially empty.

[0065] Fourthly, based on the start point, the end point, the set of path nodes and the set of exploration paths, the following processing steps are performed:

[0066] Firstly, the start point is added to the set of path nodes to generate an added set of path nodes as the set of path nodes.

[0067] Secondly, at least one adjacent node corresponding to the start point is determined as a set of adjacent nodes.

[0068] Thirdly, an adjacent node satisfying a first preset condition is selected from the set of adjacent nodes as a target adjacent node. The first preset condition can be an adjacent node closest to the end point in a straight line.

[0069] Fourthly, it is determined whether the target adjacent node satisfies a preset end point condition. The preset end point condition can be that the coordinates corresponding to the target adjacent node are the same as the coordinates corresponding to the end point.

[0070] Fifthly, in response to the target adjacent node satisfying the preset end point condition, the target adjacent node is added to the set of path nodes and the set of exploration paths respectively to update the set of path nodes and the set of exploration paths.

[0071] Sixthly, in response to the target adjacent node being the same as the end point, each path node in the set of path nodes is combined as a candidate power transmission line.

[0072] Optionally, after the fourth step, in response to the target adjacent node being different from the end point, the target adjacent node is taken as the start point, the updated set of path nodes and the set of exploration paths are taken as the set of path nodes and the set of exploration paths, and the processing steps are performed again.

[0073] The first step to the optional step above, as one of the invention points of the embodiments of the present disclosure, in combination with step 107 below, solves the technical problem that "it takes a long time to select candidate power transmission lines due to the large number of nodes between the starting point and the ending point when selecting power transmission lines". The reason for the need to take a long time to select candidate power transmission lines is as follows: when selecting power transmission lines, it takes a long time to select candidate power transmission lines due to the large number of nodes between the starting point and the ending point. If the above factors are solved, the time to select candidate power transmission lines can be reduced. To achieve this effect, the present disclosure first obtains the starting point and the ending point corresponding to the power transmission line as the initial information of the power transmission line. In this way, the positions of the starting point and the ending point of the power transmission line can be determined. Second, based on the above initial information of the power transmission line, the corresponding route node information set is determined. In this way, the route nodes existing between the starting point and the ending point can be determined. Third, the path node set and the exploration path set are generated. In this way, a collection for storing path nodes can be generated. Fourth, based on the starting point, the ending point, the path node set, and the exploration path set, the following processing steps are performed: first, the starting point is added to the path node set to generate an added path node set as the path node set; at least one adjacent node corresponding to the starting point is determined as an adjacent node set. In this way, each node adjacent to the starting point can be determined. Second, select an adjacent node that meets the first preset condition from the adjacent node set as a target adjacent node. In this way, the adjacent node closest to the ending point can be selected. Then, it is determined whether the target adjacent node meets the preset ending point condition; in response to the target adjacent node meeting the preset ending point condition, the target adjacent node is added to the path node set and the exploration path set, respectively, to update the path node set and the exploration path set. In this way, the adjacent node that meets the condition can be used as the next node of the starting point. Finally, in response to the target adjacent node being consistent with the ending point, each path node in the path node set is combined as a candidate power transmission line; in response to the target adjacent node being inconsistent with the ending point, the target adjacent node is used as the starting point, the updated path node set and the exploration path set are used as the path node set and the exploration path set, and the above processing steps are executed again. In this way, the power transmission line can be selected. In combination with the following step "step 107", based on the route evaluation data set, the target power transmission line is selected from the candidate power transmission line set, and the route drawing paper corresponding to the target power transmission line is printed. In this way, the time to select candidate power transmission lines can be reduced.

[0074] Step 106: inputting the candidate power transmission line set into the fine-tuning large model to output route evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, to obtain a route evaluation data set.

[0075] In some embodiments, the execution subject can input the candidate power transmission line set into the fine-tuning large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, and obtain a line evaluation data set.

[0076] At step 107, based on the line evaluation data set, a target power transmission line is selected from the candidate power transmission line set, and a line drawing corresponding to the target power transmission line is printed.

[0077] In some embodiments, the execution subject can select a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and print a line drawing corresponding to the target power transmission line.

[0078] Optionally, after step 107, the following steps are further included:

[0079] First, the policy update terminal sends policy update information corresponding to the initial terrain image to the execution subject.

[0080] In some embodiments, the execution subject can receive policy update information corresponding to the initial terrain image sent by the policy update terminal.

[0081] Second, the policy update information is subjected to keyword extraction processing to generate policy keyword information.

[0082] In some embodiments, the execution subject can perform keyword extraction processing on the policy update information to generate policy keyword information.

[0083] Third, the target power transmission line is subjected to splitting processing to generate a locked line and an unlocked line.

