Intelligent substation inspection method and device using laser radar and electronic equipment

Abstract

The invention discloses an intelligent substation inspection method and device using a laser radar and electronic equipment, and relates to the related field of power system monitoring and diagnosis, and the method comprises the steps: carrying out the initial scanning of a substation, and building a three-dimensional geometric model of each piece of equipment; configuring an inspection granularity template according to the three-dimensional geometric model; when the inspection equipment executes the inspection task, acquiring a currently identified first equipment object and a corresponding first inspection granularity; inputting the first inspection granularity into a scanning parameter adaptive regulator, outputting a first adaptive scanning parameter, and controlling a laser radar arranged on the inspection equipment to collect first inspection scanning data; and performing point cloud analysis on the first inspection scanning data, and identifying defect point cloud data to generate an inspection report. The technical problem that scanning parameters are difficult to flexibly adjust according to equipment characteristics in existing transformer substation intelligent inspection by using a laser radar, so that the inspection efficiency and reliability are insufficient is solved, and the technical effect of improving the inspection efficiency and reliability is achieved.

Description

Technical Field

[0001] The present application relates to fields related to power system monitoring and diagnosis, and in particular to methods, devices, and electronic equipment for intelligent inspection of substations using lidar. Background Art

[0002] In the power system, substations are key facilities, and their safe and stable operation is of vital importance. Any slight equipment defect may cause a serious accident, resulting in huge economic losses and even endangering personnel safety. Therefore, efficient and accurate substation inspections are the core link to ensure the reliability of power supply.

[0003] Currently, substation inspections primarily rely on traditional automated inspection equipment with fixed parameters. These devices scan substations according to pre-set scanning parameters. However, due to differences in structure, size, and operating status, fixed scanning parameters cannot be flexibly adjusted to suit the specific characteristics of each device. This results in insufficient inspection data accuracy for complex or critical equipment, making it difficult to accurately identify potential defects. Furthermore, for simpler or non-critical equipment, this can lead to data redundancy, increasing the data processing burden.

[0004] Among the current related technologies, intelligent substation inspection using lidar has the technical problem of difficulty in flexibly adjusting scanning parameters according to equipment characteristics, resulting in insufficient inspection efficiency and reliability. Summary of the Invention

[0005] The present application provides a method, device and electronic equipment for intelligent inspection of substations using lidar, uses lidar to perform an initial scan of the substation to establish a three-dimensional geometric model, configures an inspection granularity template according to the model, and adaptively adjusts the scanning parameters according to the equipment object during the inspection process. Technical means such as this can accurately adjust the scanning parameters of the lidar according to the characteristics of different equipment, ensure the acquisition of high-precision inspection data on complex equipment, accurately discover equipment defects, and reduce redundant data collection for simple equipment. This solves the technical problem of the existing intelligent inspection of substations using lidar that it is difficult to flexibly adjust the scanning parameters according to the characteristics of the equipment, resulting in insufficient inspection efficiency and reliability, and achieves the technical effect of improving inspection efficiency and reliability.

[0006] The present application provides a method for intelligent inspection of a substation using lidar, including: performing an initial scan on the substation to establish a three-dimensional geometric model of each device in the substation; configuring an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes inspection granularities corresponding to each device object; during the inspection device performing the inspection task of the substation, obtaining the first device object currently identified and the first inspection granularity corresponding to the first device object; inputting the first inspection granularity into a scan parameter adaptive regulator, and outputting the first adaptive scan parameter at the first inspection granularity according to the output of the scan parameter adaptive regulator to control the lidar set on the inspection device to collect the first inspection scan data of the first device object; performing point cloud analysis on the first inspection scan data to identify the defective point cloud data of the first device object and generate an inspection report.

[0007] In a possible implementation manner, when configuring the inspection granularity template according to the three-dimensional geometric model, the following processing is performed: according to the three-dimensional geometric model, extracting the key structure parameters of each device object, where the key structure parameters include geometric scale, surface curvature, the number of nested device components, and the number of external connections; according to the three-dimensional geometric model, extracting the key operation weights of each device object; calculating the inspection granularity of each device object according to the key structure parameters and the key operation weights, and outputting a set of inspection granularities; setting inspection granularity levels, and dividing the set of inspection granularities according to the inspection granularity levels to generate an inspection granularity template.

[0008] In a possible implementation manner, when inputting the first inspection granularity into the scan parameter adaptive regulator to train the scan parameter adaptive regulator, the following processing is performed: constructing inspection sample data of the substation, including combinations of lidar scan parameter samples of each device object at different inspection granularities, inspection scan parameter sample data collected by the lidar, and sample data characterizing the quality of the scan point cloud; outputting the point cloud quality labels of each device object under each granularity label; constructing a random forest, and training the random forest according to the inspection sample data corresponding to the point cloud quality labels until the optimal solutions of the inspection scan parameters corresponding to each device object are output; generating a scan parameter adaptive regulator based on the optimal solutions of the inspection scan parameters corresponding to each device object.

[0009] In a possible implementation manner, when outputting the first adaptive scan parameter at the first inspection granularity according to the scan parameter adaptive regulator, the following processing is performed: the scan parameter adaptive regulator outputs the optimal solution of the inspection scan parameters corresponding to the first device according to the first inspection granularity; obtaining the first adaptive scan parameter based on the optimal solution of the inspection scan parameters, where the first adaptive scan parameter includes point cloud sampling interval, horizontal angular resolution, vertical angular resolution, scan frequency, and scan range.

[0010] In a possible implementation, the following processing is performed: collecting historical defect samples of each device in the substation; calling the three-dimensional geometric models of each device to analyze the historical defect samples and outputting the geometric scales of the historical defects; recalculating the inspection granularity of each device object based on the geometric scales of the historical defects, the key structure parameters, and the key operation weights, and updating the inspection granularity template.

[0011] In a possible implementation, after calling the three-dimensional geometric models of each device to analyze the historical defect samples and outputting the geometric scales of the historical defects, the following processing is performed: vectorizing the historical defect samples by calling the three-dimensional geometric models of each device, including size scale, point cloud distribution density, boundary contour data, and defect depth; clustering the vectorized historical defect samples to obtain defect scale clusters with similar geometric features; calculating the scale mean of the defect scale clusters and outputting it as the geometric scale of the historical defects for each device object.

