Power transmission line intelligent inspection robot path optimization method and system and medium

Abstract

The invention discloses a power transmission line intelligent inspection robot path optimization method and system and a medium, and relates to the technical field of path optimization, and the method comprises the steps: obtaining time sequence environment perception data and called power transmission line structured geographic data, and constructing a three-dimensional channel environment model; introducing an energy constraint, and calculating an initial global path connected with a plurality of preset inspection task points by taking the weighted sum of the total path length and the estimated total energy consumption as an optimization target; analyzing an inspection target and a task of a preset inspection task point, establishing a local inspection optimization path, and identifying coordinates of nodes capable of being connected and constraint attitude information as local optimization information; and performing consistency verification and path optimization correction on the initial global path by using local optimization information to obtain a final inspection path. The technical problem that in the prior art, energy consumption is high due to the fact that energy constraints are not fully considered in routing inspection robot path planning is solved, and the technical effect that the overall energy consumption of the routing inspection robot is reduced while the routing inspection task requirement is met is achieved.

Description

Technical Field

[0001] This invention relates to the field of path optimization technology, specifically to a path optimization method, system, and medium for intelligent inspection robots of power transmission lines. Background Technology

[0002] In power transmission line inspection operations, inspection robots typically need to operate continuously over long distances, in complex terrains, and through obstacles. Path planning directly affects their operation time and energy efficiency. Existing path planning methods focus more on geometric distance or accessibility, and do not adequately consider the energy consumption characteristics of robots under different terrain slopes, operating postures, and inspection actions. This can easily lead to situations where the path is reachable but the energy consumption is too high, thus shortening the single inspection endurance and limiting the coverage and efficiency of the inspection task. Summary of the Invention

[0003] This application provides a path optimization method, system, and medium for intelligent inspection robots of transmission lines, which is used to address the technical problem of high energy consumption caused by insufficient consideration of energy constraints in the path planning of inspection robots in the prior art.

[0004] In view of the above problems, this application provides a path optimization method, system and medium for intelligent inspection robots of transmission lines.

[0005] The first aspect of this application provides a path optimization method for an intelligent inspection robot of transmission lines, the method comprising: The system acquires temporal environmental perception data collected by the lidar and vision sensors mounted on the inspection robot, as well as structured geographic data of the power transmission line, to construct a three-dimensional channel environment model that includes terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. Based on the three-dimensional channel environment model, energy constraints are introduced, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The inspection objectives and tasks of the preset inspection task points are analyzed to establish local inspection optimization paths for the task points, and the coordinates and constraint posture information of connectable nodes are identified as local optimization information. The initial global path is then used to perform consistency verification and path optimization correction using the local optimization information to obtain the final inspection path.

[0006] A second aspect of this application provides a path optimization system for intelligent inspection robots of transmission lines, the system comprising: The environment model construction module is used to acquire time-series environmental perception data collected by the LiDAR and vision sensors mounted on the inspection robot, as well as the structured geographic data of the power transmission line, to construct a three-dimensional channel environment model that includes terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. The calculation module is used to introduce energy constraints based on the three-dimensional channel environment model, and calculate the initial global path connecting multiple preset inspection task points with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The parsing module is used to parse the inspection objectives and tasks of the preset inspection task points, establish local inspection optimization paths for the task points, and identify the coordinates and constraint posture information of connectable nodes as local optimization information. The correction module is used to use the local optimization information to perform consistency verification and path optimization correction on the initial global path to obtain the final inspection path.

[0007] A third aspect of the embodiments of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the path optimization method for intelligent inspection robot of transmission lines provided in this application.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application acquires temporal environmental perception data collected by LiDAR and vision sensors mounted on an inspection robot, and calls upon structured geographic data of power transmission lines to construct a three-dimensional channel environment model including terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. Based on the three-dimensional channel environment model, energy constraints are introduced, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The inspection objectives and tasks of the preset inspection task points are analyzed, and local inspection optimization paths for the task points are established. The coordinates and constraint posture information of connectable nodes are identified as local optimization information. The initial global path is validated for consistency and optimized and corrected using the local optimization information to obtain the final inspection path. This invention solves the technical problem of high energy consumption caused by insufficient consideration of energy constraints in the path planning of inspection robots in the prior art. By introducing energy constraints on the basis of the three-dimensional channel environment model and optimizing and correcting the inspection path, the technical effect of reducing the overall energy consumption of the inspection robot while meeting the requirements of the inspection task is achieved. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of the path optimization method for intelligent inspection robot of transmission lines provided in the embodiments of this application; Figure 2 This is a schematic diagram of the path optimization system for intelligent inspection robots of power transmission lines provided in an embodiment of this application.

[0011] Figure labeling: Environment model construction module 11, calculation module 12, analysis module 13, correction module 14. Detailed Implementation

[0012] This application provides a path optimization method, system, and medium for intelligent inspection robots of transmission lines. It addresses the technical problem of high energy consumption caused by insufficient consideration of energy constraints in the path planning of inspection robots in the prior art. By introducing energy constraints on the basis of a three-dimensional channel environment model and optimizing and correcting the inspection path, it achieves the technical effect of reducing the overall energy consumption of the inspection robot while meeting the requirements of the inspection task.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides a path optimization method for intelligent inspection robots of transmission lines, the method comprising: Step S100: Obtain the temporal environmental perception data collected by the lidar and vision sensors mounted on the inspection robot, as well as the structured geographic data of the power transmission line, and construct a three-dimensional channel environment model that includes terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment.

[0016] In this embodiment, during the inspection process, a lidar mounted on an inspection robot continuously scans the power transmission line corridor to acquire lidar point cloud data reflecting changes in the spatial structure of the surrounding environment. Simultaneously, a visual sensor continuously acquires images of the corridor environment to obtain visual sensing data reflecting environmental appearance information. The lidar point cloud data and visual sensing data are arranged in chronological order of acquisition to form temporal environmental perception data. Furthermore, by accessing pre-stored structured geographic data of the power transmission line, information related to the line's alignment, tower locations, and foundation elevation is obtained.

[0017] Next, a three-dimensional tunnel environment model is constructed. This process begins by acquiring continuous temporal laser point cloud data and visual sensor data of the tunnel. These data are then temporally aligned, denoised, and spatially registered to achieve consistent representation of multi-source sensing data across time and spatial coordinates. Subsequently, image recognition is performed based on the visual sensor data to extract semantic information about tunnel obstacles and power transmission equipment. This semantic information is then mapped onto the laser point cloud data, and the point cloud is semantically annotated to form three-dimensional point cloud data with semantic attributes. Finally, structured geographic data of the power transmission line is accessed through a geographic information system (GIS). The three-dimensional point cloud data with semantic attributes is spatially registered and fused with the structured geographic data to generate a three-dimensional tunnel environment model containing terrain elevation information, semantic labels for tunnel obstacles, and structural features of power transmission equipment.

