A power transmission line route automatic planning and dynamic adjustment system based on a smart terminal

Abstract

The application provides a kind of based on intelligent terminal transmission line route automatic planning and dynamic adjustment system, the system includes data modeling module, regional assessment module, route planning module and defect identification module;The data modeling module is used to construct the three-dimensional space model of transmission line;The regional assessment module is used to divide transmission line into at least one inspection area according to geographical and structural characteristics, and dynamically assesses each inspection area to generate inspection points within the region;The route planning module is used to generate inspection routes based on the generated inspection points;The defect identification module is used to identify defects in transmission line equipment based on image data collected during unmanned aerial vehicle inspection;The application realizes the intelligentization, precision and dynamic optimization of transmission line inspection tasks, significantly improves the high-risk equipment identification efficiency and the real-time response capability of unmanned aerial vehicle route planning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power transmission line inspection, and particularly relates to a power transmission line route automatic planning and dynamic adjustment system based on an intelligent terminal. BACKGROUND

[0002] With the continuous expansion of the scale of power transmission lines and the increasing complexity of the operating environment, the traditional method of relying on manual field inspection faces many challenges in terms of efficiency, safety and comprehensive coverage. In recent years, using unmanned aerial vehicles (UAVs) to carry out power transmission line inspection has become a mainstream trend, which not only improves work efficiency but also reduces operational risks.

[0003] However, there are still the following problems in practical application: on the one hand, most of the UAV route planning is based on static paths or fixed inspection points, and lacks the ability to dynamically adjust according to the line state, making it difficult to respond to changes in equipment risks in a timely manner; on the other hand, the linkage between defect identification and task scheduling systems is insufficient, resulting in phenomena such as repeated inspection tasks, unclear priorities, and waste of resources. How to effectively combine line structure information, inspection data and task planning logic to improve the intelligence and refinement level of UAV inspection has become a technical problem that the industry urgently needs to solve.

[0004] After consulting relevant published technical solutions, a technical solution with publication number CN118758298A proposes a self-inspection dynamic planning method and system based on GPS and key point relationships, which collects power tower data to establish a tower relative position relationship model; based on the tower relative position relationship model, a UAV control system is set up to make the UAV fly along the tower trajectory in GPS mode; the model of the power tower is modeled and verified; according to the UAV control system and the power tower model, the inspection path is planned and tested by a route planning algorithm, and the UAV control system and the route planning algorithm are adjusted and optimized according to the test results; the UAV inspects according to the control of the adjusted and optimized UAV control system and route planning algorithm; this solution combines GPS and key points to automatically plan the optimal inspection path according to the key inspection points during the inspection process, ensuring the inspection quality while reducing energy consumption and prolonging the hovering time; however, this solution only relies on GPS and the relative position of the tower for path planning, lacks comprehensive assessment capability for the state of power transmission equipment and defect risks, and cannot dynamically adjust the inspection tasks and points according to the health trend of the equipment, making it difficult to achieve intelligent scheduling and route optimization based on risk priority. SUMMARY

[0005] The present application relates to the technical field of power transmission line inspection, and particularly relates to a power transmission line route automatic planning and dynamic adjustment system based on an intelligent terminal.

[0006] The present application adopts the following technical solutions:

[0007] An intelligent terminal-based power transmission line route automatic planning and dynamic adjustment system, the system comprising a data modeling module, a region evaluation module, a route planning module and a defect identification module; the data modeling module is used to construct a three-dimensional space model of the power transmission line; the region evaluation module is used to divide the power transmission line into at least one inspection region according to geographical and structural characteristics, and dynamically evaluate each inspection region to generate inspection points within the region; the route planning module is used to generate an inspection route based on the generated inspection points; and the defect identification module is used to identify defects of power transmission line equipment based on image data collected during unmanned aerial vehicle inspection.