[0012] In a possible implementation, after collecting the historical defect samples of each device in the substation, the following processing is performed: constructing a historical defect sample library, where the historical defect sample library includes a set of historical defect samples corresponding to each device object; calling the historical defect sample library to compare the first inspection scan data with the set of historical defect samples corresponding to the first device object, and identifying defect point cloud data, including defect type and point cloud positioning result; generating an inspection report based on the defect type and point cloud positioning result.

[0013] In a possible implementation, the following processing is performed: during the inspection task of the substation by the inspection device, if the currently identified object switches to a second device object and the second inspection granularity corresponding to the second device object; outputting the second adaptive scan parameters at the second inspection granularity according to the scan parameter adaptive regulator to control the lidar set on the inspection device.

[0014] The present application also provides a substation intelligent inspection device using lidar, including: a three-dimensional geometric model establishment module for initially scanning the substation to establish three-dimensional geometric models of various devices in the substation; an inspection granularity template configuration module for configuring an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes inspection granularities corresponding to respective device objects; a first inspection granularity acquisition module for acquiring a currently identified first device object and a first inspection granularity corresponding to the first device object during the inspection device's execution of the substation inspection task; a first inspection scan data acquisition module for inputting the first inspection granularity into a scan parameter adaptive regulator and outputting first adaptive scan parameters at the first inspection granularity according to the scan parameter adaptive regulator to control the lidar provided on the inspection device to acquire first inspection scan data of the first device object; and a point cloud analysis module for performing point cloud analysis on the first inspection scan data to identify defective point cloud data of the first device object and generate an inspection report.

[0015] The present application also provides an electronic device, including: a memory for storing executable instructions; and a processor for implementing the substation intelligent inspection method using lidar when executing the executable instructions stored in the memory.

[0016] It is intended to first initially scan the substation through the substation intelligent inspection method, device, and electronic device using lidar proposed in the present application to establish three-dimensional geometric models of various devices in the substation, then configure an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes inspection granularities corresponding to respective device objects, then acquire a currently identified first device object and a first inspection granularity corresponding to the first device object during the inspection device's execution of the substation inspection task, then input the first inspection granularity into a scan parameter adaptive regulator and output first adaptive scan parameters at the first inspection granularity according to the scan parameter adaptive regulator to control the lidar provided on the inspection device to acquire first inspection scan data of the first device object, and finally perform point cloud analysis on the first inspection scan data to identify defective point cloud data of the first device object and generate an inspection report. The technical effect of improving the inspection efficiency and inspection reliability is achieved. Description of the Drawings

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the devices according to the embodiments of this application. It should be understood that the operations described above or below do not necessarily need to be precisely executed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0018] Figure 1 It is a schematic flowchart of the substation intelligent inspection method using lidar provided by the embodiments of this application.

[0019] Figure 2 It is a schematic structural diagram of the substation intelligent inspection device using lidar provided by the embodiments of this application.

[0020] Figure 3 It is a schematic structural diagram of an electronic device provided by the embodiments of this application.

[0021] Explanation of reference numerals: 3D geometric model establishment module 10, inspection granularity template configuration module 20, first inspection granularity acquisition module 30, first inspection scan data acquisition module 40, point cloud analysis module 50, input device 301, processor 302, memory 303, output device 304. Detailed implementation manners

[0022] The above description is only an overview of the technical solutions of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically gives the detailed implementation manners of this application.

[0023] In order to make the purpose, technical solutions and advantages of this application clearer, the following will further describe this application in detail with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of this application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

[0024] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or server that includes a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application.

[0025] An embodiment of this application provides a substation intelligent inspection method using lidar, as Figure 1 shown, the method includes: Step S100, perform an initial scan on the substation to establish a three-dimensional geometric model of each device in the substation.

[0026] Specifically, a high-precision lidar (LiDAR) device is used and installed on an inspection robot or a fixed bracket. Among them, lidar (LiDAR) is a sensor that uses lasers for distance measurement and target detection, and obtains the three-dimensional spatial information of the target by emitting laser beams and receiving reflected signals. The lidar obtains the three-dimensional spatial information of the substation equipment by emitting laser beams and receiving reflected signals. Use professional point cloud processing software (such as the PCL library) to preprocess the point cloud data collected by the lidar, including denoising, filtering, and stitching. Adopt a three-dimensional modeling algorithm based on point clouds, such as the Marching Cubes algorithm or the Poisson reconstruction algorithm, to convert the point cloud data into a three-dimensional geometric model of the device.

[0027] For example, the lidar is installed on the top of the inspection robot, and the robot moves along a preset path. The lidar scans the equipment in the substation in all directions at 360°. The lidar emits laser beams at a frequency of 10 times per second, and each emission covers a 360° horizontal view and a 30° vertical view to obtain the point cloud data of the equipment surface. Through the point cloud processing software, the collected point cloud data is denoised to remove abnormal points caused by environmental interference. Then, the Poisson reconstruction algorithm is used to convert the point cloud data into a three-dimensional geometric model of the device.

[0028] Step S200, configure an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes the inspection granularity corresponding to each device object.

[0029] Specifically, using deep learning algorithms such as YOLOv5 (the fifth version of the YOLO algorithm) or Faster R-CNN (Fast Region-based Convolutional Neural Network), classify and label the devices in the 3D geometric model to identify devices such as transformers, lightning arresters, switches, etc. According to the device type and importance, define the inspection granularity of each device through a configuration file or database. Among them, the inspection granularity refers to the accuracy and density when scanning the device. A high inspection granularity indicates high-density scanning and is applicable to critical devices; a low inspection granularity indicates low-density scanning and is applicable to non-critical areas. For example, the inspection granularity of the lightning arrester is set to high, and the inspection granularity of the open space is set to low. Combine the device classification results and the inspection granularity configuration to generate an inspection granularity template.