[0018] Furthermore, the method provided in the application embodiments, which constructs a three-dimensional channel environment model including terrain elevation information, semantic tags of channel obstacles, and structural features of power transmission equipment, also includes: The system acquires continuous temporal laser point cloud data and visual sensing data of the transmission line, and performs temporal alignment, denoising, and spatial registration processing on the laser point cloud data and visual sensing data. Based on the visual sensing data, image recognition is performed to generate semantic recognition information of obstacles and power transmission equipment in the transmission line. The semantic recognition information is used to semantically annotate the laser point cloud data to obtain three-dimensional point cloud data with semantic attributes. By calling the structured geographic data of the power transmission line through a geographic information system, the three-dimensional point cloud data with semantic attributes is spatially registered and fused with the structured geographic data to generate the three-dimensional transmission line environment model.

[0019] In this embodiment, the inspection robot continuously scans the channel area using its onboard LiDAR, acquiring spatial distance information of various reflective targets in the environment based on the laser time-of-flight ranging principle, forming laser point cloud data arranged in chronological order of acquisition. Simultaneously, a visual sensor continuously acquires images of the same channel area, obtaining visual sensing data reflecting the appearance characteristics of the environment. Both the laser point cloud data and the visual sensing data are recorded using a unified timestamp, constituting continuous temporal data of the channel. Subsequently, timestamp matching is used to align the laser point cloud data and the visual sensing data to ensure that data acquired by different sensors at the same time correspond to each other. Statistical filtering is then used to denoise the laser point cloud data to eliminate outliers caused by measurement errors. Finally, spatial registration is performed on the laser point cloud data and the visual sensing data based on the extrinsic parameter matrix obtained from sensor calibration, mapping the multi-source data to the same spatial coordinate system.

[0020] Next, a pre-trained target detection and semantic segmentation model is introduced based on visual sensing data. This model, trained offline on typical scene samples of power transmission line corridors, is used to perform pixel-by-pixel semantic discrimination on the input image. The model then performs image recognition processing on the visual sensing data, generating semantic recognition information for corridor obstacles (describing object categories such as ground, vegetation, and foreign objects) and semantic recognition information for power transmission equipment (describing structural attributes such as towers, conductors, and insulators). This semantic recognition information characterizes the semantic categories of different targets in the corridor environment. Subsequently, based on the known extrinsic parameter relationship between the visual sensor and the lidar, the semantic recognition information for corridor obstacles and power transmission equipment is mapped point-by-point to the corresponding lidar point cloud data. The spatial points in the point cloud are assigned corresponding semantic category attributes, completing the semantic annotation of the lidar point cloud data. This yields 3D point cloud data that simultaneously contains 3D spatial coordinate information and semantic attribute information.

[0021] After obtaining 3D point cloud data with semantic attributes, the structured geographic data of the transmission line is accessed through a geographic information system (GIS). This structured geographic data is pre-collected and maintained by the power operation and maintenance system, storing line alignment, tower locations, and terrain elevation information according to a unified geographic coordinate reference system to provide a global spatial benchmark. Subsequently, a two-dimensional plane translation and yaw angle correction method based on control points is used to achieve spatial registration and fusion. Specifically, firstly, point cloud regions of at least two towers are extracted from the 3D point cloud data with semantic attributes based on the semantic recognition information of the transmission equipment, and the coordinates of the center point of each tower's point cloud region on the horizontal plane are calculated as point cloud-side control points. Simultaneously, the geographic coordinates of the towers corresponding one-to-one are read from the structured geographic data, and their projected coordinates on the horizontal plane are taken as geographic-side control points. Then, the yaw angle deviation of the point cloud-side coordinate system relative to the geographic coordinate reference system is calculated using the azimuth difference between the two sets of control points, and the 3D point cloud data with semantic attributes is then subjected to vertical correction. The rotation correction of the axis is performed. Then, after the rotation correction is completed, the coordinate difference between the control points on the point cloud side and the control points on the geographic side on the horizontal plane is calculated to obtain the translation vector. The translation vector is then applied to the rotated and corrected 3D point cloud data with semantic attributes to complete the position alignment. Finally, the route alignment and terrain elevation information in the structured geographic data are fused with the semantic labels of passage obstacles and the structural features of power transmission equipment in the 3D point cloud data with semantic attributes. This allows for spatial registration and fusion of the 3D point cloud data with semantic attributes and the structured geographic data to generate a 3D passage environment model.

[0022] Step S200: Based on the three-dimensional channel environment model, energy constraints are introduced, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective.

[0023] In this embodiment, when performing global path planning based on a 3D channel environment model, energy constraint information is first constructed based on the historical operation data of the inspection robot, and the energy timing of the inspection robot is dynamically constrained. Then, using the 3D channel environment model as the spatial base for path planning, a path search space containing channel terrain and obstacle constraints is constructed, and the energy constraint information is introduced into the path search space to predict and evaluate the energy consumption of each candidate path during the path search process. Next, multiple preset inspection task points are mapped into the 3D channel environment model, and the weighted sum of the total path length and the estimated total energy consumption is used as the optimization objective. Path search and energy consumption evaluation are performed through the path search space, and the inspection path with the optimal optimization objective is selected as the initial global path connecting the preset inspection task points.

[0024] Furthermore, in the method provided in the application embodiment, energy constraints are introduced based on the three-dimensional channel environment model, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The method further includes: Based on the operational data of the inspection robot, energy constraint information is constructed, and the energy timing of the inspection robot is dynamically constrained and configured. Using the three-dimensional channel environment model as a base, a path search space is constructed, and the energy constraint information is added to the path search space to predict and evaluate the energy consumption in path optimization. The preset inspection task points are projected onto the three-dimensional channel environment model, and the weighted sum of the total path length and the estimated total energy consumption is used as the optimization objective. Path search and energy consumption evaluation are performed through the path search space to obtain the inspection path that maximizes the optimization objective, which is used as the initial global path.