[0008] The data modeling module comprises a point cloud data processing unit, an image registration unit, a device account synchronization unit and a three-dimensional model construction unit; the point cloud data processing unit is used to import point cloud data collected by a laser radar of an unmanned aerial vehicle, and identify and classify the power transmission line equipment structure therein; the image registration unit is used to spatially register image data taken by the unmanned aerial vehicle with the identified position of the power transmission line equipment structure; the device account synchronization unit is used to access a power equipment management terminal, and import structured information of the power transmission line equipment, the structured information including the name, coordinates and device number of the power transmission line equipment; and the three-dimensional model construction unit is used to construct an interactive three-dimensional space model of the power transmission line based on the power transmission line equipment structure classification results output by the point cloud data processing unit, the image spatial positioning information generated by the image registration unit and the structured information imported by the device account synchronization unit.

[0009] Further, the region evaluation module comprises a line region division unit, an inspection point generation unit and an inspection point activation unit; the line region division unit divides the power transmission line into at least one inspection region based on the three-dimensional space model of the power transmission line, in combination with the geographical distribution characteristics and structural layout characteristics of the power transmission line, each inspection region containing at least one power transmission line equipment structure; the inspection point generation unit is used to generate regular inspection points in combination with the spatial positions of the power transmission line equipment structures within the inspection region in the three-dimensional space model of the power transmission line, and further identify and label key inspection points in combination with the fault type distribution in the historical defect records; and the inspection point activation unit is used to dynamically activate the key inspection points in combination with the historical inspection data analysis, and output a set of inspection points to be actually executed within the current task period.

[0010] Further, the inspection point activation unit completes the prediction of the health trend of each power transmission line equipment structure through a health prediction model, and determines the key inspection points of the corresponding power transmission line equipment structure in combination with the health trend of each power transmission line equipment structure.

[0011] Further, the health prediction model is established in the following manner:

[0012] S11: Obtain historical inspection data corresponding to each power transmission line equipment structure; the historical inspection data includes an inspection timestamp and a defect identification record output by a defect identification module;

[0013] S12: Perform feature processing on the historical inspection data to construct a training set of the health prediction model; the feature processing process includes: for each piece of inspection data in the historical inspection data, performing: accessing a Deepseek model to perform semantic analysis on a natural language description in the defect identification record, and converting a semantic analysis result into a structured feature input vector as an input of the health prediction model; manually labeling a judgment result of a historical defect evolution trend of the equipment by an operation and maintenance expert to construct a health trend label corresponding to each piece of inspection data as a supervised output label of the health prediction model;

[0014] S13: Construct a health prediction model framework based on time series modeling, and train the health prediction model using the training set constructed in the previous step; a loss function of the model in the training process satisfies:

[0015]

[0016] Wherein, L is a loss function value of the model, n is the total number of training samples, ω i is a fusion weight of the i-th sample data, y i is a true label value corresponding to the i-th sample data, that is, a supervised output label of the inspection data corresponding to the sample data, is a predicted output label of the model for the i-th sample data; L base is a mean square error loss function;

[0017]

[0018] Wherein, α is a time decay factor, used to control the influence degree of the time distance on the importance of the sample; Δt i is a difference between a current training timestamp and an inspection timestamp corresponding to the i-th sample data.

[0019] Further, the inspection point activation unit sets a health trend activation threshold, and when a predicted output label output by the health prediction model corresponding to a power transmission line equipment structure is lower than the health trend activation threshold, it is determined that the power transmission line equipment structure has a failure deterioration trend, and the corresponding key inspection point of the power transmission line equipment structure is activated; the regular inspection point and all activated key inspection points are merged as a set of inspection points to be actually executed in a current task period.

[0020] Further, the route planning module generates an optimal unmanned aerial vehicle inspection route covering all inspection points based on a set of inspection points to be actually executed in the current task period, a path optimization algorithm, point space distribution characteristics and flight constraints.