[0030] For example, use a pre-trained YOLOv5 model to classify the devices in the 3D geometric model, and the model outputs the category and location information of the devices. The inspection granularity configuration is as follows: Transformer: The inspection granularity is set to medium, and the scanning density is 100 points per square meter; Lightning arrester: The inspection granularity is set to high, and the scanning density is 200 points per square meter; Open space: The inspection granularity is set to low, and the scanning density is 50 points per square meter. Store the classification results and the inspection granularity configuration in the database to generate an inspection granularity template.

[0031] In a possible implementation manner, according to the 3D geometric model to configure the inspection granularity template, step S200 further includes step S210, according to the 3D geometric model, extract the key structure parameters of each device object, and the key structure parameters include geometric scale, surface curvature, the number of nested device components, and the number of external connections. Specifically, the key structure parameters refer to the geometric feature parameters of the device, including geometric scale, surface curvature, the number of component nestings, and the number of external connections, which are used to describe the physical structure complexity of the device. Using the point cloud data of the 3D geometric model, calculate the size parameters of the device, such as length, width, height, etc. For example, the Bounding Box algorithm can be used to quickly obtain the geometric scale of the device. By calculating the local curvature of the point cloud data, identify the shape features of the device surface. For example, the normal vector and curvature estimation algorithm, such as the PCA (Principal Component Analysis) algorithm, can be used to calculate the curvature of each point. Analyze the structural hierarchy of the device and count the number of nested components inside the device. For example, the depth-first search (DFS) or breadth-first search (BFS) algorithm can be used to traverse the geometric model of the device. Count the number of connection points between the device and other external devices. For example, it can be achieved by analyzing the geometric model and connection relationship diagram of the device.

[0032] Examples of extracting key structure parameters are as follows: For example, for a transformer device, using the bounding box algorithm, its length is calculated to be 2 meters, width is 1.5 meters, and height is 1.2 meters. For a lightning arrester, using the PCA algorithm to calculate the curvature of its surface points, and identifying the tip parts with larger curvature. For a complex switch device, using the DFS algorithm to count the number of internal nested components as 5. For a transformer, analyzing its geometric model and connection relation diagram, and counting the number of external connections as 3.

[0033] Step S220: Extract the key operation weights of each device object according to the three-dimensional geometric model. Specifically, the key operation weight is an important operation index of the device, used to measure the importance of the device in the operation of the substation. According to the operation importance, failure risk, and maintenance frequency of the device, assign an operation weight to each device. The operation weight can be obtained through expert experience or historical data statistics. For example, using the operation historical data, failure records, and maintenance plans of the device, calculate the operation weight of each device, and use the weighted average method or machine learning algorithms (such as decision trees) to extract the operation weight. For example, by analyzing the historical failure records and maintenance plans of a transformer, calculate its operation weight as 0.8 using the weighted average method.

[0034] Step S230: Calculate the inspection granularity of each device object according to the key structure parameters and the key operation weights, and output a set of inspection granularities. Specifically, define an inspection granularity calculation formula that comprehensively considers the key structure parameters and operation weights. For example: Inspection granularity = × Geometric scale + × Surface curvature + × Number of component nestings + × Number of external connections + × Operation weight, where , , , , are weight coefficients, which are adjusted according to actual needs. Substitute the extracted key structure parameters and operation weights into the formula to calculate the inspection granularity of each device, which is used to guide the adjustment of the scanning parameters of the lidar.

[0035] Step S240: Set the inspection granularity level, and divide the set of inspection granularities according to the inspection granularity level to generate an inspection granularity template. Specifically, according to the importance of the device and the inspection requirements, set different inspection granularity levels. For example, the inspection granularity can be divided into three levels: high, medium, and low. According to the set inspection granularity level, divide the calculated set of inspection granularities, and threshold division or clustering algorithms such as K-Means (K-means algorithm) can be used. Corresponding the divided inspection granularities to the device objects to generate an inspection granularity template.

[0036] For example, the inspection granularity level is set as follows: the inspection granularity level is set to high (2.0 and above), medium (1.0 - 2.0), and low (below 1.0). Using the threshold division method, the transformer with an inspection granularity of 2.5 is classified as a high granularity level, the lightning arrester with an inspection granularity of 1.8 is classified as a medium granularity level, and the open space with an inspection granularity of 1.0 is classified as a low granularity level. An inspection granularity template is generated to record the inspection granularity level of each device. This implementation method reasonably allocates inspection resources according to the key structural parameters and operation weights of the devices. More resources are allocated to important devices for high-precision inspections, while resource consumption is reduced in non-critical areas, avoiding resource waste caused by high-density scanning of the entire area, and achieving the technical effect of optimizing resource allocation.

[0037] In a possible implementation manner, step S200 further includes step S250 of collecting historical defect samples of each device in the substation. Specifically, historical defect samples of each device are extracted from the maintenance records of the substation, historical inspection reports, and the defect database of the devices. These samples include the location, type, size, and related image or point cloud data of the defects. The collected historical defect samples are standardized to unify the data format and ensure the integrity and consistency of the data.

[0038] For example, historical defect samples of the transformer are extracted from the maintenance system of the substation, including cracks on the surface of the transformer oil tank, corrosion of the radiator, etc. The length, width, and location information of the cracks are organized into structured data. For example, the crack length is 10 cm, the width is 0.5 cm, and the coordinates on the oil tank surface are (X = 1.2 m, Y = 0.8 m).

[0039] Step S260, calling the three-dimensional geometric models of each device to analyze the historical defect samples and outputting the geometric scale of the historical defects. Specifically, the historical defect samples are matched with the three-dimensional geometric models of the devices, and point cloud processing algorithms, such as the ICP algorithm (Iterative Closest Point algorithm), are used to align the point cloud data of the defect samples with the device models. By analyzing the aligned point cloud data, the geometric scale of the historical defects, including length, width, depth, etc., is calculated.

[0040] For example, the ICP algorithm is used to align the point cloud data of the crack on the surface of the transformer oil tank with the three-dimensional geometric model of the transformer. The calculated crack length is 10 cm, the width is 0.5 cm, and the depth is 0.1 cm.