[0025] In this embodiment, when constructing energy constraint information based on the inspection robot's operational data, the energy consumption time series and corresponding operational posture data reflecting the operation process are first extracted from the inspection robot's historical operation records. Then, based on the energy consumption time series and operational posture data, a constraint relationship between time-series energy consumption and operational posture is established to characterize the correlation between operational posture and energy consumption changes at different times. Subsequently, the constraint relationship between time-series energy consumption and operational posture is parameterized to form energy constraint information, which is used to describe the correspondence between energy consumption and operational posture during the inspection robot's operation, as well as the energy constraint conditions under different operational postures.

[0026] After obtaining the energy constraint information, the energy timing of the inspection robot is dynamically constrained and configured, and the energy consumption is updated along the path using a segment-by-segment accumulation method. Specifically, the upper limit of the available energy of the inspection robot is first set to E. max The remaining energy at the starting point of the path is initialized to E0=E max Subsequently, when the path moves from one node to the next, this movement is considered as a path segment k. Attitude parameters (such as slope, steering, and speed) are extracted from the corresponding running attitude of this path segment and substituted into the energy consumption coefficient per unit distance given by the energy constraint information to calculate the energy consumption c per unit distance for this path segment. k Then calculate the spatial distance d of the path segment. k Thus, the energy consumption e of this path segment can be obtained. k =c k ×d k Then update the energy time series so that the remaining energy is recursively calculated according to the time series. and E k ≥0 is used as a dynamic constraint to eliminate candidate path segments that would lead to energy overdraft during the path planning process.

[0027] In global path planning, a 3D channel environment model is used as the base to construct the path search space. A rasterization discretization method is employed to divide the 3D channel environment model into several walkable units on a horizontal plane. Each walkable unit corresponds to a path node, and path edges between nodes are established based on the connectivity between adjacent grids, thus forming a path search space usable for searching. Subsequently, energy constraint information is added to the path search space, so that each path edge simultaneously possesses two attributes: distance cost and energy cost. The distance cost is calculated from the Euclidean distance between adjacent nodes, and the energy cost is calculated cumulatively according to the aforementioned step-by-step accumulation method. k This allows for the prediction and evaluation of energy consumption during path optimization, and the utilization of remaining energy E during the search process. k ≥0 constrains and filters candidate expansion nodes.

[0028] When determining the necessary points along the path, multiple pre-set inspection task points are projected onto the 3D channel environment model. The nearest neighbor mapping method is used to select the corresponding node in the path search space for each pre-set inspection task point. Specifically, the spatial distance between the pre-set inspection task point and all path nodes is calculated, and the node with the smallest distance is selected as the projection node for that pre-set inspection task point. Then, using the weighted sum of the total path length and the estimated total energy consumption as the optimization objective, candidate paths connecting multiple pre-set inspection task points are searched and evaluated. In this process, the total path length L = ∑d is first calculated for any candidate path. k Based on the aforementioned segmented energy consumption calculation, the estimated total energy consumption E=∑e is obtained. k Then, weighting coefficients α and β are set, and the optimization objective value J = αL + βE is constructed, where α is used to adjust the importance of the total path length and β is used to adjust the importance of the estimated total energy consumption. The optimization objective value J is calculated for different candidate paths in the path search space, and under the premise of satisfying the dynamic energy time-series constraint configuration, the inspection path with the optimal optimization objective value is selected as the initial global path connecting multiple preset inspection task points.

[0029] Furthermore, the method provided in the application embodiments, which constructs energy constraint information based on the operating data of the inspection robot, further includes: Based on the historical operation records of the inspection robot, extract the energy consumption time series and operation posture data of the inspection robot; based on the energy consumption time series and operation posture data, establish the constraint relationship between time series energy consumption and operation posture; use the constraint relationship between time series energy consumption and operation posture to construct the energy constraint information, which is used to describe the relationship between energy consumption and operation posture during the operation of the inspection robot and the energy constraint conditions corresponding to the operation posture.

[0030] In this embodiment, during the construction of the energy constraint information for the inspection robot, the original operating information is first obtained from the historical operating record data generated by the inspection robot in previous inspection operations. The historical operating record data continuously records the changes in battery power, motor current, and voltage of the inspection robot during operation, indexed by timestamps. Based on the historical operating record data, the battery power changes are differentially calculated according to the time sequence to obtain the energy consumption time series characterizing the energy change per unit time. The energy consumption time series is used to reflect the energy consumption pattern of the inspection robot during continuous operation. Simultaneously, the operating posture data corresponding to each time point is extracted from the historical operating record data. The operating posture data is collected by posture sensors and is used to characterize the motion posture parameters of the inspection robot at each moment, including travel speed, turning angle, and slope status.

[0031] After obtaining the energy consumption time series and operating posture data, the two types of data are aligned by time index to form a one-to-one corresponding data sample set. A grouped statistical analysis method is then used to analyze the energy consumption characteristics under different operating posture conditions. Specifically, the operating posture data is classified according to preset posture intervals, and the mean of the corresponding energy consumption time series within each posture interval is calculated to obtain the average energy consumption level under specific operating posture conditions, thereby establishing a mapping relationship describing the change in energy consumption with operating posture. Through the above processing, a constraint relationship between time series energy consumption and operating posture is formed to characterize the energy consumption variation law of the inspection robot under different operating posture conditions. This constraint relationship is expressed in the form of operating posture parameter - energy consumption per unit time.

[0032] After obtaining the constraint relationship between time-series energy consumption and operating posture, the constraint relationship is organized into a set of parameters that can be used in the path planning process. The energy consumption parameters corresponding to each operating posture interval are uniformly encoded to construct complete energy constraint information. This energy constraint information is used during path planning and energy assessment to directly retrieve the corresponding energy consumption constraint parameters based on the current or predicted operating posture of the inspection robot. This enables a quantitative description of the relationship between energy consumption and operating posture during the inspection robot's operation, as well as effective constraints on the energy consumption range under different operating posture conditions.

[0033] Step S300: Analyze the inspection targets and tasks of the preset inspection task points, establish local inspection optimization paths for the task points, and identify the coordinates and constraint attitude information of connectable nodes as local optimization information.

