Three-dimensional point cloud-based intelligent management and control method and system for uninterrupted operation of distribution network

Abstract

The invention discloses a three-dimensional point cloud-based intelligent management and control method and system for non-power-cut operation of a distribution network, and relates to the technical field of power operation management and control, and the method comprises the steps: integrating distribution network line data through a GIS, and eliminating a no-fly zone and an interference zone to generate an unmanned plane inspection route; a high-definition camera and a laser radar are used for collecting images and environment parameters, and a three-dimensional point cloud model is constructed. And calculating a shooting angle and classifying images based on the position relationship between the unmanned aerial vehicle and the tower, and fusing multi-source data to optimize the model precision. Effective images are screened, optimal angle parameters are determined, an optimal route is generated, the abnormal state of equipment is automatically recognized according to a route real-time collection data comparison model, fault alarm is triggered, maintenance resources are dispatched, and dynamic monitoring and closed-loop management and control are achieved. The method has the advantages that three-dimensional modeling, intelligent path planning and real-time analysis are integrated, the inspection efficiency and safety are remarkably improved, the manual intervention requirement is reduced, and the method is suitable for complex environment distribution network operation and maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network control, and specifically relates to an intelligent control method and system for live working on distribution networks based on three-dimensional point clouds. Background Art

[0002] With the increasing demand for electricity, live working on distribution networks has become increasingly crucial. Traditional manual inspection of distribution networks is time-consuming and laborious, and it is difficult to accurately detect hidden hazards. Driven by technology, the combination of drones and three-dimensional point cloud technology has emerged. This technology uses drones to collect data to construct a three-dimensional point cloud model, which can comprehensively and three-dimensionally present the details of distribution network equipment, realize real-time monitoring and intelligent analysis of equipment status. It not only greatly improves the inspection efficiency, but also provides strong support for the accurate control of live working on distribution networks, ensuring power supply reliability.

[0003] Existing intelligent control methods and systems for live working on distribution networks rely on experience for the path planning of drones, are easily affected by the environment, and it is difficult to comprehensively capture hidden hazards such as equipment deformation in three-dimensional space during data collection. Therefore, an intelligent control method and system for live working on distribution networks based on three-dimensional point clouds are needed to solve the above problems. Summary of the Invention

[0004] To solve the above technical problems, an intelligent control method and system for live working on distribution networks based on three-dimensional point clouds are provided. The present technical solution solves the problems of the existing intelligent control methods and systems for live working on distribution networks, which rely on experience for the path planning of drones, are easily affected by the environment, and it is difficult to comprehensively capture hidden hazards such as equipment deformation in three-dimensional space during data collection.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] An intelligent control method for live working on distribution networks based on three-dimensional point clouds, comprising:

[0007] S1. Obtain the distribution information of distribution network lines and the information of distribution network equipment, and determine the operation area information. Then, according to the distribution information of distribution network lines and the operation area information, determine the preset information of the drone inspection route;

[0008] S2. Based on the preset information of the drone inspection route, call the drone equipped with a high-definition camera and a lidar to collect the initial two-dimensional image data of the operation area, and record the initial collection environment information, the initial collection time information, and the initial collection position information at the same time;

[0009] Among them, the initial collection position information includes the position information of the initial collection drone, the distance information between the initial collection drone and the top of the pole tower, and the shortest radial distance information between the initial collection drone and the pole tower;

[0010] S3. Obtain the initial capture angle of the first-stage UAV based on the initial environment information, distribution information of the power distribution lines, position information of the first-stage UAV, distance information between the first-stage UAV and the tower top, and shortest radial distance information between the first-stage UAV and the tower pole, classify the two-dimensional image data, and divide the two-dimensional image data with the same first-stage UAV capture angle parameters into the same two-dimensional image data set;

[0011] S4. Construct a three-dimensional point cloud model based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information, and position information of the first-stage UAV;

[0012] S5. Determine the preferred UAV capture angle parameters based on the two-dimensional image data set, and based on the preferred UAV capture angle parameters, determine the corresponding position information of the first-stage UAV to obtain the preferred UAV position information. Then, through the preferred UAV position information and the three-dimensional point cloud model, determine the preferred information of the UAV inspection route;

[0013] S6. According to the preferred information of the UAV inspection route, call the UAV to collect real-time two-dimensional image data, determine the real-time point cloud data, monitor the status of the power distribution equipment to obtain the real-time status information of the power distribution equipment, and then based on the real-time two-dimensional image data and the real-time status information of the power distribution equipment, perform intelligent control on the power distribution equipment.