[0041] Step S270, recalculating the inspection granularity of each device object based on the geometric scale of the historical defects, the key structural parameters, and the key operation weights, and updating the inspection granularity template. Specifically, an inspection granularity calculation formula that comprehensively considers the geometric scale of the historical defects, the key structural parameters, and the operation weights is defined. For example: Inspection granularity = × Geometric scale + × Surface curvature + × Number of component nestings + × Number of external connections + × Operating weight + × Geometric scale of historical defects, where , , , , , are weight coefficients, adjusted according to actual requirements. Substitute the geometric scale of historical defects and other parameters into the formula, recalculate the inspection granularity of each device, and update the inspection granularity template. By introducing the geometric scale of historical defects in this implementation method, the inspection system can more accurately identify potential problem areas of devices. For example, for a transformer that has had cracks, the system will increase the scanning density of that area to improve the accuracy of defect detection. At the same time, historical defect data provides an important reference basis for inspections, making inspections no longer a general strategy based on device type and structure, but customized inspections according to the actual situation of each device. For example, for an arrester that has had rust many times, the system will focus on its surface condition to promptly detect new rust points. Moreover, historical defect data enables more reasonable allocation of inspection resources. For devices with more historical defects or more serious defects, the system will allocate more resources for high-precision inspections; for devices with fewer historical defects, the resource input will be appropriately reduced, thereby improving the overall inspection efficiency.

[0042] In a possible implementation method, call the three-dimensional geometric models of each device to analyze the historical defect samples, and output the geometric scale of historical defects. Step S260 further includes step S261, which calls the three-dimensional geometric models of each device to perform vector representation on the historical defect samples, including dimensional scale, point cloud distribution density, boundary contour data, and defect depth. Specifically, match and align the point cloud data of the historical defect samples with the three-dimensional geometric models of the devices, and use the Iterative Closest Point (ICP) algorithm or other point cloud registration algorithms to ensure that the coordinate systems of the point clouds of the defect samples and the device models are consistent. Extract the geometric features of the defects from the aligned point cloud data, including dimensional scale, point cloud distribution density, boundary contour data, and defect depth. Convert the extracted geometric features into vector form. For example, the dimensional scale can be represented by vectors of length, width, and depth; the point cloud distribution density can be represented by the number of points per unit volume; and the boundary contour data can be represented by the coordinate sequence of the contour points.

[0043] For example, the ICP algorithm is used to align the point cloud data of the cracks on the surface of the transformer oil tank with the 3D geometric model of the transformer. The extracted features are as follows: Dimension scale: The crack length is 10 cm, the width is 0.5 cm, and the depth is 0.1 cm. Point cloud distribution density: The point cloud density in the crack area is 10 points per cubic centimeter. Boundary contour data: The boundary contour of the crack is represented by a series of coordinate points, such as: (1.2, 0.8), (1.3, 0.85), …. Defect depth: The maximum depth of the crack is 0.1 cm. The above features are represented in vector form, for example: Defect vector = [10, 0.5, 0.1, 10, (1.2, 0.8), (1.3, 0.85), …] Step S262: Cluster the vectorized historical defect samples to obtain defect scale clusters with similar geometric features. Specifically, since the dimensions of different features are different, it is necessary to normalize the features to ensure the accuracy of the clustering results. Input the normalized feature vectors into clustering algorithms such as K-Means, DBSCAN (DBSCAN density clustering), or hierarchical clustering, and divide the defect samples into different clusters according to similarity.

[0044] For example, normalize features such as dimension scale and point cloud distribution density to the interval [0, 1]. Input the normalized feature vectors into the K-Means algorithm to obtain 3 defect scale clusters: Cluster 1: The crack length is short (<5 cm), and the width is narrow (<0.3 cm). Cluster 2: The crack length is medium (5 - 10 cm), and the width is medium (0.3 - 0.8 cm). Cluster 3: The crack length is long (>10 cm), and the width is wide (>0.8 cm).

[0045] Step S263: Calculate the scale mean of the defect scale clusters and output it as the historical defect geometric scale of each device object. Specifically, for each defect scale cluster, calculate the mean of its geometric features, including length, width, depth, etc. Use the calculated scale mean as the historical defect geometric scale of each device object for subsequent inspection granularity calculation. This implementation method classifies defects with similar geometric features into one category through clustering analysis and calculates their scale mean, which can more accurately reflect the defect characteristics of the device. This makes the calculation of the inspection granularity more scientific and reasonable, and improves the pertinence and accuracy of the inspection.

[0046] Step S300: During the inspection task of the substation performed by the inspection device, obtain the currently identified first device object and the corresponding first inspection granularity of the first device object.

[0047] Specifically, the inspection robot uses lidar and an inertial measurement unit (IMU) for real-time positioning and map matching (SLAM, Simultaneous Localization and Mapping) to determine its own position and the positions of surrounding devices. It uses deep learning algorithms to identify devices from the point cloud data scanned in real time, matches the device objects in the inspection granularity template, and obtains the inspection granularity of the current device. The recognition results and inspection granularity information are transmitted to the control center through a wireless communication module (such as Wi-Fi or 4G) and stored in a database.

[0048] For example, the inspection robot uses lidar and IMU for SLAM to update its own position and the position information of surrounding devices in real time. During the inspection process, the lidar scans the devices in real time, and the deep learning algorithm identifies the current device as a lightning arrester. According to the inspection granularity template, the high inspection granularity corresponding to the lightning arrester is obtained (the scanning density is 200 points per square meter). The recognition results and inspection granularity information are transmitted to the control center through Wi-Fi and stored in the database.

[0049] Step S400: Input the first inspection granularity into the scan parameter adaptive regulator, and output the first adaptive scan parameters at the first inspection granularity according to the output of the scan parameter adaptive regulator, so as to control the lidar set on the inspection device to collect the first inspection scan data of the first device object.

[0050] Specifically, develop an adaptive regulator module that uses algorithms such as PID (Proportional-Integral-Derivative) control algorithm or fuzzy logic control algorithm to dynamically adjust the scan parameters of the lidar, such as scan density, scan frequency, and scan range, according to the input inspection granularity. In addition, this module can also be trained through machine learning methods to optimize the adjustment strategy of scan parameters, so as to more accurately adapt to different devices and inspection requirements. Send the adjusted scan parameters through the control interface of the lidar (such as API or SDK) to control the scanning behavior of the lidar.