[0034] In this embodiment, when parsing the inspection targets and tasks of preset inspection task points, the inspection equipment targets and task types corresponding to each preset inspection task point are first parsed. Combined with the current operating status information of the task point obtained by the edge computing module deployed in the area corresponding to the task point, the specific execution requirements of the inspection task are analyzed. Then, based on the parsed inspection task requirements, the core inspection node used by the task point to meet the inspection target and its corresponding constraint task acquisition parameters are determined, as well as the auxiliary inspection node associated with the core inspection node and its corresponding optional acquisition parameters. Next, based on the core inspection node and its constraint task acquisition parameters, and the auxiliary inspection node and its optional acquisition parameters, the channel path related to the task point is mapped to the three-dimensional channel environment model. The channel path coordinates of the task point are projected, and the path parameters are configured according to the constraint strength of the acquisition parameters. Subsequently, based on the three-dimensional channel environment model, constraint acquisition paths that meet the acquisition parameter constraint requirements are identified, and path node coherence constraint analysis is performed on the constraint acquisition paths to determine connectable nodes that can be connected to other path segments and obtain the corresponding connectable node coordinates. Simultaneously, the attitude parameters in the constrained acquisition path are classified into levels based on the constraint strength of the acquired parameters, forming multi-level constraint attitude information. Finally, the constrained acquisition path, the coordinates of connectable nodes, and the multi-level constraint attitude information are combined as local optimization information reflecting the local execution characteristics of the inspection task points.

[0035] Furthermore, the method provided in the application embodiment, which parses the inspection target and task of the preset inspection task point, establishes a local inspection optimization path for the task point, and identifies the coordinates and constraint posture information of connectable nodes, also includes: The system analyzes the inspection equipment targets and task types at each preset inspection task point, and analyzes the inspection task requirements based on the current operating status information of the task point obtained by the edge computing module deployed in the corresponding area of ​​the task point. Based on the analyzed inspection task requirements, it determines the core inspection node of the task point and its corresponding constraint task acquisition parameters, and determines the auxiliary inspection node associated with the core inspection node and its corresponding optional acquisition parameters. Based on the core inspection node, constraint task acquisition parameters, auxiliary inspection nodes, and optional acquisition parameters, it projects the channel path coordinates of the task point and configures the parameters according to the constraint strength of the acquisition parameters. Based on the three-dimensional channel environment model, it identifies the constraint acquisition path of the task point, performs path node coherence constraint analysis based on the constraint acquisition path, determines connectable nodes and obtains the coordinates of the corresponding connectable nodes, and classifies the attitude parameters of the acquisition path according to the constraint strength of the acquisition parameters, setting multi-level constraint attitude information.

[0036] In this embodiment, when processing inspection task points, the corresponding inspection equipment target and task type are first read from the inspection task configuration for each preset inspection task point. The inspection equipment target clarifies the specific power transmission equipment location to be inspected, and the task type distinguishes different inspection operation methods such as photo capture, status detection, or detailed observation. Simultaneously, the current operating status information of the area where the task point is located is obtained. This information includes whether the equipment is energized, whether the passage is unobstructed, and whether the environmental conditions meet the inspection requirements. The current operating status information is then compared item by item with the inspection equipment target and task type, and the specific inspection task requirements that the inspection task needs to meet in the current state are determined through conditional judgment.

[0037] After analyzing the requirements of the inspection tasks, spatial locations that directly meet the requirements of the inspection tasks are selected from the 3D channel environment model based on the requirements of different inspection tasks for acquisition accuracy and observation angle. At least one core inspection node is determined using distance matching and viewpoint matching methods. This core inspection node serves as the primary acquisition location for completing the inspection tasks. Subsequently, based on the minimum acquisition quality requirements of the inspection tasks, constraint acquisition parameters that must be strictly met are configured for the core inspection node. These constraint acquisition parameters include a fixed acquisition distance range, a limited acquisition angle range, and attitude stability requirements. Simultaneously, several alternative acquisition locations are selected around the core inspection node according to spatial adjacency relationships to serve as auxiliary inspection nodes associated with the core inspection node. Optional acquisition parameters that can be adjusted within a certain range are configured for these auxiliary inspection nodes to provide more feasible options during path planning.

[0038] After determining the core inspection nodes, auxiliary inspection nodes, and their corresponding acquisition parameters, the core inspection nodes, constrained task acquisition parameters, auxiliary inspection nodes, and optional acquisition parameters are uniformly mapped into the three-dimensional channel environment model, and the channel path coordinates corresponding to the task points are projected. Subsequently, parameters are configured according to the constraint strength of the acquisition parameters. Path nodes corresponding to constrained task acquisition parameters are marked as high constraint levels, and path nodes corresponding to optional acquisition parameters are marked as low constraint levels. This allows different path nodes to have different constraint priorities and adjustment space during path planning, thereby reflecting differentiated constraints of acquisition requirements at the channel path level.

[0039] Next, when identifying the constraint acquisition path for task points based on the 3D channel environment model, the relevant information of the task nodes is first digitally represented using the 3D channel environment model. The inspection equipment targets, inspection node location distribution, acquisition parameters, and their corresponding constraint strengths corresponding to each task node are uniformly mapped into calculable digital parameters. Then, based on the digital parameters of the task nodes, different constraint levels are divided according to the constraint strength of the acquisition parameters. Multi-level inspection path search is performed in the 3D channel environment model. During the path search process, candidate paths are analyzed segment by segment based on the continuity constraints of the acquisition parameters, and paths that can continuously meet the acquisition parameter constraint requirements are selected as constraint acquisition paths. Afterwards, path node continuity constraint analysis is performed on the constraint acquisition paths, identifying indivisible path intervals that must be continuously executed and cannot be interrupted during the acquisition process in the multi-level inspection paths. Based on these indivisible path intervals, connectable nodes that can serve as the start or end point of the path are extracted from the multi-level inspection paths. Connectable nodes include any node excluding the indivisible path interval and the starting node of the indivisible path interval, and the corresponding coordinates of the connectable nodes are extracted.

[0040] After obtaining the coordinates of connectable nodes, the attitude control requirements in the constrained acquisition path are refined, and the attitude parameters of the acquisition path are classified into levels according to the constraint strength of the acquisition parameters, setting multi-level constraint attitude information. Specifically, based on the path nodes with completed parameter configurations, the acquisition parameters associated with the path nodes are used as the direct basis for attitude constraints. The corresponding set of attitude parameters is extracted for each path node, and the attitude parameters are classified according to the constraint strength of the acquisition parameters associated with that node. Path nodes corresponding to acquisition parameters with high constraint levels are classified as high-strength constraint levels, requiring strict adherence to a predetermined attitude range during path execution. Path nodes corresponding to acquisition parameters with low constraint levels are classified as low-strength constraint levels, allowing attitude adjustments within a preset tolerance range. When the same path node is associated with multiple acquisition parameters, the acquisition parameter with the highest constraint strength is used as the basis for determining the attitude level of that node. Through the above processing, a clear multi-level attitude constraint division is formed on the constraint acquisition path, and the attitude constraints of each level are uniformly organized into multi-level strength constraint attitude information. This information is used in the subsequent connection and consistency verification of global and local paths to accurately control the attitude constraint range of different path nodes, ensuring the acquisition quality of key inspection nodes while taking into account the flexibility of path optimization.