[0014] In an alternative embodiment, step S1 specifically includes:

[0015] Obtain the basic data of the power distribution lines from the geographic information system of the power company. The basic data of the power distribution lines includes power distribution voltage level information, line routing information, tower pole position information, substation position information, and line connection relationship information;

[0016] Perform data processing on the line routing information, tower pole position information, substation position information, and line connection relationship information, and unify the data formats in sequence to obtain the distribution information of the power distribution lines;

[0017] Based on the distribution information of the power distribution lines, determine the maximum influence reference range information, and according to the maximum influence reference range information, obtain the natural protection area information and population density area information within the maximum influence reference range in the geographic information system;

[0018] Based on the natural protection area information and population density area information within the maximum influence reference range, determine the UAV flight prohibited area information;

[0019] Remove the UAV flight prohibited area from the maximum influence reference range, and use the remaining area as the candidate operation area information;

[0020] Determine the UAV flight interference area information according to the distribution information of the power distribution lines and the power distribution voltage level information;

[0021] Based on the information of the to-be-selected operation area and the UAV flight interference area, remove the UAV flight interference area from the information of the to-be-selected operation area, and use the remaining area as the operation area information;

[0022] Based on the operation area information and the distribution information of the power distribution lines, traverse the position information of all the points closest to the power distribution lines within the operation area to obtain the reference points for the preset UAV inspection route;

[0023] According to the pole position information and the reference for the preset UAV inspection route, obtain the preset UAV inspection route, so as to determine the preset information of the UAV inspection route.

[0024] In an alternative embodiment, step S3 specifically includes:

[0025] According to the distribution information of the power distribution lines, the position information of the initial UAV, and the shortest radial distance information of the poles, determine the position information of the nearest pole. The position information of the nearest pole is the position information of the pole closest to the UAV corresponding to the position information of the initial UAV in the distribution information of the power distribution lines;

[0026] Based on the position information of the initial UAV and the position information of the nearest pole, construct a right triangle for determining the shooting angle with the two points of the position of the initial UAV and the position of the nearest pole as the hypotenuse of the right triangle;

[0027] Construct a space coordinate system for determining the shooting angle with the vertex where the right angle of the right triangle for determining the shooting angle is located as the origin;

[0028] According to the position information of the initial UAV and the position information of the nearest pole, determine the position coordinates (x0, y0, z0) of the initial UAV and the position coordinates (x1, y1, z1) of the nearest pole in the space coordinate system for determining the shooting angle;

[0029] According to the Pythagorean theorem, obtain the differences Δx = x1 - x0 and Δy = y1 - y0 between the projected coordinates of the pole on the horizontal plane and the projected coordinates of the UAV on the horizontal plane, so as to determine the projected distance between the UAV and the pole on the horizontal plane

[0030] Through Obtain the azimuth angle θ of the initial UAV shooting;

[0031] Obtain the vertical distance Δz = z1 - z0 between the UAV and the top of the pole;

[0032] Through Obtain the pitch angle of the initial UAV shooting;

[0033] Integrate the azimuth angle of the initial UAV shooting and the pitch angle of the initial UAV shooting to obtain the shooting angle of the initial UAV (θi , , where θ i is the initial acquisition azimuth angle of the initial acquisition UAV position corresponding to the i-th two-dimensional image data, and is the initial acquisition pitch angle of the initial acquisition UAV position corresponding to the i-th two-dimensional image data;

[0034] Initialize multiple empty two-dimensional image data sets, and each two-dimensional image data set corresponds to a shooting angle range;

[0035] Traverse each two-dimensional image data and its corresponding initial acquisition UAV shooting angle (θ i , , and check whether each two-dimensional image data set contains (θ i , . If so, divide the two-dimensional image data into the corresponding two-dimensional image data set. If not, initialize a new data set and divide the two-dimensional image data into it.

[0036] In an optional embodiment, step S4 specifically includes:

[0037] Based on the initial acquisition time information and the initial environment information, obtain the light intensity value, color temperature value, atmospheric transparency, and air humidity corresponding to the two-dimensional image data;

[0038] Obtain the light compensation factor according to the light intensity value and the color temperature value;

[0039] Obtain the atmospheric scattering compensation factor according to the atmospheric transparency and the air humidity;

[0040] Based on the light compensation factor and the atmospheric scattering compensation factor, compensate the two-dimensional image data to obtain the two-dimensional image standard data;

[0041] Use the SIFT algorithm to extract the feature points in the two-dimensional image standard data, and perform brute-force matching on the feature points between different two-dimensional image standard data to obtain the image matching feature point pairs;

[0042] Based on the initial acquisition UAV position information, obtain the control center position information, and construct a three-dimensional point cloud coordinate system with the control center position as the origin;

[0043] Based on the two-dimensional image standard data, the initial acquisition UAV shooting angle parameters, and the image matching feature point pairs, obtain the three-dimensional coordinates of the feature points of the image matching feature point pairs in the three-dimensional point cloud coordinate system;

[0044] Combine the three-dimensional coordinates of the feature points to form the initial three-dimensional point cloud data, and then perform optimization processing on the initial three-dimensional point cloud data to obtain the initial three-dimensional point cloud standard data;

[0045] Fuse the initial three-dimensional point cloud standard data to obtain a three-dimensional point cloud model.