[0051] For example, develop an adaptive regulator module based on the PID control algorithm, with the input being the inspection granularity and the output being the scan parameters. For example, for a lightning arrester, input the high inspection granularity, and the regulator outputs a scan density of 200 points per square meter and a scan frequency of 10 times per second. For an open area, input the low inspection granularity, and the regulator outputs a scan density of 50 points per square meter and a scan frequency of 5 times per second. Send the adjusted scan parameters to the lidar through the API interface of the lidar to control its scanning behavior.

[0052] In a possible implementation, the first inspection granularity is input into the scan parameter adaptive regulator to train the scan parameter adaptive regulator. Step S400 further includes step S410 of constructing the inspection sample data of the substation, including the lidar scan parameter sample combinations of each device object at different inspection granularities, the inspection scan parameter sample data collected by the lidar, and the sample data characterizing the quality of the scan point cloud. Specifically, key devices in the substation (such as transformers, lightning arresters, switches, etc.) are selected as the research objects. The scan parameters of the lidar at different inspection granularities are recorded, including scan density, scan frequency, scan range, etc. The point cloud data collected by the lidar under different parameters is recorded. The quality of the point cloud data is evaluated, including point cloud density, noise level, coverage range, etc.

[0053] Step S420, output the point cloud quality labels of each device object under each granularity label. Specifically, develop a point cloud quality evaluation algorithm, and assign quality labels (such as high, medium, low) to each sample data according to indicators such as point cloud density, noise level, and coverage range. The evaluation result is used as the point cloud quality label and recorded in the sample data. For example, use a comprehensive scoring formula for point cloud density, noise level, and coverage range: quality score = a × point cloud density - b × noise level + c × coverage range, where a, b, and c are weight coefficients.

[0054] Step S430, construct a random forest, and train the random forest according to the inspection sample data corresponding to the point cloud quality labels until the optimal solutions of the inspection scan parameters corresponding to each device object are output. Specifically, construct a random forest model. A random forest is an ensemble learning algorithm that improves the accuracy of classification or regression by constructing multiple decision trees and synthesizing their results. Use the inspection sample data (including scan parameters and point cloud quality labels) to train the random forest, and the goal is to predict the quality of the point cloud under different scan parameters. Optimize the performance of the random forest model through cross-validation and hyperparameter tuning.

[0055] For example, construct a random forest model containing 100 decision trees. Use the sample data to train the random forest model, with the input features including scan density, scan frequency, scan range, etc., and the output being the target quality label (high, medium, low). Adjust the hyperparameters of the random forest, such as the number of trees, maximum depth, etc., through cross-validation, and finally obtain an optimized model.

[0056] Step S440: Generate a scan parameter adaptive regulator based on the optimal solutions of the inspection scan parameters corresponding to each device object. Specifically, extract the optimal scan parameters of each device object at different inspection granularities from the trained random forest model. Package the extracted optimal scan parameters into the scan parameter adaptive regulator, and the regulator dynamically adjusts the scan parameters of the lidar according to the input inspection granularity. This implementation method can provide the optimal scan parameters for each device object at different inspection granularities through random forest model training, ensuring that the scan parameters of the lidar can dynamically adapt to the actual needs of the device and improving the scan efficiency and quality.

[0057] In a possible implementation, the scan parameter adaptive regulator outputs the first adaptive scan parameter at the first inspection granularity, and step S400 further includes step S450. The scan parameter adaptive regulator outputs the optimal solution of the inspection scan parameters corresponding to the first device according to the first inspection granularity. Specifically, extract the optimal scan parameters of each device object at different inspection granularities from the trained random forest model. Map the extracted optimal scan parameters to specific device objects and inspection granularities.

[0058] Step S460: Obtain the first adaptive scan parameter based on the optimal solution of the inspection scan parameters, where the first adaptive scan parameter includes the point cloud sampling interval, horizontal angular resolution, vertical angular resolution, scan frequency, and scan range. Specifically, convert the optimal scan parameters into the adaptive scan parameters of the lidar, including the point cloud sampling interval, horizontal angular resolution, vertical angular resolution, scan frequency, and scan range. Set the adaptive scan parameters through the control interface (such as API or SDK) of the lidar to control the scan behavior of the lidar.

[0059] For example, the scanning parameters of the transformer at high granularity are as follows: the point cloud sampling interval is 0.05 meters, the horizontal angular resolution is 0.1 degree, the vertical angular resolution is 0.1 degree, the scanning frequency is 10 times per second, and the scanning range is 360 degrees. The scanning parameters at medium granularity are: the point cloud sampling interval is 0.1 meters, the horizontal angular resolution is 0.2 degrees, the vertical angular resolution is 0.2 degrees, the scanning frequency is 5 times per second, and the scanning range is 360 degrees. The scanning parameters at low granularity are: the point cloud sampling interval is 0.2 meters, the horizontal angular resolution is 0.4 degrees, the vertical angular resolution is 0.4 degrees, the scanning frequency is 2 times per second, and the scanning range is 360 degrees. The scanning parameters of the lightning arrester at high granularity are: the point cloud sampling interval is 0.07 meters, the horizontal angular resolution is 0.15 degrees, the vertical angular resolution is 0.15 degrees, the scanning frequency is 8 times per second, and the scanning range is 360 degrees. The scanning parameters at medium granularity are: the point cloud sampling interval is 0.12 meters, the horizontal angular resolution is 0.3 degrees, the vertical angular resolution is 0.3 degrees, the scanning frequency is 4 times per second, and the scanning range is 360 degrees. The scanning parameters at low granularity are: the point cloud sampling interval is 0.25 meters, the horizontal angular resolution is 0.6 degrees, the vertical angular resolution is 0.6 degrees, the scanning frequency is 1 time per second, and the scanning range is 360 degrees. This implementation method can provide optimal scanning parameters for each device object at different inspection granularities through random forest model training, ensuring that the scanning parameters of the lidar can dynamically adapt to the actual needs of the device and improving the scanning efficiency and quality.