[0041] Furthermore, in the method provided in the application embodiments, the constraint acquisition path for identifying task points based on the three-dimensional channel environment model further includes: The three-dimensional channel environment model digitally represents the inspection equipment target, inspection node location distribution, acquisition parameters and their constraint strength of the task node; based on the digital parameters of the task node, a multi-level inspection path search is performed according to the constraint strength of the acquisition parameters. In the search inspection path, the path that meets the continuity requirement is identified and screened according to the continuity constraint analysis of the acquisition parameters, and the constrained acquisition path is obtained.

[0042] In this embodiment, when identifying the constraint acquisition path of task points, the relevant information of task nodes is first digitally represented based on a three-dimensional channel environment model. Specifically, the inspection equipment target corresponding to each task node is converted into clear target identification information to distinguish different types of power transmission equipment. The distribution of inspection node positions of each task node is represented as spatial coordinates in the three-dimensional channel environment model to determine the specific position of the task node in the channel. At the same time, the acquisition parameters corresponding to the inspection task are associated with the corresponding inspection node in numerical or interval form, and the corresponding constraint strength is marked for each acquisition parameter to indicate the strictness to which the acquisition parameter needs to be satisfied during the path planning process, thereby forming a digital description of the task node that can directly participate in path calculation.

[0043] After completing the digital representation of the task nodes, a multi-level inspection path search is performed in the 3D channel environment model based on the digital parameters of the task nodes and the constraint strength of the acquisition parameters. Specifically, the acquisition parameter with the highest constraint strength is first used as the primary condition for path search. Path nodes that can meet the requirements of this high-constraint acquisition parameter are selected in the 3D channel environment model, and path expansion is only performed between these path nodes. Subsequently, while ensuring that the high-constraint acquisition parameter is continuously satisfied, acquisition parameters with lower constraint strength are introduced to participate in path expansion. This allows the path search to gradually increase the range of optional paths while meeting key acquisition requirements, thereby forming multiple candidate inspection paths under different constraint levels.

[0044] During the inspection path search process, a continuity constraint analysis of the acquisition parameters is performed on candidate inspection paths. In this process, along the travel direction of the candidate path, the acquisition parameters associated with adjacent path nodes are compared one by one to determine whether the changes in acquisition distance, acquisition angle, and attitude between adjacent nodes are within the allowable continuous variation range of the acquisition parameters. When the changes in acquisition parameters between adjacent path nodes exceed the allowable range, the path segment is determined not to meet the continuity constraint, and the candidate path containing this path segment is removed. When the changes in acquisition parameters between adjacent path nodes are within the allowable range, the path segment is determined to meet the continuity constraint, and the path segment is retained. By performing continuity screening on each segment of the candidate inspection paths, the path that can continuously meet the acquisition parameter constraint requirements throughout the entire path progression is finally identified and selected, and this path is determined as the constrained acquisition path.

[0045] Furthermore, in the method provided in the application embodiments, the method further includes performing path node coherence constraint analysis based on the constraint acquisition path to determine connectable nodes and obtain the coordinates of the corresponding connectable nodes: In a multi-level inspection path, a coherent constraint analysis is performed on the constraint acquisition path to identify indivisible path intervals. Based on these indivisible path intervals, connectable nodes are extracted from the multi-level inspection path. A connectable node is any node excluding the indivisible path interval, or the starting node of the indivisible path interval. The connectable node can serve as the starting or ending point of the path for connecting with other paths. Based on the determined connectable nodes, the corresponding spatial coordinates are extracted to obtain the coordinates of the connectable nodes.

[0046] In this embodiment, within a multi-level inspection path, a path node coherence constraint analysis is first performed on the determined constraint acquisition path. During this process, adjacent path nodes are sequentially selected according to their order in the constraint acquisition path. The acquisition parameters and attitude parameters corresponding to each node are read, and the changes in acquisition distance, acquisition angle, and attitude angle between the two nodes are calculated. These changes are then compared to the maximum allowable change threshold for the acquisition parameters. When the change between adjacent path nodes exceeds the allowable threshold, it is determined that the acquisition process between the two nodes cannot be interrupted and must be executed continuously. By performing the above judgment on all adjacent path nodes in the constraint acquisition path, all adjacent nodes that must be executed continuously and cannot be interrupted are combined to form one or more continuous path segments. Each such continuous path segment is defined as an indivisible path interval, representing the path range that must be continuously traversed during the inspection process.

[0047] After identifying indivisible path intervals, connectable nodes are extracted from the multi-level inspection paths based on these intervals. This process involves first traversing all path nodes within the multi-level inspection paths and determining whether each node is located within any indivisible path interval. If a path node is within an indivisible path interval but is not its starting node, it is marked as unusable for path connection. If a path node does not belong to any indivisible path interval, it is marked as usable for path connection. Additionally, for each indivisible path interval, its starting node is also marked as usable for path connection. Through this judgment and marking process, path nodes that neither disrupt the continuity of indivisible path intervals nor prevent connection to other path segments are identified and designated as connectable nodes. These connectable nodes serve as the start or end point of a path in subsequent path connections.

[0048] After identifying the connectable nodes, the spatial position data of each connectable node in the 3D channel environment model is read one by one according to the correspondence between the connectable nodes and the model coordinate system. The 3D coordinates of the node in the model coordinate system are then extracted, thus obtaining the coordinates of each connectable node. Through the above steps, the identification of indivisible path intervals in the constrained acquisition path, the determination of connectable nodes, and the acquisition of their coordinates are completed.

[0049] Step S400: Use the local optimization information to perform consistency verification and path optimization correction on the initial global path to obtain the final inspection path.

[0050] In this embodiment, when using local optimization information to perform consistency verification and path optimization correction on the initial global path, firstly, based on the multi-level constraint attitude information and multi-level inspection path in the local optimization information, path nodes are matched with the initial global path to determine the corresponding matching constraint attitude, connectable nodes, and constraint acquisition path in the initial global path. Then, combining the matching constraint attitude, connectable nodes, and constraint acquisition path, and introducing the aforementioned constructed energy constraint information, constraint consistency verification is performed on the initial global path to determine whether each path segment in the initial global path simultaneously meets the acquisition constraint requirements and energy constraint conditions of the inspection task. During the constraint consistency verification process, for path segments with suboptimal energy consumption, path optimization correction is implemented without violating the acquisition constraint requirements of the inspection task, and the energy consumption of the corrected path segments is evaluated. Finally, based on the energy consumption evaluation results and the intermediate results in the path optimization correction process, candidate paths are comprehensively screened, prioritizing the path combinations that meet the acquisition requirements of the inspection task while having lower energy consumption as the final inspection path.