[0021] Further, the defect identification module includes an image processing unit, a defect identification unit and a result output unit; the image processing unit is used for receiving image data collected by the unmanned aerial vehicle in the inspection task, and performing denoising and alignment processing on the collected image data; the defect identification unit is deployed on an image processing terminal configured with a 24G graphics memory, and performs defect identification on the processed image data through a pre-trained deep neural network model; and the result output unit is used for outputting defect identification records corresponding to each power transmission line equipment structure.

[0022] The present application has the following beneficial effects:

[0023] The present application realizes fine management and differentiated scheduling of the inspection points of the power transmission line by constructing an intelligent terminal integrating three-dimensional space modeling, health trend prediction and dynamic task adjustment, thereby dynamically generating an optimal route and focusing on high-risk equipment areas; the health prediction model is used to predict the health trend of each power transmission line equipment structure, and a time decay weight mechanism is introduced in the model training process, so that the model pays more attention to the changes of recent sample data and enhances the perception ability of the short-term evolution trend of the equipment state, thereby realizing early discovery of potential fault risks, and further improving the initiative of task scheduling and the inspection efficiency, and optimizing the real-time response and path decision effect of the inspection task. BRIEF DESCRIPTION OF DRAWINGS

[0024] The application can be further understood from the following description in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but the emphasis is placed on showing the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0025] Figure 1 It is a schematic diagram of the overall module of the present application.

[0026] Figure 2 It is a schematic diagram of the health prediction model establishment process of the present application.

[0027] Figure 3 It is a function diagram of ω under different values of the time decay factor a when the total number of fixed training samples is 100. i with Δt i

[0028] Figure 4 It is a schematic diagram of the specific update process of the time decay factor of the present application. DETAILED DESCRIPTION ​

[0029] In order to make the objectives, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with its embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application; for those skilled in the art, other systems, methods and / or features of the embodiments will become apparent after reading the following detailed description; all such additional systems, methods, features and advantages are intended to be included within the scope of the present application; included within the scope of the present application and protected by the appended claims; further features of the disclosed embodiments are described in the following detailed description, and will be apparent from the following detailed description.

[0030] The same or similar reference numerals in the drawings of the embodiments of the present application correspond to the same or similar components; in the description of the present application, it should be understood that if the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right" and the like are based on the orientations or positional relationships shown in the drawings, they are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or components referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore the terms describing the positional relationships in the drawings are only used for exemplary illustration, and cannot be understood as a limitation on the present patent, for those skilled in the art, the specific meanings of the above terms can be understood according to the specific circumstances.

[0031] Embodiment one:

[0032] As shown in Figure 1 The present embodiment provides a power transmission line route automatic planning and dynamic adjustment system based on a smart terminal, the system comprises a data modeling module, a region evaluation module, a route planning module and a defect identification module; the data modeling module is used to construct a three-dimensional space model of the power transmission line; the region evaluation module is used to divide the power transmission line into at least one inspection region according to geographical and structural characteristics, and dynamically evaluate each inspection region to generate inspection points within the region; the route planning module is used to generate an inspection route based on the generated inspection points; the defect identification module is used to identify defects of power transmission line equipment based on image data collected during unmanned aerial vehicle inspection;

[0033] The data modeling module comprises a point cloud data processing unit, an image registration unit, a device account synchronization unit and a three-dimensional model construction unit; the point cloud data processing unit is configured to import point cloud data collected by a UAV laser radar and to identify and classify the structure of power transmission line equipment in the point cloud data; the image registration unit is configured to perform spatial registration of image data captured by the UAV and the position of the identified power transmission line equipment structure; the device account synchronization unit is configured to access a power equipment management terminal, import structured information of the power transmission line equipment, and the structured information comprises the name, coordinates and equipment number of the power transmission line equipment; and the three-dimensional model construction unit is configured to construct an interactive three-dimensional space model of the power transmission line based on the classification results of the power transmission line equipment structure output by the point cloud data processing unit, the image spatial positioning information generated by the image registration unit and the structured information imported by the device account synchronization unit.