[0084] In some embodiments, the execution subject can perform splitting processing on the target power transmission line to generate a locked line and an unlocked line. The locked line can be a line that does not need to be updated. The unlocked line can be a line affected by the policy update information. For example, the policy update information can be the establishment of an ecological protection zone, and the unlocked line is within the ecological protection zone.

[0085] Fourth, based on the policy keyword information, a section corresponding to the policy keyword information is selected from the unlocked line as a target section.

[0086] In some embodiments, the execution subject can select a section corresponding to the policy keyword information from the unlocked line as a target section based on the policy keyword information.

[0087] In the fifth step, the path re-planning process is performed on the target section to generate at least one re-planned path, thereby obtaining a re-planned path set.

[0088] In some embodiments, the execution subject can perform the path re-planning process on the target section to generate at least one re-planned path, thereby obtaining a re-planned path set. In practice, the path re-planning process can be performed on the target section by using a gradient descent algorithm to generate at least one re-planned path, thereby obtaining a re-planned path set.

[0089] In the sixth step, for each re-planned path in the re-planned path set, a path evaluation corresponding to the re-planned path is determined.

[0090] In some embodiments, the execution subject can determine, for each re-planned path in the re-planned path set, a path evaluation corresponding to the re-planned path. In practice, each re-planned path in the re-planned path set can be input into the fine-tuned large model to generate a path evaluation corresponding to the re-planned path.

[0091] In the seventh step, based on the determined path evaluations, a re-planned path satisfying a preset selection condition is selected from the re-planned path set as a target re-planned path.

[0092] In some embodiments, the execution subject can select, based on the determined path evaluations, a re-planned path satisfying a preset selection condition from the re-planned path set as a target re-planned path.

[0093] In the eighth step, the target re-planned path, the locked line, and the non-locked line excluding the target section are combined to form an updated power transmission line.

[0094] In some embodiments, the execution subject can combine the target re-planned path, the locked line, and the non-locked line excluding the target section to form an updated power transmission line.

[0095] In the ninth step, based on the updated power transmission line and the target power transmission line, a line update difference report is generated.

[0096] In some embodiments, the execution subject can generate, based on the updated power transmission line and the target power transmission line, a line update difference report. The line update difference report can represent the difference between the data corresponding to the updated power transmission line and the target power transmission line.

[0097] The above various embodiments of the present disclosure have the following beneficial effects: through the power transmission line selection method based on the large language model of some embodiments of the present disclosure, the time for selecting the power transmission line is reduced, and the cost of constructing the power transmission line is reduced. Specifically, the reason for spending a long time to select the power transmission line and the high cost of constructing the power transmission line is that when selecting the power transmission line by professional technicians, a long time is spent for selection, and the subjectivity of selection is strong, and the power transmission line generated by the path planning algorithm does not consider the green penalty and other problems, resulting in a high cost of constructing the power transmission line. Based on this, the power transmission line selection method based on the large language model of some embodiments of the present disclosure, first, acquires remote sensing images and radar point cloud data; based on the remote sensing images and the radar point cloud data, performs data fusion processing to generate four-dimensional fusion data. Thus, the fused terrain data can be generated. Second, based on a preset green penalty rule set, a green penalty expert knowledge base is constructed; according to the four-dimensional fusion data and the green penalty expert knowledge base, the initial large language model is fine-tuned to obtain a fine-tuned large model. Thus, the large language model can be fine-tuned through the green penalty knowledge base, so as to include the power transmission line cost in the selection strategy. Then, the initial terrain image is processed for path planning to generate at least one candidate power transmission line, obtaining a candidate power transmission line set. Thus, multiple candidate power transmission lines can be generated for selection. Then, the candidate power transmission line set is input into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, obtaining a line evaluation data set. Thus, each candidate line can be evaluated. Finally, based on the line evaluation data set, a target power transmission line is selected from the candidate power transmission line set, and a line drawing corresponding to the target power transmission line is printed. Thus, the image of the selected power transmission line can be printed. The time for selecting the power transmission line is reduced, and the cost of constructing the power transmission line is reduced.

[0098] Further reference Figure 2 , as an implementation of the method shown in the above figures, the present disclosure provides some embodiments of a power transmission line selection device based on a large language model, which device embodiments correspond to the method embodiments shown in Figure 1 , the power transmission line selection device based on a large language model can be applied to various electronic devices.

[0099] As Figure 2As shown, the power transmission line selection device 200 based on the large language model of some embodiments includes an acquisition unit 201, an execution unit 202, a construction unit 203, a fine-tuning unit 204, a path planning unit 205, an input unit 206, and a selection unit 207. Among them, the acquisition unit 201 is configured to acquire remote sensing images and radar point cloud data, wherein the radar point cloud data includes elevation data and ground object classification information; the execution unit 202 is configured to perform data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data; the construction unit 203 is configured to construct a green compensation expert knowledge base based on a preset green compensation rule set; the fine-tuning unit 204 is configured to fine-tune an initial large language model according to the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model; the path planning unit 205 is configured to perform path planning processing on the initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set; the input unit 206 is configured to input the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, thereby obtaining a line evaluation data set; and the selection unit 207 is configured to select a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and print a line drawing corresponding to the target power transmission line.