[0051] Furthermore, in the method provided in the application embodiments, the initial global path is validated for consistency and optimized and corrected using the local optimization information to obtain the final inspection path, which further includes: Based on the multi-level constraint attitude information and multi-level inspection paths in the local optimization information, path node matching is performed with the initial global path to obtain matched constraint attitudes, connectable nodes, and constraint acquisition paths. Using the matched constraint attitudes, connectable nodes, and constraint acquisition paths, combined with energy constraint information, constraint consistency verification is performed on the initial global path to determine whether each path segment in the initial global path simultaneously meets the inspection task acquisition constraint requirements and energy constraint conditions. For path segments with suboptimal energy consumption during constraint consistency verification, path optimization and correction are performed on these path segments while still meeting the inspection task acquisition requirements, and energy consumption is evaluated on the corrected path segments. Based on the intermediate results of the energy consumption evaluation and path optimization and correction, path combinations that meet the inspection task acquisition requirements while having lower energy consumption are preferentially retained as the final inspection path.

[0052] In this embodiment, when processing the initial global path using local optimization information, path node matching is performed first. Specifically, according to the order of path nodes in the initial global path, the spatial coordinates of each path node in the 3D channel environment model are read one by one, and these spatial coordinates are compared one by one with the spatial coordinates of path nodes included in the multi-level inspection path in the local optimization information. When the spatial distance between an initial global path node and a certain multi-level inspection path node is less than a preset spatial matching threshold, the two nodes are determined to be spatially corresponding nodes, and node matching is completed. After node matching is completed, the constraint acquisition path identifier and multi-level constraint attitude information corresponding to the matched node in the multi-level inspection path are mapped to the initial global path, thereby obtaining the matching constraint attitude corresponding to the local inspection requirements in the initial global path, and simultaneously identifying the nodes that are allowed to serve as connection boundaries in the multi-level inspection path, and determining the corresponding connectable nodes in the initial global path.

[0053] After completing path node matching, constraint consistency verification is performed on the initial global path. Specifically, according to the travel order of the initial global path, the path is divided into several continuous path segments, each consisting of several adjacent path nodes. For each path segment, it is first checked whether the path segment completely contains the required constraint acquisition path nodes, and whether the path segment meets the attitude range defined by the matching constraint attitude at each path node, that is, whether the running attitude corresponding to the path node falls within the allowable range specified by the multi-level constraint attitude information. Under the premise that the acquisition constraint conditions are met, the energy usage of the path segment is calculated according to the energy constraint information. During the calculation, according to the order of the path nodes in the path segment, the path segment is divided into several sub-segments between adjacent nodes. For each sub-segment, its corresponding running attitude parameters are read, and the unit distance energy consumption parameter of the sub-segment is determined according to the correspondence between running attitude and unit distance energy consumption in the energy constraint information. Then, the spatial distance between the two nodes of the sub-segment is calculated, and the spatial distance is multiplied by the corresponding unit distance energy consumption parameter to obtain the energy usage of the sub-segment during execution. By summing the energy usage of all sub-segments within a path segment, the total energy usage of that path segment during execution is obtained. This result is then compared with the energy constraints to determine whether the path segment meets the energy constraints. Through the above segment-by-segment calculation and judgment, the verification is completed to ensure that each path segment in the initial global path simultaneously meets the inspection task's data collection constraints and energy constraints.

[0054] During the constraint consistency verification process, for path segments identified that meet the inspection task's data acquisition constraints but have unsatisfactory energy usage, path optimization and correction are performed while ensuring that the corresponding constraint acquisition path and matching constraint posture are not disrupted. In this process, using the connectable nodes at both ends of the path segment as fixed start and end points, alternative paths connecting these points are re-searched in the 3D channel environment model. During the search, the energy usage of each candidate alternative path is calculated for its sub-segments using the same calculation method as described above, and the overall energy usage of the candidate paths is accumulated. Subsequently, among all candidate paths that meet the acquisition constraints, the path with the best overall energy usage is selected as the corrected path segment. After completing the path optimization and correction, the energy usage of the corrected path segment is calculated again using the same calculation steps to confirm the correction effect.

[0055] After optimizing and calculating the energy usage of all path segments with unsatisfactory energy consumption, a comprehensive evaluation is performed on the multiple candidate paths formed by combining these path segments. Specifically, it is first confirmed whether the candidate paths completely cover all constraint acquisition paths corresponding to the inspection tasks and whether the corresponding multi-level constraint attitude information is satisfied at all path nodes. Among the candidate paths that meet the above conditions, their overall energy consumption is accumulated and calculated, and then sorted in ascending order of overall energy consumption. Finally, the path with the optimal overall energy consumption while meeting the acquisition requirements of the inspection tasks is selected as the final inspection path.

[0056] In summary, the embodiments of this application have at least the following technical effects: This application acquires temporal environmental perception data collected by LiDAR and vision sensors mounted on an inspection robot, and calls upon structured geographic data of power transmission lines to construct a three-dimensional channel environment model including terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. Based on the three-dimensional channel environment model, energy constraints are introduced, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The inspection objectives and tasks of the preset inspection task points are analyzed, and local inspection optimization paths for the task points are established. The coordinates and constraint posture information of connectable nodes are identified as local optimization information. The initial global path is validated for consistency and optimized and corrected using the local optimization information to obtain the final inspection path. This invention solves the technical problem of high energy consumption caused by insufficient consideration of energy constraints in the path planning of inspection robots in the prior art. By introducing energy constraints on the basis of the three-dimensional channel environment model and optimizing and correcting the inspection path, the technical effect of reducing the overall energy consumption of the inspection robot while meeting the requirements of the inspection task is achieved.

[0057] Example 2, based on the same inventive concept as the path optimization method for the intelligent inspection robot of transmission lines in the foregoing examples, such as... Figure 2 As shown, this application provides a path optimization system for intelligent inspection robots of transmission lines. The system and method embodiments in this application are based on the same inventive concept. The system includes: The environment model construction module 11 is used to acquire time-series environmental perception data collected by the lidar and vision sensors mounted on the inspection robot, as well as the structured geographic data of the power transmission line, to construct a three-dimensional channel environment model including terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. The calculation module 12 is used to introduce energy constraints based on the three-dimensional channel environment model, and calculate the initial global path connecting multiple preset inspection task points with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The parsing module 13 is used to parse the inspection objectives and tasks of the preset inspection task points, establish local inspection optimization paths for the task points, and identify the coordinates and constraint posture information of connectable nodes as local optimization information. The correction module 14 is used to use the local optimization information to perform consistency verification and path optimization correction on the initial global path to obtain the final inspection path.