[0046] In an alternative embodiment, step S5 specifically includes:

[0047] Based on the two-dimensional image dataset and the power distribution equipment information, determine the area values of the power distribution equipment and the two-dimensional image area values in all two-dimensional images in the two-dimensional image dataset;

[0048] According to the power distribution equipment information, determine the wire type information, and based on the wire type information, obtain the wire area values in all two-dimensional images;

[0049] Obtain the average value of the wire area values in all two-dimensional images, and use the average value of the wire area values as the two-dimensional image effective reference index threshold;

[0050] Use the ratio of the power distribution equipment area value to the two-dimensional image area value as the two-dimensional image effective reference index;

[0051] Obtain the two-dimensional image effective reference indexes corresponding to all two-dimensional image data in the two-dimensional image dataset, mark the two-dimensional image data with the two-dimensional image effective reference index greater than or equal to the two-dimensional image effective reference index threshold to obtain two-dimensional image marked data;

[0052] Use the proportion of the two-dimensional image marked data in the corresponding two-dimensional image dataset to obtain the important index of the UAV shooting angle;

[0053] Set the important index threshold of the UAV shooting angle, and use the two-dimensional image dataset with the important index of the UAV shooting angle greater than or equal to the important index threshold of the UAV shooting angle as the two-dimensional image preferred dataset;

[0054] Use the initial acquisition UAV shooting angle parameters corresponding to the two-dimensional image data in the two-dimensional image preferred dataset as the preferred UAV shooting angle parameters;

[0055] Based on the preferred UAV shooting angle parameters, determine the corresponding initial acquisition UAV position information, so as to obtain the preferred UAV position information;

[0056] Based on the preferred UAV position information and the three-dimensional point cloud model, determine the preferred UAV position coordinates in the three-dimensional point cloud model;

[0057] Combine the preferred UAV position coordinates in the three-dimensional point cloud model to obtain the preferred information of the UAV inspection route.

[0058] In an alternative embodiment, step S6 specifically includes:

[0059] Based on the preferred information of the UAV inspection route, combined with the corresponding initial acquisition time information, obtain the total UAV inspection time;

[0060] Based on the optimized information of the UAV inspection route, determine the corresponding distribution information of the distribution network lines, so as to determine the number of poles and the pole height information;

[0061] According to the number of poles and the pole height information, determine the pole interval information and the pole inspection time information;

[0062] Based on the pole interval information, the pole inspection time information and the total UAV inspection time, determine the UAV inspection interval time information, so as to determine the UAV call interval time information;

[0063] Call the UAV in turn according to the UAV call interval time information to collect two-dimensional image real-time data and point cloud real-time data;

[0064] Based on the two-dimensional image real-time data, determine the corresponding UAV real-time position information, compare the two-dimensional image real-time data and the point cloud real-time data corresponding to the same UAV real-time position information, and obtain the real-time status information of the distribution network equipment;

[0065] According to the real-time status information of the distribution network equipment, judge whether the distribution network equipment fails. If not, continue to monitor the status of the distribution network equipment. If so, obtain the distribution network equipment failure information and the failure location information;

[0066] According to the distribution network equipment failure information and the failure location information, determine the operation scenario information. Then, according to the operation scenario information, dispatch maintenance personnel to repair the distribution network equipment, and according to the failure location information and the UAV real-time position information, call the nearest UAV to monitor the maintenance process of the maintenance personnel, and complete the intelligent control of the distribution network equipment.

[0067] Furthermore, a distribution network live working intelligent control system based on three-dimensional point cloud is proposed, which is used to implement the control method as described in any one of the above, including:

[0068] A data acquisition and preprocessing module, which is used to obtain the distribution network line distribution information and the distribution network equipment information, and determine the operation area information. Then, according to the distribution network line distribution information and the operation area information, determine the preset information of the UAV inspection route. Based on the preset information of the UAV inspection route, call the UAV equipped with a high-definition camera and a lidar to collect the initial two-dimensional image data of the operation area, and record the initial acquisition environment information, the initial acquisition time information and the initial acquisition position information;

[0069] 3D modeling and angle classification module, which is used to obtain the initial capture angles of the UAV according to the initial environment information, distribution information of the power distribution lines, position information of the initial capture UAV, distance information between the initial capture UAV and the top of the pole tower, and shortest radial distance information between the initial capture UAV and the pole tower, classify the two-dimensional image data, divide the two-dimensional image data with the same initial capture angle parameters of the UAV into the same two-dimensional image dataset, and construct a 3D point cloud model based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information, and position information of the initial capture UAV;

[0070] Optimal path planning module, which is used to determine the preferred UAV capture angle parameters based on the two-dimensional image dataset, determine the corresponding position information of the initial capture UAV based on the preferred UAV capture angle parameters to obtain the preferred UAV position information, and then determine the preferred UAV inspection route information through the preferred UAV position information and the 3D point cloud model;

[0071] Real-time monitoring and fault response module, which is used to call the UAV to collect real-time two-dimensional image data according to the preferred UAV inspection route information, determine the real-time point cloud data, monitor the status of the power distribution equipment to obtain the real-time status information of the power distribution equipment, and then perform intelligent control on the power distribution equipment based on the real-time two-dimensional image data and the real-time status information of the power distribution equipment.