[0100] It can be understood that the units described in the power transmission line selection device 200 based on the large language model correspond to the respective steps in the method described above. Figure 1 Therefore, the operations, features and beneficial effects described above for the method also apply to the power transmission line selection device 200 based on the large language model and the units contained therein, which will not be described here.

[0101] Reference is made below to Figure 3 which shows a structural schematic diagram of an electronic device 300 suitable for implementing some embodiments of the present disclosure. The electronic device in some embodiments of the present disclosure can include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet PCs), PMPs (Portable Multimedia Players), vehicle terminals (such as vehicle navigation terminals), and the like, as well as fixed terminals such as digital TVs, desktop computers, and the like. Figure 3 The electronic device shown is only an example and should not impose any limitations on the functions and use range of embodiments of the present disclosure.

[0102] As Figure 3As shown, the electronic device 300 can include a processing device 301 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 302 or loaded into a random access memory (RAM) 303 from a storage device 308. Various programs and data required for the operation of the electronic device 300 are also stored in the RAM 303. The processing device 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0103] Generally, the following devices can be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 309. The communication devices 309 can allow the electronic device 300 to communicate wirelessly or wired with other devices to exchange data. Although Figure 3 The electronic device 300 is shown with various devices, but it should be understood that all of the illustrated devices are not required, and more or fewer devices can alternatively be implemented. Figure 3 Each block shown in the flowcharts can represent a device, or multiple devices, as necessary.

[0104] In particular, processes described above with reference to the flowcharts can be implemented as a computer software program according to some embodiments of the present disclosure. For example, some embodiments of the present disclosure include a computer program product including a computer program carried on a computer readable medium, the computer program containing program code for performing the methods illustrated by the flowcharts. In some such embodiments, the computer program can be downloaded and installed from a network through the communication devices 309, or installed from the storage devices 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the above-described functions defined in the methods of some embodiments of the present disclosure are performed.

[0105] Note that the computer readable medium in some embodiments of the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination thereof. The computer readable storage medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or any suitable combination of the foregoing. More specific examples of the computer readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, the computer readable storage medium can be any tangible medium that contains or stores a program used by an instruction execution system, apparatus or device, or that can be used by or in connection with an instruction execution system, apparatus or device. In some embodiments of the present disclosure, the computer readable signal medium can include a computer readable program code propagated in or on a carrier medium, in which the computer readable program code is embodied. Such propagated computer readable program code can take many forms, including but not limited to, an electromagnetic signal, an optical signal or any suitable combination of the foregoing. The computer readable signal medium can also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate or transport a program for use by or in connection with an instruction execution system, apparatus or device. Program code embodied on a computer readable medium can be transmitted using any suitable medium, including but not limited to, wire, cable, wireless, RF, infrared or any suitable combination of the foregoing.

[0106] In some embodiments, the client, server, or both can communicate using any current known or future developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet, and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any current known or future developed networks.

[0107] The computer readable medium can be included in the electronic device, or can exist separately from the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquire remote sensing image and radar point cloud data, wherein the radar point cloud data includes elevation data and feature classification information; perform data fusion processing based on the remote sensing image and the radar point cloud data to generate four-dimensional fusion data; construct a green compensation expert knowledge base based on a preset green compensation rule set; fine-tune an initial large language model based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model; perform path planning processing on an initial terrain image to generate at least one candidate power transmission line to obtain a candidate power transmission line set; and input the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set to obtain a line evaluation data set. Based on the line evaluation data set, a target power transmission line is selected from the candidate power transmission line set, and a line drawing corresponding to the target power transmission line is printed.

[0108] Computer program code for carrying out operations of some embodiments of the disclosure can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0109] The computer program product of the first aspect can include a computer readable storage medium. The computer readable storage medium can include instructions. The instructions can include one or both of: instructions for causing a computer to implement a method as described above; and instructions for causing a computer to operate based on a system as described above. The computer readable storage medium can include one or more types of computer readable storage media. For example, the computer readable storage medium can include at least one of the following: a hard disk; a floppy disk; a CD-ROM; a DVD; a Blu-ray Disc; a memory stick; a memory card; a RAM; and a ROM. The computer readable storage medium can include a non-transitory computer readable storage medium.

[0110] The units described in some embodiments of the present disclosure can be implemented by means of software, or by means of hardware. The described units can also be provided in a processor, for example, it can be described that: a processor includes an acquisition unit, an execution unit, a construction unit, a fine-tuning unit, a path planning unit, an input unit and a selection unit. Among them, the names of these units do not constitute a limitation to the units themselves in some cases, for example, the acquisition unit can also be described as "a unit for acquiring remote sensing image and radar point cloud data".