[0060] Step S500, perform point cloud analysis on the first inspection scanning data to identify the defective point cloud data of the first device object and generate an inspection report.

[0061] Specifically, use point cloud processing algorithms (such as voxel filtering, region growing segmentation, curvature analysis, etc.) to analyze the scanning data, and train deep learning models such as PointNet (point cloud network) or PointCNN (point cloud convolutional neural network) for detecting the defective point cloud data of the device. Develop a report generation module to organize the detected defect information into an inspection report, including the defect location, type, and severity. In addition, by calling the historical defect sample library, the first inspection scanning data can be compared with the historical defect samples of the corresponding device to further improve the accuracy of defect identification and the detail of the report.

[0062] For example, use the voxel filtering algorithm to downsample the scanning data, and then use the region growing segmentation algorithm to extract the region of interest on the surface of the device. Use the pre-trained PointNet model to analyze the extracted point cloud data to identify defects such as cracks and rust on the surface of the device. Organize the detected defect information into an inspection report, which includes the position coordinates of the defect, the type (such as crack, rust), and the severity (such as minor, medium, severe).

[0063] In a possible implementation, after collecting the historical defect samples of each device in the substation, step S500 further includes step S510 of constructing a historical defect sample library, where the historical defect sample library includes a set of historical defect samples corresponding to each device object. Specifically, the collected historical defect samples of each device are sorted out, including information such as the type, location, size, and point cloud data of the defects. The sorted defect samples are classified according to the device object to construct a historical defect sample library. Each device object corresponds to a set of historical defect samples, which is convenient for subsequent comparison and analysis.

[0064] For example, the set of historical defect samples of a transformer includes types such as cracks and rust, and each sample contains the location, size, and point cloud data of the defect. The set of historical defect samples of a lightning arrester includes types such as surface wear and local deformation, and each sample contains the location, size, and point cloud data of the defect.

[0065] Step S520: Call the historical defect sample library and compare the first inspection scan data with the set of historical defect samples corresponding to the first device object to identify the defect point cloud data, including the defect type and the point cloud positioning result. Specifically, use a point cloud comparison algorithm (such as the ICP algorithm or a feature-based matching algorithm) to compare the first inspection scan data with the point cloud data in the set of historical defect samples. Through the comparison result, identify the defect point cloud data in the first inspection scan data, including the defect type and the point cloud positioning result.

[0066] For example, use the ICP algorithm to compare the first inspection scan data with the historical crack samples of the transformer. Identify the crack type, determine that the position coordinates of the crack are (1.2 meters, 0.8 meters, 0.1 meters), the length is 10 centimeters, and the width is 0.5 centimeters.

[0067] Step S530: Generate an inspection report according to the defect type and the point cloud positioning result. Specifically, develop a report generation module to sort out the identified defect type and the point cloud positioning result into an inspection report. The report contains information such as the device name, defect type, defect location, defect size, and severity. Save the generated inspection report as a file (such as PDF or Excel format) and send it to the maintenance personnel. This implementation can more accurately identify the defect type and location by comparing the inspection scan data with the historical defect sample library. The historical defect sample library provides rich reference data, enabling the comparison algorithm to more precisely match and identify defects.

[0068] In a possible implementation, the method further includes: during the inspection device performing the inspection task of the substation, if the currently identified object switches to a second device object and the second inspection granularity corresponding to the second device object, output the second adaptive scanning parameters at the second inspection granularity according to the scanning parameter adaptive regulator to control the lidar provided on the inspection device.

[0069] Specifically, when the inspection device moves within the substation and identifies a new device object (such as switching from a transformer to a lightning arrester), the system dynamically adjusts the scanning parameters of the lidar through the scanning parameter adaptive regulator according to the inspection granularity of this device object (the second inspection granularity). The scanning parameter adaptive regulator outputs the optimal scanning parameters corresponding to the second device object according to the device type and inspection requirements. These parameters include the point cloud sampling interval, horizontal angular resolution, vertical angular resolution, scanning frequency, and scanning range, etc. In this way, the lidar can dynamically adjust the scanning parameters according to different device objects and inspection requirements, thereby improving the inspection efficiency and accuracy.

[0070] For example, when the inspection device switches from a transformer (the first device object) to a lightning arrester (the second device object), the scanning parameter adaptive regulator outputs scanning parameters with a scanning density of 150 points per square meter and a scanning frequency of 8 times per second according to the high inspection granularity of the lightning arrester. These parameters are sent to the lidar through the API interface of the lidar to control its scanning behavior to collect the point cloud data of the lightning arrester.

[0071] The embodiments of the present application use a lidar to perform an initial scan on the substation to establish a three-dimensional geometric model, configure an inspection granularity template according to the model, and adaptively adjust the scanning parameters according to the device object during the inspection process and other technical means. It can accurately adjust the scanning parameters of the lidar according to the characteristics of different devices, ensure high-precision inspection data is obtained on complex devices, accurately detect device defects, and at the same time reduce redundant data collection for simple devices. It solves the technical problem that the existing intelligent inspection of substations using lidar is difficult to flexibly adjust the scanning parameters according to the device characteristics, resulting in insufficient inspection efficiency and inspection reliability, and achieves the technical effect of improving the inspection efficiency and inspection reliability.

[0072] In the above text, reference is made to Figure 1 The intelligent substation inspection method using lidar according to the embodiments of the present invention is described in detail. Next, reference will be made to Figure 2 Describe the intelligent substation inspection device using lidar according to the embodiments of the present invention.

[0073] The substation intelligent inspection device using lidar according to an embodiment of the present invention is used to solve the technical problem that the existing substation intelligent inspection using lidar is difficult to flexibly adjust scanning parameters according to the characteristics of equipment, resulting in insufficient inspection efficiency and inspection reliability, and achieves the technical effect of improving inspection efficiency and inspection reliability. The substation intelligent inspection device using lidar includes: a three-dimensional geometric model establishment module 10, an inspection granularity template configuration module 20, a first inspection granularity acquisition module 30, a first inspection scan data acquisition module 40, and a point cloud analysis module 50.