[0072] In an alternative embodiment, the data acquisition and preprocessing module includes:

[0073] Geographic information processing unit, which is used to obtain the distribution information of the power distribution lines and the information of the power distribution equipment, determine the operation area information, and then determine the preset UAV inspection route information according to the distribution information of the power distribution lines and the operation area information;

[0074] UAV acquisition unit, which is used to call the UAV equipped with a high-definition camera and a lidar to collect the initial two-dimensional image data of the operation area based on the preset UAV inspection route information, and record the initial acquisition environment information, initial acquisition time information, and initial acquisition position information at the same time.

[0075] In an alternative embodiment, the 3D modeling and angle classification module includes:

[0076] Calculation unit, which is used to obtain the initial capture angles of the UAV according to the initial environment information, distribution information of the power distribution lines, position information of the initial capture UAV, distance information between the initial capture UAV and the top of the pole tower, and shortest radial distance information between the initial capture UAV and the pole tower, classify the two-dimensional image data, and divide the two-dimensional image data with the same initial capture angle parameters of the UAV into the same two-dimensional image dataset;

[0077] A building unit, which is used to construct a three-dimensional point cloud model based on two-dimensional image data, initial acquisition time information, initial acquisition environment information, and initial UAV position information.

[0078] In an alternative embodiment, the optimal path planning module includes:

[0079] An evaluation unit, which is used to determine the preferred UAV shooting angle parameters based on the two-dimensional image data set;

[0080] A flight path generation unit, which is used to determine the corresponding initial UAV position information based on the preferred UAV shooting angle parameters to obtain the preferred UAV position information, and then determine the preferred UAV inspection flight path information through the preferred UAV position information and the three-dimensional point cloud model.

[0081] Compared with the prior art, the beneficial effects of the present invention are:

[0082] The intelligent management and control method and system for power distribution non-stop operation based on three-dimensional point cloud proposed in this solution uses the GIS system to integrate basic data such as power distribution lines and poles, dynamically excludes no-fly zones and electromagnetic interference zones, generates a safe and efficient UAV inspection flight path, and optimizes the shooting angle by calculating the pitch angle and azimuth angle between the UAV and the pole to reduce redundant data acquisition;

[0083] The intelligent management and control method and system for power distribution non-stop operation based on three-dimensional point cloud proposed in this solution constructs a three-dimensional point cloud model by fusing lidar ranging and high-definition image feature point matching (SIFT algorithm), accurately captures hidden dangers that are difficult to identify in two-dimensional images such as conductor sag and insulator tilt, and combines light compensation and environmental parameter correction to improve the model stability in complex weather;

[0084] The intelligent management and control method and system for power distribution non-stop operation based on three-dimensional point cloud proposed in this solution compares the three-dimensional point cloud model with the real-time collected data, automatically identifies abnormal states such as equipment deformation and displacement, triggers a fault alarm and schedules the nearest UAV to monitor the repair process, realizes the closed-loop management of "discovery - response - repair", optimizes the real-time data processing ability through edge computing, adapts to the requirements of multi-UAV collaborative operation, dynamically allocates tasks to avoid path conflicts, and significantly improves the inspection efficiency and safety in large-scale power distribution scenarios. Description of the Drawings

[0085] Figure 1 It is a flowchart of the intelligent management and control method for power distribution non-stop operation based on three-dimensional point cloud proposed by the present invention;

[0086] Figure 2 It is a flowchart for obtaining the three-dimensional point cloud model in the present invention;

[0087] Figure 3 It is a flowchart for obtaining the optimal information of the UAV inspection route in the present invention;

[0088] Figure 4 It is a system framework diagram of the intelligent management and control system for power distribution live working based on 3D point cloud proposed by the present invention. Detailed implementation manners

[0089] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0090] Referring to Figure 1 - Figure 3 As shown, the intelligent management and control method for power distribution live working based on 3D point cloud includes:

[0091] S1. Obtain the power distribution line distribution information and power distribution equipment information, and determine the operation area information. Then, according to the power distribution line distribution information and the operation area information, determine the preset information of the UAV inspection route;

[0092] Further, step S1 specifically includes:

[0093] Obtain the basic data of the power distribution line from the geographic information system of the power company. The basic data of the power distribution line includes power distribution voltage level information, line routing information, pole position information, substation location information, and line connection relationship information;

[0094] Among them, the power distribution voltage level information indicates the voltage level of the line (such as 10kV, 35kV), the line routing information is the geographical path of the line (such as a sequence of longitude and latitude coordinates), the pole position information is the coordinates and height of the pole, the substation location information is the geographical location of the substation, and the line connection relationship information is the line topology structure (such as the connection relationship between poles). The data source is the internal GIS database of the power company or the industry standard data format (such as Shp, GeoJSON), providing basic geographical and power grid topology information for subsequent route planning;

[0095] Perform data processing on the line routing information, pole position information, substation location information, and line connection relationship information, and unify the data format in sequence to obtain the power distribution line distribution information. Use tools such as ArcGIS Pro to convert data from different sources (such as CAD drawings, database tables, satellite images) into a unified format (such as GIS's Feature Class) to ensure data compatibility and facilitate subsequent spatial analysis and processing;