[0111] The functions described above in the present document can be performed, at least in part, by one or more hardware logic components. For example, and without limitation, example types of hardware logic components that can be used include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip systems (SOCs), complex programmable logic devices (CPLDs), etc.

[0112] The above description is merely some of the preferred embodiments of the present disclosure and a description of the principles of the technology used. Those skilled in the art should understand that the scope of the application involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or equivalent features without departing from the above inventive concept. For example, the above features are replaced with each other to form a technical solution with similar functions disclosed in the embodiments of the present disclosure (but not limited to).

Claims

1. A power transmission line selection method based on a large language model, comprising: obtaining remote sensing images and radar point cloud data, wherein the radar point cloud data comprises elevation data and surface feature classification information; performing data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data; constructing a green compensation expert knowledge base based on a preset green compensation rule set; fine-tuning an initial large language model based on the four-dimensional fusion data and the green compensation expert knowledge base to obtain a fine-tuned large model; performing path planning processing on an initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set; inputting the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, thereby obtaining a line evaluation data set; selecting a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and printing a line drawing corresponding to the target power transmission line.

2. The method of claim 1, wherein, The data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data comprises: performing coordinate conversion processing on the radar point cloud data to generate converted point cloud data; determining an image resolution corresponding to the remote sensing images based on the remote sensing images; performing resampling processing on the converted point cloud data based on bilinear interpolation and the image resolution to generate resampled point cloud data; performing channel splicing processing on the remote sensing images and the resampled point cloud data to generate a four-dimensional tensor as four-dimensional fusion data.

3. The method of claim 1, wherein, After obtaining the remote sensing images and the radar point cloud data, the method further comprises: extracting at least one image feature point corresponding to the remote sensing images to obtain an image feature point set; eliminating mis-matching points in the image feature point set included in the remote sensing images based on a preset algorithm to generate an eliminated remote sensing image; performing denoising processing on the radar point cloud data through bilateral filtering to generate denoised radar point cloud data; determining geometric features of the denoised radar point cloud data, wherein the geometric features comprise normal vectors and curvatures.

4. The method of claim 1, wherein, The path planning processing on the initial terrain image to generate at least one candidate power transmission line, thereby obtaining a candidate power transmission line set, comprises: performing three-dimensional modeling processing on the initial terrain image to generate a three-dimensional initial terrain model; performing discretization processing on the three-dimensional initial terrain model to generate a discretized three-dimensional terrain model, wherein the discretized three-dimensional terrain model comprises at least one terrain grid; generating at least one candidate power transmission line based on a preset path planning algorithm and the at least one terrain grid included in the discretized three-dimensional terrain model, thereby obtaining a candidate power transmission line set.

5. The method of claim 1, wherein, The method further comprises: receiving policy update information corresponding to the initial terrain image sent by a policy update terminal; performing keyword extraction processing on the policy update information to generate policy keyword information; performing splitting processing on the target power transmission line to generate a locked line and a non-locked line; select a section corresponding to the policy keyword information from the non-locked line as a target section based on the policy keyword information; perform path re-planning processing on the target section to generate at least one re-planned path, to obtain a re-planned path set; determine a path evaluation corresponding to each re-planned path in the re-planned path set; select a re-planned path satisfying a preset selection condition from the re-planned path set as a target re-planned path based on the determined path evaluations; combine the target re-planned path, the locked line, and the non-locked line without the target section into an updated power transmission line; generate a line update difference report based on the updated power transmission line and the target power transmission line.

6. A power transmission line selection device based on a large language model, comprising: an acquisition unit configured to acquire remote sensing images and radar point cloud data, wherein the radar point cloud data includes elevation data and feature classification information; an execution unit configured to perform data fusion processing based on the remote sensing images and the radar point cloud data to generate four-dimensional fusion data; a construction unit configured to construct a compensation expert knowledge base based on a preset compensation rule set; a fine-tuning unit configured to fine-tune an initial large language model based on the four-dimensional fusion data and the compensation expert knowledge base to obtain a fine-tuned large model; a path planning unit configured to perform path planning processing on an initial terrain image to generate at least one candidate power transmission line, to obtain a candidate power transmission line set; an input unit configured to input the candidate power transmission line set into the fine-tuned large model to output line evaluation data corresponding to each candidate power transmission line in the candidate power transmission line set, to obtain a line evaluation data set; a selection unit configured to select a target power transmission line from the candidate power transmission line set based on the line evaluation data set, and to print a line drawing corresponding to the target power transmission line.

7. An electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon; when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-5.

8. A computer readable medium having stored thereon a computer program, wherein, The program is executed by the processor to implement the method of any one of claims 1-5. The program is executed by the processor to implement the method of any one of claims 1-5.

Images