[0074] The three-dimensional geometric model establishment module 10 is used to perform an initial scan on the substation and establish three-dimensional geometric models of various equipment in the substation; the inspection granularity template configuration module 20 is used to configure an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes inspection granularities corresponding to each equipment object; the first inspection granularity acquisition module 30 is used to obtain a currently identified first equipment object and a first inspection granularity corresponding to the first equipment object during the inspection of the substation by the inspection device; the first inspection scan data acquisition module 40 is used to input the first inspection granularity into a scan parameter adaptive regulator, output a first adaptive scan parameter at the first inspection granularity according to the scan parameter adaptive regulator, and control the lidar set on the inspection device to collect first inspection scan data of the first equipment object; the point cloud analysis module 50 is used to perform point cloud analysis on the first inspection scan data and identify defective point cloud data of the first equipment object to generate an inspection report.

[0075] Next, the specific configuration of the inspection granularity template configuration module 20 will be described in detail. As described above, an inspection granularity template is configured according to the three-dimensional geometric model. The inspection granularity template configuration module 20 may further include: a key structure parameter extraction unit for extracting key structure parameters of each equipment object according to the three-dimensional geometric model, where the key structure parameters include geometric scale, surface curvature, the number of nested equipment components, and the number of external connections; a key operation weight extraction unit for extracting key operation weights of each equipment object according to the three-dimensional geometric model; an inspection granularity calculation unit for calculating the inspection granularity of each equipment object according to the key structure parameters and the key operation weights and outputting an inspection granularity set; and an inspection granularity set division unit for setting inspection granularity levels and dividing the inspection granularity set according to the inspection granularity levels to generate an inspection granularity template.

[0076] Next, the specific configuration of the first inspection scan data acquisition module 40 will be described in detail. As described above, the first inspection granularity is input into the scan parameter adaptive regulator to train the scan parameter adaptive regulator. The first inspection scan data acquisition module 40 may further include: an inspection sample data construction unit for constructing the inspection sample data of the substation, including the lidar scan parameter sample combinations of each device object at different inspection granularities, the inspection scan parameter sample data collected by the lidar, and the sample data characterizing the quality of the scan point cloud; a point cloud quality label output unit for outputting the point cloud quality labels of each device object under each granularity label; a training unit for constructing a random forest and training the random forest according to the inspection sample data corresponding to the point cloud quality labels until the optimal solution of the inspection scan parameters corresponding to each device object is output; a scan parameter adaptive regulator generation unit for generating a scan parameter adaptive regulator based on the optimal solution of the inspection scan parameters corresponding to each device object.

[0077] Among them, according to the scan parameter adaptive regulator to output the first adaptive scan parameters at the first inspection granularity, the first inspection scan data acquisition module 40 may further include: an inspection scan parameter optimal solution output unit for the scan parameter adaptive regulator to output the optimal solution of the inspection scan parameters corresponding to the first device according to the first inspection granularity; a first adaptive scan parameter acquisition unit for obtaining the first adaptive scan parameters based on the optimal solution of the inspection scan parameters, where the first adaptive scan parameters include the point cloud sampling interval, the horizontal angular resolution, the vertical angular resolution, the scan frequency, and the scan range.

[0078] Among them, the inspection granularity template configuration module 20 may further include: a historical defect sample collection unit for collecting the historical defect samples of each device in the substation; a historical defect geometric scale output unit for analyzing the historical defect samples by calling the three-dimensional geometric models of each device and outputting the historical defect geometric scale; an inspection granularity template update unit for recalculating the inspection granularity of each device object with the historical defect geometric scale, the key structure parameters, and the key operation weights, and updating the inspection granularity template.

[0079] Among them, the three-dimensional geometric models of each device are called to analyze the historical defect samples, and the historical defect geometric scales are output. The historical defect geometric scale output unit may further include: a vectorization representation subunit for calling the three-dimensional geometric models of each device to perform vectorization representation on the historical defect samples, including size scale, point cloud distribution density, boundary contour data, and defect depth; a clustering subunit for clustering the vectorized historical defect samples to obtain defect scale clusters with similar geometric features; and a scale mean calculation subunit for calculating the scale mean of the defect scale clusters and outputting the historical defect geometric scales for each device object.

[0080] Next, the specific configuration of the point cloud analysis module 50 will be described in detail. As described above, after collecting the historical defect samples of each device in the substation, the point cloud analysis module 50 may further include: a historical defect sample library construction unit for constructing a historical defect sample library, where the historical defect sample library includes a set of historical defect samples corresponding to each device object; a defect point cloud data recognition unit for calling the historical defect sample library to compare the first inspection scan data with the set of historical defect samples corresponding to the first device object and identify the defect point cloud data, including the defect type and the point cloud positioning result; and an inspection report generation unit for generating an inspection report according to the defect type and the point cloud positioning result.

[0081] Among them, the device may further include: an inspection device object switching module for, during the inspection task of the substation by the inspection device, if the currently identified object is switched to a second device object and a second inspection granularity corresponding to the second device object, outputting the second adaptive scan parameters at the second inspection granularity according to the scan parameter adaptive regulator to control the lidar provided on the inspection device.

[0082] The substation intelligent inspection device using lidar provided by the embodiments of the present invention can execute the substation intelligent inspection method using lidar provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0083] Although various references are made to certain modules in the device according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or the server. The included units and modules are only divided according to the functional logic, but are not limited to the above division as long as the corresponding functions can be realized; in addition, the specific names of the functional units are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0084] Based on the foregoing embodiments, an electronic device is further provided in the embodiments of the present application. Figure 3It is a schematic structural diagram of an electronic device provided by an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing the embodiments of the present invention. Figure 3 The displayed electronic device is only an example and should not impose any limitations on the functions and scope of use of the embodiments of the present invention. The electronic device is presented in the form of a general-purpose computing device, and its components may include, but are not limited to, an input device 301, a processor 302, a memory 303, and an output device 304. Among them, the processor 302 may be one or more; the memory 303 may include a computer-readable medium and at least one program product, and this program product has a set (at least one) of program modules, and these program modules are configured to execute the functions of the embodiments of the present application.