[0058] Furthermore, the system is also used to implement the following functions: The system acquires continuous temporal laser point cloud data and visual sensing data of the transmission line, and performs temporal alignment, denoising, and spatial registration processing on the laser point cloud data and visual sensing data. Based on the visual sensing data, image recognition is performed to generate semantic recognition information of obstacles and power transmission equipment in the transmission line. The semantic recognition information is used to semantically annotate the laser point cloud data to obtain three-dimensional point cloud data with semantic attributes. By calling the structured geographic data of the power transmission line through a geographic information system, the three-dimensional point cloud data with semantic attributes is spatially registered and fused with the structured geographic data to generate the three-dimensional transmission line environment model.

[0059] Furthermore, the system is also used to implement the following functions: Based on the operational data of the inspection robot, energy constraint information is constructed, and the energy timing of the inspection robot is dynamically constrained and configured. Using the three-dimensional channel environment model as a base, a path search space is constructed, and the energy constraint information is added to the path search space to predict and evaluate the energy consumption in path optimization. The preset inspection task points are projected onto the three-dimensional channel environment model, and the weighted sum of the total path length and the estimated total energy consumption is used as the optimization objective. Path search and energy consumption evaluation are performed through the path search space to obtain the inspection path that maximizes the optimization objective, which is used as the initial global path.

[0060] Furthermore, the system is also used to implement the following functions: Based on the historical operation records of the inspection robot, extract the energy consumption time series and operation posture data of the inspection robot; based on the energy consumption time series and operation posture data, establish the constraint relationship between time series energy consumption and operation posture; use the constraint relationship between time series energy consumption and operation posture to construct the energy constraint information, which is used to describe the relationship between energy consumption and operation posture during the operation of the inspection robot and the energy constraint conditions corresponding to the operation posture.

[0061] Furthermore, the system is also used to implement the following functions: The system analyzes the inspection equipment targets and task types at each preset inspection task point, and analyzes the inspection task requirements based on the current operating status information of the task point obtained by the edge computing module deployed in the corresponding area of ​​the task point. Based on the analyzed inspection task requirements, it determines the core inspection node of the task point and its corresponding constraint task acquisition parameters, and determines the auxiliary inspection node associated with the core inspection node and its corresponding optional acquisition parameters. Based on the core inspection node, constraint task acquisition parameters, auxiliary inspection nodes, and optional acquisition parameters, it projects the channel path coordinates of the task point and configures the parameters according to the constraint strength of the acquisition parameters. Based on the three-dimensional channel environment model, it identifies the constraint acquisition path of the task point, performs path node coherence constraint analysis based on the constraint acquisition path, determines connectable nodes and obtains the coordinates of the corresponding connectable nodes, and classifies the attitude parameters of the acquisition path according to the constraint strength of the acquisition parameters, setting multi-level constraint attitude information.

[0062] Furthermore, the system is also used to implement the following functions: The three-dimensional channel environment model digitally represents the inspection equipment target, inspection node location distribution, acquisition parameters and their constraint strength of the task node; based on the digital parameters of the task node, a multi-level inspection path search is performed according to the constraint strength of the acquisition parameters. In the search inspection path, the path that meets the continuity requirement is identified and screened according to the continuity constraint analysis of the acquisition parameters, and the constrained acquisition path is obtained.

[0063] Furthermore, the system is also used to implement the following functions: In a multi-level inspection path, a coherent constraint analysis is performed on the constraint acquisition path to identify indivisible path intervals. Based on these indivisible path intervals, connectable nodes are extracted from the multi-level inspection path. A connectable node is any node excluding the indivisible path interval, or the starting node of the indivisible path interval. The connectable node can serve as the starting or ending point of the path for connecting with other paths. Based on the determined connectable nodes, the corresponding spatial coordinates are extracted to obtain the coordinates of the connectable nodes.

[0064] Furthermore, the system is also used to implement the following functions: Based on the multi-level constraint attitude information and multi-level inspection paths in the local optimization information, path node matching is performed with the initial global path to obtain matched constraint attitudes, connectable nodes, and constraint acquisition paths. Using the matched constraint attitudes, connectable nodes, and constraint acquisition paths, combined with energy constraint information, constraint consistency verification is performed on the initial global path to determine whether each path segment in the initial global path simultaneously meets the inspection task acquisition constraint requirements and energy constraint conditions. For path segments with suboptimal energy consumption during constraint consistency verification, path optimization and correction are performed on these path segments while still meeting the inspection task acquisition requirements, and energy consumption is evaluated on the corrected path segments. Based on the intermediate results of the energy consumption evaluation and path optimization and correction, path combinations that meet the inspection task acquisition requirements while having lower energy consumption are preferentially retained as the final inspection path.

[0065] In Example 3, based on the same inventive concept as the path optimization method for intelligent inspection robots of transmission lines in the foregoing embodiments, this application also provides a computer-readable storage medium storing a computer program, which, when executed, implements the steps of any of the methods described in Example 1 above.

[0066] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0067] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A path optimization method for intelligent inspection robots of power transmission lines, characterized in that, include: Acquire temporal environmental perception data collected by the lidar and vision sensors mounted on the inspection robot, as well as the structured geographic data of the power transmission line, and construct a three-dimensional channel environment model that includes terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. Based on the three-dimensional channel environment model, energy constraints are introduced, and the initial global path connecting multiple preset inspection task points is calculated with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The inspection targets and tasks of the preset inspection task points are analyzed, a local inspection optimization path for the task points is established, and the coordinates and constraint attitude information of the connectable nodes are identified as local optimization information. The initial global path is validated for consistency and optimized using the local optimization information to obtain the final inspection path.

2. The path optimization method for intelligent inspection robot of transmission lines according to claim 1, characterized in that, A three-dimensional corridor environment model was constructed, including terrain elevation information, semantic labels of corridor obstacles, and structural features of power transmission equipment. Acquire continuous temporal laser point cloud data and visual sensing data of the channel, and perform temporal alignment, denoising and spatial registration processing on the laser point cloud data and visual sensing data; Image recognition is performed based on the visual sensing data to generate semantic recognition information of channel obstacles and power transmission equipment. The semantic recognition information is then used to semantically annotate the laser point cloud data to obtain three-dimensional point cloud data with semantic attributes. By calling the structured geographic data of the power transmission line through the geographic information system, the three-dimensional point cloud data with semantic attributes is spatially registered and fused with the structured geographic data to generate the three-dimensional channel environment model.