[0096] Based on the distribution information of the distribution network lines, determine the maximum impact reference range information, and according to the maximum impact reference range information, obtain the natural protection area information and population dense area information within the maximum impact reference range in the geographic information system;

[0097] Based on the natural protection area information and population dense area information within the maximum impact reference range, determine the no-fly area information for drones;

[0098] Among them, determining the maximum impact reference range information can be understood as setting a buffer radius according to the line voltage level and equipment importance (such as the buffer for high-voltage lines ≥ 500 meters). Specifically, a rectangular or circular range covering the distribution network lines and their surrounding areas can be generated using GIS spatial analysis tools (such as "buffer analysis"). The maximum impact reference range information is to define the geographical boundary that needs to be covered by the drone inspection to avoid missing key equipment. For the natural protection area information (obtaining data such as ecological protection areas and national parks from the environmental protection department) and population dense area information (defining high-risk areas based on population density layers, such as night light data), the no-fly area information for drones can be determined using GIS overlay analysis (such as "intersection" operation) to exclude the no-fly areas for drones and ensure flight safety and compliance;

[0099] Remove the no-fly area for drones from the maximum impact reference range, and use the remaining area as the candidate operation area information. Use GIS tools (such as the "clipping" function) to cut out the boundary of the candidate area to narrow the drone inspection range and focus on the areas where safe flight is possible around the power grid equipment;

[0100] According to the distribution information of the distribution network lines and the distribution network voltage level information, determine the drone flight interference area information. The drone flight interference area information is the areas where electromagnetic interference is likely to occur, such as near substations and high-voltage lines. Different distribution network voltage levels have different interference ranges for drones. For example, for high-voltage lines (≥ 35 kV), in the "Measures for the Air Traffic Management of Civil Unmanned Aerial Vehicle Systems" of the Civil Aviation Administration of China (CAAC), the recommended safe distance is 300 - 500 meters, and in this embodiment, 400 meters is adopted to avoid significant interference of the electromagnetic field on the drone navigation (GPS, magnetic compass) and communication (2.4 GHz frequency band). If it is medium- and low-voltage lines (10 - 35 kV), in the internal specifications of some power enterprises, the recommended safe distance is 50 - 200 meters. Since it is observed in practice that the electromagnetic field intensity within 50 meters may cause drone positioning drift or communication delay, 100 meters is taken in this embodiment;

[0101] Based on the candidate operation area information and the drone flight interference area information, remove the drone flight interference area from the candidate operation area information, and use the remaining area as the operation area information;

[0102] Based on the job area information and the distribution information of the distribution network lines, traverse the position information of all the points closest to the distribution network lines within the job area to obtain the reference points for the preset UAV inspection route. In this embodiment, it is necessary to use a GIS tool (such as the nearest neighbor analysis) within the job area to find the points closest to the distribution network lines (such as directly below the pole tower, at the line turning point), but it is necessary to ensure that the distance between the point closest to the distribution network lines and the distribution network lines is greater than or equal to 400 or 100, depending on the voltage level of the distribution network lines. At the same time, the height of the reference points for the preset UAV inspection route needs to be the same as the height of the pole tower;

[0103] According to the pole tower position information and the preset UAV inspection route reference, obtain the preset UAV inspection route, so as to determine the preset information of the UAV inspection route. For the preset UAV inspection route, when the reference point of the preset UAV inspection route is directly opposite the pole tower position or the substation position, the UAV needs to hover at this point, and make the equipped high-definition camera rotate up and down along the pole tower. The hovering time and the rotation time of the high-definition camera are determined according to the height of the pole tower. The hovering time = the height of the pole tower × the photographing frequency of the high-definition camera, and the rotation time = tan -1 (the height of the pole tower / the distance between the reference point of the preset UAV inspection route and the pole tower position). The angle of the high-definition camera is defaulted to 0 degrees (directly opposite the pole tower). Step S5 is to obtain the preferred UAV shooting angle parameters after the angle of the UAV is offset due to environmental influence (wind speed), and use the preferred UAV shooting angle parameters as the shooting angle of the UAV at the preferred UAV position.

[0104] S2. Based on the preset information of the UAV inspection route, call the UAV equipped with a high-definition camera and a lidar to collect the initial two-dimensional image data of the job area, and record the initial collection environment information, the initial collection time information, and the initial collection position information at the same time;

[0105] Among them, the initial collection position information includes the initial UAV position information, the distance between the initial UAV and the top of the pole tower, and the shortest radial distance information between the initial UAV and the pole tower;