[0034] Further, the area evaluation module comprises a line area division unit, an inspection point generation unit and an inspection point activation unit; the line area division unit divides the power transmission line into at least one inspection area based on the three-dimensional space model of the power transmission line, in combination with the geographical distribution characteristics and the structural layout characteristics of the power transmission line, each inspection area containing at least one power transmission line equipment structure; the inspection point generation unit is configured to generate a regular inspection point in combination with the spatial position of the power transmission line equipment structure in the inspection area in the three-dimensional space model of the power transmission line, and further identify and label a key inspection point in combination with the fault type distribution in the historical defect record; and the inspection point activation unit is configured to dynamically activate the key inspection point in combination with the historical inspection data analysis, and output a set of inspection points to be actually executed in the current task cycle.

[0035] Further, the inspection point activation unit predicts the health trend of each power transmission line equipment structure by using a health prediction model, and determines to activate the key inspection point of the corresponding power transmission line equipment structure in combination with the health trend of each power transmission line equipment structure.

[0036] Further, as shown in Figure 2 , Figure 3 The health prediction model is established in the following manner:

[0037] S11: Obtain historical inspection data corresponding to each power transmission line equipment structure; the historical inspection data comprises an inspection timestamp and a defect identification record output by a defect identification module;

[0038] S12: Feature processing is performed on historical inspection data to construct a training set for a health prediction model; the feature processing process includes: for each piece of inspection data in the historical inspection data, performing: accessing the Deepseek model to perform semantic analysis on the natural language description in the defect identification record, extracting semantic elements including but not limited to fault type keywords, fault occurrence scene features, and severity expressions, and converting the semantic analysis result into a structured feature input vector as an input of the health prediction model, the feature input vector including but not limited to fault type code, defect level value and key description semantic embedding vector; manually labeling the judgment result of the device historical defect evolution trend by the operation and maintenance expert, and constructing a health trend label corresponding to each piece of inspection data as a supervised output label of the health prediction model;

[0039] S13: A health prediction model framework based on time series modeling is constructed, and the health prediction model is trained using the training set constructed in the previous step; the loss function of the model in the training process satisfies:

[0040]

[0041] Wherein, L is the loss function value of the model, n is the total number of training samples, ω i is the fusion weight of the i-th sample data, y i is the true label value corresponding to the i-th sample data, that is, the supervised output label of the inspection data corresponding to the sample data, is the predicted output label of the i-th sample data by the model; L base is the mean square error loss function;

[0042]

[0043] Wherein, α is a time decay factor, used to control the influence degree of time distance on the importance of the sample; Δt i is the difference between the current training timestamp and the inspection timestamp corresponding to the i-th sample data;

[0044] Further, the inspection point activation unit sets a health trend activation threshold, and when the predicted output label output by the health prediction model corresponding to a power transmission line equipment structure is lower than the health trend activation threshold, it is determined that the power transmission line equipment structure has a fault deterioration trend, and the key inspection point corresponding to the power transmission line equipment structure is activated; by merging the regular inspection points and all activated key inspection points as a set of inspection points to be actually executed in the current task period;

[0045] The scheme can realize differentiated inspection scheduling of key equipment by introducing health trend prediction and dynamic point activation mechanism, and significantly improve the utilization efficiency of inspection resources; the route generated by the path optimization algorithm ensures that the unmanned aerial vehicle task covers the high-risk area first and the overall path is optimal under the flight constraint condition, thereby improving the intelligent level, accuracy and real-time response capability of the inspection task;

[0046] Further, the route planning module generates an optimal unmanned aerial vehicle inspection route covering all inspection points based on the set of inspection points to be actually executed in the current task period, through a path optimization algorithm combined with point spatial distribution characteristics and flight constraint conditions;

[0047] Further, the defect identification module includes an image processing unit, a defect identification unit and a result output unit; the image processing unit is used to receive image data collected by the unmanned aerial vehicle in the inspection task, and to perform denoising and alignment processing on the collected image data; the defect identification unit is deployed on an image processing terminal configured with 24G graphics memory, and performs defect identification on the processed image data through a pre-trained deep neural network model; the result output unit is used to output the defect identification record corresponding to each power transmission line equipment structure.