3. The path optimization method for intelligent inspection robot of transmission lines according to claim 1, characterized in that, Based on the aforementioned three-dimensional channel environment model, energy constraints are introduced. Using the weighted sum of the total path length and the estimated total energy consumption as the optimization objective, the initial global path connecting multiple preset inspection task points is calculated, including: Based on the operational data of the inspection robot, energy constraint information is constructed, and the energy timing of the inspection robot is dynamically constrained and configured. Using the three-dimensional channel environment model as a base, a path search space is constructed, and the energy constraint information is added to the path search space to predict and evaluate the energy consumption in path optimization. The preset inspection task points are projected into the three-dimensional channel environment model. The weighted sum of the total path length and the estimated total energy consumption is used as the optimization objective. The path search space is used to perform path search and energy consumption assessment to obtain the inspection path that maximizes the optimization objective, which is then used as the initial global path.

4. The path optimization method for intelligent inspection robot of transmission lines according to claim 3, characterized in that, Based on the operational data of the inspection robot, energy constraint information is constructed, including: Based on the historical operation records of the inspection robot, extract the energy consumption time sequence and operation posture data of the inspection robot; Based on energy consumption time series and operational attitude data, a constraint relationship between time series energy consumption and operational attitude is established; By utilizing the constraint relationship between the time-series energy consumption and the running posture, the energy constraint information is constructed to describe the relationship between the energy consumption and the running posture of the inspection robot, as well as the energy constraint conditions corresponding to the running posture.

5. The path optimization method for intelligent inspection robot of transmission lines according to claim 3, characterized in that, The inspection targets and tasks of the preset inspection task points are analyzed, local inspection optimization paths for the task points are established, and the coordinates and constraint attitude information of connectable nodes are identified, including: The inspection equipment targets and task types of each preset inspection task point are analyzed, and the inspection task requirements are analyzed based on the current operating status information of the task point obtained by the edge computing module deployed in the corresponding area of ​​the task point. Based on the inspection task requirements obtained from the analysis, the core inspection nodes of the task points and their corresponding constraint task acquisition parameters are determined, and the auxiliary inspection nodes associated with the core inspection nodes and their corresponding optional acquisition parameters are determined. Based on the core inspection nodes, the constraint task acquisition parameters, the auxiliary inspection nodes, and the optional acquisition parameters, the channel path coordinates of the task points are projected, and the parameters are configured according to the constraint strength of the acquisition parameters. Based on the three-dimensional channel environment model, the constraint acquisition path of the task point is identified. According to the constraint acquisition path, the path node coherence constraint analysis is performed to determine the connectable nodes and obtain the coordinates of the corresponding connectable nodes. According to the constraint strength of the acquisition parameters, the attitude parameters of the acquisition path are classified into levels, and multi-level constraint attitude information is set.

6. The path optimization method for intelligent inspection robot of transmission lines according to claim 5, characterized in that, The constraint acquisition path for identifying task points based on the aforementioned 3D channel environment model includes: The three-dimensional channel environment model digitally represents the inspection equipment target, the location distribution of the inspection node, the acquisition parameters and their constraint strength of the task node; Based on the digital parameters of the task nodes, a multi-level inspection path search is performed according to the constraint strength of the collected parameters. In the search inspection path, the path that meets the continuity requirement is identified and screened according to the continuity constraint analysis of the collected parameters, and the constrained collection path is obtained.

7. The path optimization method for intelligent inspection robot of transmission lines according to claim 6, characterized in that, Based on the constraint acquisition path, perform path node coherence constraint analysis to determine connectable nodes and obtain the coordinates of the corresponding connectable nodes, including: In multi-level inspection paths, coherent constraint analysis is performed on the constraint acquisition path to identify indivisible path intervals. Based on the indivisible path interval, connectable nodes are extracted from the multi-level inspection path. The connectable node is any node excluding the indivisible path interval, or the starting node of the indivisible path interval. The connectable node can be used as the starting point or ending point of the path to connect with other paths. Based on the identified connectable nodes, the corresponding spatial coordinates are extracted to obtain the coordinates of the connectable nodes.

8. The path optimization method for intelligent inspection robot of transmission lines according to claim 6, characterized in that, The initial global path is validated for consistency and optimized using the local optimization information to obtain the final inspection path, including: Based on the multi-level constraint attitude information and multi-level inspection path in the local optimization information, the path node is matched with the initial global path to obtain the matching constraint attitude, connectable nodes and constraint acquisition path. Using the matching constraint posture, connectable nodes, and constraint acquisition path, combined with energy constraint information, the initial global path is constrained to verify consistency, and it is determined whether each path segment in the initial global path simultaneously satisfies the inspection task acquisition constraint requirements and energy constraint conditions. For path segments with suboptimal energy consumption during constraint consistency verification, the path segments are optimized and corrected while meeting the data collection requirements of the inspection task, and the energy consumption of the corrected path segments is evaluated. Based on the intermediate results of the energy consumption assessment and path optimization correction, the path combinations that meet the data collection requirements of the inspection task while having low energy consumption are preferentially retained as the final inspection path.

9. A path optimization system for intelligent inspection robots of power transmission lines, characterized in that, The system is used to execute the path optimization method for the intelligent inspection robot of transmission lines as described in any one of claims 1-8, and the system includes: The environment model building module is used to acquire the time-series environmental perception data collected by the lidar and vision sensors mounted on the inspection robot, as well as the structured geographic data of the power transmission line, to build a three-dimensional channel environment model that includes terrain elevation information, semantic labels of channel obstacles, and structural features of power transmission equipment. The calculation module is used to introduce energy constraints based on the three-dimensional channel environment model, and to calculate the initial global path connecting multiple preset inspection task points with the weighted sum of the total path length and the estimated total energy consumption as the optimization objective. The parsing module is used to parse the inspection targets and tasks of the preset inspection task points, establish local inspection optimization paths for the task points, and identify the coordinates and constraint attitude information of connectable nodes as local optimization information. The correction module is used to perform consistency verification and path optimization correction on the initial global path using the local optimization information to obtain the final inspection path.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the path optimization method for intelligent inspection robots of transmission lines as described in any one of claims 1-8.

Images