[0106] Specifically, step S2 is as follows. Based on the preset route information (such as longitude and latitude coordinates, flight altitude, hovering points) generated in step S1, flight instructions are issued through a drone control platform (such as PX4, DJI Pilot). The high-definition camera can set the resolution (such as 4K), frame rate (such as 30fps), and exposure parameters (automatic / manual mode) before the drone takes off. A dynamic high-definition camera can be used to dynamically adjust the camera parameters according to the surrounding environment. In this embodiment, the drone is also equipped with sensors to record the initial acquisition environment information. The initial two-dimensional image data is the images of equipment such as poles, conductors, and insulators taken by the high-definition camera at regular intervals while the drone flies along the preset information of the drone inspection route. The initial two-dimensional image data is the images of equipment such as poles, conductors, and insulators taken by the high-definition camera at regular intervals while the drone flies along the preset information of the drone inspection route. The initial acquisition environment information includes light intensity, color temperature value, atmospheric transparency, and air humidity, which are used for subsequent image light compensation (S4) and atmospheric scattering correction. The initial acquisition position information includes the position information of the drone when it acquires the initial two-dimensional image data, the vertical distance between the drone and the top of the pole (calculated by lidar ranging), and the straight-line horizontal distance between the drone and the center point of the pole. The initial acquisition time information and the initial acquisition position information are used to associate the image / point cloud data with the physical space position, supporting subsequent angle calculation (S3) and three-dimensional modeling (S4).

[0107] S3. According to the initial environment information, the distribution information of the distribution network lines, the initial position information of the drone for acquisition, the distance information between the initial acquisition drone and the top of the pole, and the shortest radial distance information between the initial acquisition drone and the pole, obtain the shooting angle of the initial acquisition drone, and classify the two-dimensional image data. The two-dimensional image data with the same shooting angle parameters of the initial acquisition drone is divided into the same two-dimensional image data set;

[0108] Further, step S3 specifically includes:

[0109] According to the distribution information of the distribution network lines, the initial position information of the drone for acquisition, and the shortest radial distance information of the pole, determine the position information of the nearest pole. The position information of the nearest pole is the position information of the pole closest to the drone corresponding to the initial position information of the drone in the distribution information of the distribution network lines. Determining the position information of the nearest pole is to screen out the pole closest to the current position of the drone from the distribution information of the distribution network lines (i.e., the Euclidean distance is the smallest);

[0110] Based on the initial position information of the drone for acquisition and the position information of the nearest pole, construct a right triangle for determining the shooting angle with the two points of the initial position of the drone and the position of the nearest pole as the hypotenuse of the right triangle. For the two points on the hypotenuse of the right triangle, one of the points, the position of the nearest pole, is the bottom coordinate of the pole. The right triangle for determining the shooting angle is used to determine the actual angle when the drone uses the high-definition camera for shooting;

[0111] Taking the vertex where the right angle of the right triangle is located as the origin according to the shooting angle to construct a shooting angle determination space coordinate system, which is used to determine the actual angle of the UAV during shooting with a high-definition camera;

[0112] According to the initial acquisition UAV position information and the nearest tower position information, determine the initial acquisition UAV position coordinates (x0, y0, z0) and the nearest tower position coordinates (x1, y1, z1) in the shooting angle determination space coordinate system;

[0113] According to the Pythagorean theorem, obtain the differences Δx = x1 - x0 and Δy = y1 - y0 between the projection coordinates of the tower on the horizontal plane and the projection coordinates of the UAV on the horizontal plane, so as to determine the projection distance between the UAV and the tower on the horizontal plane

[0114] Through Obtain the initial acquisition UAV shooting azimuth angle θ;

[0115] Obtain the vertical distance Δz = Z1 - z0 between the UAV and the top of the tower;

[0116] Through Obtain the initial acquisition UAV shooting pitch angle;

[0117] Integrate the initial acquisition UAV shooting azimuth angle and the initial acquisition UAV shooting pitch angle to obtain the initial acquisition UAV shooting angle (θ i , , where θ i is the initial acquisition UAV shooting azimuth angle of the position of the initial acquisition UAV corresponding to the i-th two-dimensional image data, is the initial acquisition UAV shooting pitch angle of the position of the initial acquisition UAV corresponding to the i-th two-dimensional image data;

[0118] Initialize multiple empty two-dimensional image data sets, and each two-dimensional image data set corresponds to a shooting angle range;

[0119] Traverse each two-dimensional image data and its corresponding initial acquisition UAV shooting angle (θ i , , and check whether each two-dimensional image data set contains (θ i , . If so, divide the two-dimensional image data into the corresponding two-dimensional image data set. If not, initialize a new data set and divide the two-dimensional image data into it.

[0120] Specifically, after initializing multiple empty two-dimensional image data sets, and each two-dimensional image data set corresponds to a shooting angle range, it is necessary to set an angle tolerance based on the initial environment information. The angle tolerance includes the azimuth angle tolerance Δθ i and the pitch angle tolerance Then, traverse each piece of two-dimensional image data and its corresponding initial acquisition UAV shooting angle (θ i , , and check whether each two-dimensional image data set contains (θ i , . If so, divide the two-dimensional image data into the corresponding two-dimensional image data set. If not, initialize a new data set and divide the two-dimensional image data into it. It can be understood as judging |θ i -θ set |≤Δθ i and (where θ set and is the central value of the initial acquisition UAV shooting angle corresponding to this data set).