[0048] Embodiment two:

[0049] This embodiment should be understood as at least containing all the features of any one of the preceding embodiments, and further improving on the basis thereof;

[0050] The embodiment provides a power transmission line route automatic planning and dynamic adjustment system based on an intelligent terminal, the system includes a data modeling module, a region evaluation module, a route planning module and a defect identification module; the data modeling module is used to construct a three-dimensional space model of the power transmission line; the region evaluation module is used to divide the power transmission line into at least one inspection region according to geographical and structural characteristics, and dynamically evaluate each inspection region to generate inspection points within the region; the route planning module is used to generate an inspection route based on the generated inspection points; the defect identification module is used to identify defects of power transmission line equipment based on image data collected during the unmanned aerial vehicle inspection process;

[0051] Further, to ensure the timeliness and accuracy of the health trend prediction of each power transmission line equipment structure, the health prediction model uses the newly entered inspection data to incrementally train the health prediction model when the newly entered inspection data reaches a specified update size of data volume, to ensure that the prediction results generated by the model in the subsequent task period are always based on the latest power transmission line state information;

[0052] Further, as Figure 4As shown, the incremental training of the health prediction model also includes continuous updates to the time decay factor α; the specific update method for the time decay factor α is as follows:

[0053] S21: When the amount of newly entered inspection data in the system reaches the specified update scale, the actual defects of all activated key inspection points are verified by calculating the hit rate index. The specific calculation method of the hit rate index is as follows:

[0054]

[0055] Among them, Hit j K represents the hit rate indicator corresponding to the j-th inspection data in the newly entered inspection data of the system. j ′ represents the number of key inspection points with defects identified in the inspection route corresponding to the j-th inspection data; K j This represents the total number of key inspection points activated in the inspection route corresponding to the j-th inspection data.

[0056] S22: Adaptively update the time decay factor based on the hit rate of newly entered inspection data in the system.

[0057]

[0058] Where, α t+1 This is the updated time decay factor, α, which will be used in subsequent incremental training of the health prediction model. t+1 The fusion weights of the sample data are calculated as a time decay factor; α t H represents the time decay factor in the system before the update, η is the preset learning rate, and H... target The hit rate indicator is a pre-set value for the system, with a range of (0.5, 1); m is the total number of newly entered inspection data entries in the system.

[0059] This solution introduces a hit rate feedback mechanism to achieve adaptive dynamic adjustment of the time decay factor α, thereby improving the model's ability to correct actual prediction deviations. Based on the actual recognition effect, the model's weight allocation for new and old samples is adjusted to make the health trend prediction results closer to the current real state of the device. This maintains the model's long-term learning stability while enhancing its short-term response sensitivity, thereby improving the overall intelligence level and prediction accuracy of the system.

[0060] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. A power transmission line route automatic planning and dynamic adjustment system based on a smart terminal, characterized in that, The system includes a data modeling module, a regional assessment module, a route planning module, and a defect identification module. The data modeling module is used to construct a three-dimensional spatial model of the transmission line. The regional assessment module is used to divide the transmission line into at least one inspection area based on geographical and structural characteristics, and dynamically assess each inspection area to generate inspection points within the area. The route planning module is used to generate inspection routes based on the generated inspection points. The defect identification module is used to identify defects in power transmission line equipment based on image data collected during UAV inspections. The data modeling module includes a point cloud data processing unit, an image registration unit, an equipment ledger synchronization unit, and a 3D model construction unit. The point cloud data processing unit is used to import point cloud data collected by UAV lidar and to identify and classify the power transmission line equipment structure therein. The image registration unit is used to spatially register the image data captured by the UAV with the identified power transmission line equipment structure location. The equipment ledger synchronization unit is used to connect to the power equipment management terminal and import the structured information of the transmission line equipment. The structured information includes the name, coordinates and equipment number of the transmission line equipment. The three-dimensional model building unit is used to build an interactive three-dimensional spatial model of the transmission line based on the transmission line equipment structure classification results output by the point cloud data processing unit, the image spatial positioning information generated by the image registration unit, and the structured information imported by the equipment ledger synchronization unit. The regional assessment module includes a line area division unit, an inspection point generation unit, and an inspection point activation unit. The line area division unit, based on a three-dimensional spatial model of the transmission line, combines the geographical distribution and structural layout characteristics of the transmission line to spatially divide the transmission line and generate at least one inspection area. Each inspection area contains at least one transmission line equipment structure. The inspection point generation unit is used to generate regular inspection points by combining the spatial location of the transmission line equipment structures within the inspection area in the three-dimensional spatial model of the transmission line, and further identify and mark key inspection points by combining the distribution of fault types in historical defect records. The inspection point activation unit is used to dynamically activate key inspection points by combining historical inspection data analysis, and output the set of inspection points that need to be actually executed within the current task cycle. The inspection point activation unit uses a health prediction model to predict the health trend of each transmission line equipment structure, and combines the health trend of each transmission line equipment structure to determine and activate the key inspection points of the corresponding transmission line equipment structure.