[0085] The memory 303 shown in the embodiments of the present invention may adopt any combination of one or more computer-readable media; the computer-readable storage medium may be, but is not limited to, infrared rays, semiconductor systems, devices or components, or any combination of the above, for storing software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the substation intelligent inspection method using lidar in the embodiments of the present invention. The processor 302 executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory 303, that is, implements the above-mentioned substation intelligent inspection method using lidar.

[0086] The above specific embodiments do not constitute a limitation to the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application. In some cases, the actions or steps recorded in the present application may be executed in a different order from that in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or consecutive order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for intelligent inspection of a substation using lidar, characterized in that The method includes: Performing an initial scan on the substation to establish a three-dimensional geometric model of each device in the substation; Configuring an inspection granularity template according to the three-dimensional geometric model, where the inspection granularity template includes inspection granularities corresponding to each device object; During the inspection task of the substation by the inspection device, obtaining the currently identified first device object and the first inspection granularity corresponding to the first device object; Inputting the first inspection granularity into the scan parameter adaptive regulator, and outputting the first adaptive scan parameter at the first inspection granularity according to the scan parameter adaptive regulator to control the lidar set on the inspection device to collect the first inspection scan data of the first device object; Performing point cloud analysis on the first inspection scan data to identify the defective point cloud data of the first device object and generate an inspection report.

2. The intelligent substation inspection method using lidar according to claim 1, characterized in that Configuring an inspection granularity template according to the three-dimensional geometric model, the method includes: Extracting the key structure parameters of each device object according to the three-dimensional geometric model, where the key structure parameters include geometric scale, surface curvature, the number of nested device components, and the number of external connections; Extracting the key operation weights of each device object according to the three-dimensional geometric model; Calculating the inspection granularity of each device object according to the key structure parameters and the key operation weights, and outputting a set of inspection granularities; Setting an inspection granularity level, and dividing the set of inspection granularities according to the inspection granularity level to generate an inspection granularity template.

3. The intelligent substation inspection method using lidar according to claim 1, characterized in that, Inputting the first inspection granularity into the scan parameter adaptive regulator to train the scan parameter adaptive regulator, the method includes: Constructing the inspection sample data of the substation, including the lidar scan parameter sample combinations of each device object at different inspection granularities, the inspection scan parameter sample data collected by the lidar, and the sample data characterizing the quality of the scanned point cloud; Outputting the point cloud quality labels of each device object under each granularity label; Constructing a random forest, and training the random forest according to the inspection sample data corresponding to the point cloud quality labels until the optimal solution of the inspection scan parameters corresponding to each device object is output; Generating a scan parameter adaptive regulator based on the optimal solution of the inspection scan parameters corresponding to each device object.

4. The substation intelligent inspection method using lidar according to claim 3, characterized in that Outputting the first adaptive scan parameter at the first inspection granularity according to the scan parameter adaptive regulator, the method includes: The scan parameter adaptive regulator outputs the optimal solution of the inspection scan parameters corresponding to the first device according to the first inspection granularity; Obtaining the first adaptive scan parameter based on the optimal solution of the inspection scan parameters, where the first adaptive scan parameter includes point cloud sampling interval, horizontal angular resolution, vertical angular resolution, scan frequency, and scan range.

5. The intelligent substation inspection method using lidar according to claim 2, wherein The method includes: Collecting the historical defect samples of each device in the substation; Invoking the three-dimensional geometric model of each device to analyze the historical defect samples and outputting the historical defect geometric scale; Recalculating the inspection granularity of each device object with the historical defect geometric scale, the key structure parameters, and the key operation weights, and updating the inspection granularity template.

6. The intelligent substation inspection method using lidar according to claim 5, characterized in that, The three-dimensional geometric model of each device is called to analyze the historical defect sample and output the geometric scale of the historical defect. The method includes: Calling the three-dimensional geometric model of each device to vectorize the historical defect samples, including size scale, point cloud distribution density, boundary contour data and defect depth; Cluster the vectorized historical defect samples to obtain defect scale clusters with similar geometric features; The scale mean of the defect scale cluster is calculated and output as the historical defect geometric scale of each equipment object.

7. The substation intelligent inspection method using lidar according to claim 5, wherein, After collecting historical defect samples of each device in the substation, the method includes: Building a historical defect sample library, wherein the historical defect sample library includes a set of historical defect samples corresponding to each device object; Calling the historical defect sample library, comparing the first inspection scan data with a historical defect sample set corresponding to the first device object, and identifying defect point cloud data, including defect type and point cloud positioning results; Generate an inspection report based on the defect type and point cloud positioning results.

8. The intelligent inspection method for substations using laser radar according to claim 1, characterized in that: The method comprises: When the inspection device performs an inspection task of the substation, if the currently identified object is switched to a second device object, and a second inspection granularity corresponding to the second device object; The scanning parameter adaptive regulator outputs the second adaptive scanning parameter at the second inspection granularity to control the laser radar set on the inspection device.

9. The intelligent substation inspection device using lidar is characterized in that The device is used to implement the substation intelligent inspection method using laser radar according to any one of claims 1 to 8, and the device includes: A three-dimensional geometric model building module is used to perform an initial scan of the substation and build a three-dimensional geometric model of each device in the substation; An inspection granularity template configuration module, configured to configure an inspection granularity template according to the three-dimensional geometric model, wherein the inspection granularity template includes an inspection granularity corresponding to each device object; a first inspection granularity acquisition module, configured to acquire, when an inspection device performs an inspection task of the substation, a currently identified first device object and a first inspection granularity corresponding to the first device object; a first patrol scanning data acquisition module, configured to input the first patrol granularity into a scanning parameter adaptive regulator, and output a first adaptive scanning parameter at the first patrol granularity according to the scanning parameter adaptive regulator, so as to control a laser radar provided on the patrol device to collect first patrol scanning data of the first device object; The point cloud analysis module is used to perform point cloud analysis on the first inspection scanning data, identify defect point cloud data of the first equipment object and generate an inspection report.

10. An electronic device, characterized in that, The electronic device comprises: a memory for storing executable instructions; The processor is configured to implement the intelligent substation inspection method using lidar as described in any one of claims 1 to 8 when executing the executable instructions stored in the memory.

Images