[0121] Among them, the calculation formulas for the azimuth tolerance and pitch tolerance are as follows:

[0122]

[0123] In the formula, Δθ i is the azimuth tolerance, α1, α2, α3, β1, β2, and β3 are all environmental parameter weight coefficients, and W, L, and H are the wind speed, light intensity, and atmospheric humidity respectively;

[0124] The specific parameter values are shown in Table 1 and Table 2 below:

[0125] Reference Table for Dynamic Angle Tolerance Weight Coefficient

[0126]

[0127]

[0128]

[0129] Table 1

[0130]

[0131]

[0132] Table 2

[0133] S4. Based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information, and initial acquisition UAV position information, construct a three-dimensional point cloud model;

[0134] Furthermore, step S4 specifically includes:

[0135] Based on the initial acquisition time information and initial environment information, obtain the light intensity value, color temperature value, atmospheric transparency, and air humidity corresponding to the two-dimensional image data;

[0136] Obtain a light compensation factor according to the light intensity value and the color temperature value;

[0137] Obtain an atmospheric scattering compensation factor according to the atmospheric transparency and the air humidity;

[0138] Based on the light compensation factor and the atmospheric scattering compensation factor, compensate the two-dimensional image data to obtain two-dimensional image standard data;

[0139] Use the SIFT algorithm to extract the feature points in the two-dimensional image standard data, and perform brute-force matching on the feature points between different two-dimensional image standard data to obtain image matching feature point pairs. SIFT extracts the local invariant feature points of the image through scale-space extreme value detection, and has rotation, translation, and scale invariance. Brute-force matching means using the Euclidean distance to calculate the similarity between feature points, and finding all possible matching pairs through brute-force traversal. Therefore, the image matching feature point pairs are the similarity points between two different two-dimensional images. When determining the similarity points, a threshold needs to be set. Since the area where the networking device is located belongs to a high-texture scene, the threshold can be set to 5-10 pixels;

[0140] Based on the initial collection UAV position information, obtain the control center position information, and construct a three-dimensional point cloud coordinate system with the control center position as the origin;

[0141] Based on the two-dimensional image standard data, the initial collection UAV shooting angle parameters, and the image matching feature point pairs, obtain the three-dimensional coordinates of the feature points of the image matching feature point pairs in the three-dimensional point cloud coordinate system. The initial UAV shooting angle parameters are used to determine the two-dimensional image standard data corresponding to the image matching feature point pairs;

[0142] Combine the three-dimensional coordinates of the feature points to form initial three-dimensional point cloud data, and then perform optimization processing on the initial three-dimensional point cloud data to obtain initial three-dimensional point cloud standard data. The corresponding feature points between two groups of images obtained by SIFT matching are converted into global three-dimensional coordinates through the high-definition camera projection model according to the initial collection UAV shooting angle parameters, and the three-dimensional coordinates of all feature points are stored as a list or array in the format of (X, Y, Z). Convert the point cloud in the local coordinate system (UAV perspective) to the global coordinate system (such as WGS84 or a custom reference system), and use methods such as voxel grid filtering to optimize the initial three-dimensional point cloud data. Then delete duplicate points through a hash table, and then calculate the normal vector to provide geometric information for subsequent fault detection (such as surface concavity and convexity analysis) to obtain the initial three-dimensional point cloud standard data;

[0143] Fuse the initial three-dimensional point cloud standard data to obtain a three-dimensional point cloud model.

[0144] Specifically, the light compensation factor includes brightness compensation and color temperature compensation. Based on the light compensation factor and the atmospheric scattering compensation factor, the process of compensating the two-dimensional image data to obtain the standard two-dimensional image data is as follows:

[0145] For the light compensation factor: brightness compensation factor = light intensity value / ideal light intensity. Then, for the brightness value of the pixel points in the two-dimensional image data, the compensated brightness value = the brightness value of the pixel points in the two-dimensional image data × brightness compensation factor. According to the ambient color temperature and the standard color temperature, some color space conversion and correction algorithms can be used to calculate the color compensation factor. Since the power distribution equipment is made of metal and has bright colors and high reflectivity in the captured images, the ideal light intensity is taken as 500 - 1000, and the value can be determined according to the location of the project implementation in this embodiment. For example, in the RGB color space, the deviation of the color temperature can be calculated to adjust the values of the three RGB channels to different degrees. For example, a simple approximation method is to calculate the compensation coefficients k R 、k G 、k B of the RGB channels according to the empirical relationship between the color temperature and the RGB gain, and adjust the RGB values of the pixels. For the RGB values (R, G, B) of the pixels in the two-dimensional image data, the adjusted RGB values (R1, G1, B1) of the standard two-dimensional image data are R1 = Rk R , G1 = Gk G , B1 = Bk B ; For the atmospheric scattering compensation factor: In the defogging algorithm based on the dark channel prior, first calculate the dark channel of the image where J c represents the c channel of the two-dimensional image J, and Ω(x) is a local window centered on x. Assuming the atmospheric transparency is A, the transmittance t(x) can be expressed as where ω is an empirical parameter, usually taken around 0.95. The transmittance here can be regarded as part of the atmospheric scattering compensation factor, which is used to defog the image. The restored two-dimensional standard image where I(x) is the two-dimensional image, t0 is a preset minimum transmittance threshold to prevent the denominator from being too small, and μ is the air humidity.