2. The automatic planning and dynamic adjustment system for power transmission line routes based on intelligent terminals according to claim 1, characterized in that, The health prediction model is established as follows: S11: Obtain historical inspection data corresponding to the equipment structure of each transmission line; the historical inspection data includes inspection timestamps and defect identification records output by the defect identification module; S12: Perform feature processing on historical inspection data to build a training set for the health prediction model; The feature processing process includes: for each inspection data in the historical inspection data, the following steps are performed: the Deepseek model is accessed to perform semantic analysis on the natural language description in the defect identification record, and the semantic analysis results are converted into a structured feature input vector as the input of the health prediction model; the judgment results of the historical defect evolution trend of the equipment are manually labeled, and a health trend label corresponding to each inspection data is constructed as the supervision output label of the health prediction model. S13: Construct a health prediction model framework based on time series modeling, and train the health prediction model using the training set constructed in the previous step; the loss function of the model during training satisfies: ； in, The loss function value of the model. The total number of training samples, For the first The fusion weights of the sample data For the first The true label value corresponding to each sample data point is the supervisory output label of the inspection data corresponding to that sample data point. For the model to the first Predicted output labels for each sample data point; Let the mean squared error loss function be used. ； in, This is the time decay factor, used to control the degree of influence of time distance on the importance of the sample; For the current training timestamp and the first The difference between the timestamps of each sample data point and the inspection timestamp.

3. The automatic planning and dynamic adjustment system for power transmission line routes based on intelligent terminals according to claim 1, characterized in that, The inspection point activation unit sets a health trend activation threshold, and when the predicted output label value of the health prediction model corresponding to a certain transmission line equipment structure is lower than the health trend activation threshold, it determines that the transmission line equipment structure has a fault deterioration trend and activates the key inspection points corresponding to the transmission line equipment structure. By merging regular inspection points with all activated key inspection points, a set of inspection points that need to be actually executed within the current task cycle is formed.

4. The automatic planning and dynamic adjustment system for power transmission line routes based on intelligent terminals according to claim 1, characterized in that, The route planning module generates the optimal UAV inspection route covering all inspection points based on the set of inspection points that need to be actually executed within the current task cycle, using a path optimization algorithm that combines the spatial distribution characteristics of the inspection points with flight constraints.

5. The automatic planning and dynamic adjustment system for power transmission line routes based on a smart terminal according to claim 1, characterized in that, The defect identification module includes an image processing unit, a defect identification unit, and a result output unit; the image processing unit is used to receive image data collected by the UAV during the inspection mission, and to perform noise reduction and alignment processing on the collected image data; The defect identification unit is deployed on an image processing terminal equipped with 24G of video memory. It performs defect identification on the processed image data through a pre-trained deep neural network model. The result output unit is used to output the defect identification records corresponding to the structure of each transmission line equipment.

Images