[0146] S5. Based on the two-dimensional image dataset, determine the preferred UAV shooting angle parameters, and based on the preferred UAV shooting angle parameters, determine the corresponding initial acquisition UAV position information to obtain the preferred UAV position information. Then, through the preferred UAV position information and the three-dimensional point cloud model, determine the preferred UAV inspection route information;

[0147] Further, step S5 specifically includes:

[0148] Based on the two-dimensional image dataset and the power distribution equipment information, determine the area values of the power distribution equipment and the area values of the two-dimensional images in all the two-dimensional images in the two-dimensional image dataset;

[0149] According to the power distribution equipment information, determine the wire type information, and based on the wire type information, obtain the wire area values in all the two-dimensional images. The wire types include overhead lines, cables, etc.;

[0150] Obtain the average value of the wire area values in all the two-dimensional images, and use the average value of the wire area values as the two-dimensional image effective reference index threshold. The two-dimensional image effective reference index threshold is used to determine the presence of power distribution equipment (such as cables) in the two-dimensional image and the integrity of the power distribution equipment;

[0151] Use the ratio of the power distribution equipment area value to the two-dimensional image area value as the two-dimensional image effective reference index;

[0152] Obtain the two-dimensional image effective reference indexes corresponding to all the two-dimensional image data in the two-dimensional image dataset, and mark the two-dimensional image data with the two-dimensional image effective reference index greater than or equal to the two-dimensional image effective reference index threshold to obtain the two-dimensional image marked data;

[0153] Use the proportion of the two-dimensional image marked data in the corresponding two-dimensional image dataset to obtain the important index of the UAV shooting angle;

[0154] Set the important index threshold of the UAV shooting angle, and use the two-dimensional image dataset with the important index of the UAV shooting angle greater than or equal to the important index threshold of the UAV shooting angle as the two-dimensional image preferred dataset. The important index threshold of the UAV shooting angle is recommended to be greater than or equal to 0.8 in actual calculation, which can determine the proportion of the preferred images;

[0155] Use the initial acquisition UAV shooting angle parameters corresponding to the two-dimensional image data in the two-dimensional image preferred dataset as the preferred UAV shooting angle parameters;

[0156] Based on the preferred UAV shooting angle parameters, determine the corresponding initial acquisition UAV position information, so as to obtain the preferred UAV position information;

[0157] Based on the preferred UAV position information and the three-dimensional point cloud model, determine the preferred UAV position coordinates in the three-dimensional point cloud model;

[0158] Combine the preferred UAV position coordinates in the three-dimensional point cloud model to obtain the preferred UAV inspection route information. Connect the preferred UAV position coordinates in the three-dimensional point cloud model in sequence to obtain the preferred UAV inspection route information. If there are points with the same perpendicular line (points parallel to the tower pole), only take the highest point.

[0159] Specifically, in step S5, by analyzing the two-dimensional image data, the images crucial for the inspection task are screened out, their shooting parameters and location information are extracted, and an optimized flight path is generated in the three-dimensional point cloud model. This method significantly reduces the amount of invalid data, improves the efficiency and coverage of UAV inspections, and is applicable to the intelligent operation and maintenance of complex distribution network environments.

[0160] S6. According to the UAV inspection route optimization information, call the UAV to collect real-time two-dimensional image data, determine the real-time point cloud data, monitor the status of distribution network equipment to obtain the real-time status information of distribution network equipment, and then based on the real-time two-dimensional image data and the real-time status information of distribution network equipment, conduct intelligent control of distribution network equipment.

[0161] Further, step S6 specifically includes:

[0162] Based on the UAV inspection route optimization information and combined with the corresponding initial acquisition time information, obtain the total UAV inspection time;

[0163] Based on the UAV inspection route optimization information, determine the corresponding distribution network line distribution information, and thus determine the number of pole towers information and the pole tower height information;

[0164] According to the number of pole towers information and the pole tower height information, determine the pole tower interval information and the pole tower inspection time information;

[0165] Based on the pole tower interval information, the pole tower inspection time information and the total UAV inspection time, determine the UAV inspection interval time information, and thus determine the UAV call interval time information;

[0166] Call the UAV in sequence according to the UAV call interval time information to collect real-time two-dimensional image data and real-time point cloud data;

[0167] Based on the real-time two-dimensional image data, determine the corresponding real-time UAV position information, compare the real-time two-dimensional image data and the real-time point cloud data corresponding to the same real-time UAV position information to obtain the real-time status information of distribution network equipment. The real-time UAV position information is used to record the position information of different UAVs during the inspection task. For example, if two different UAVs take pictures of the same pole tower at different time nodes and the two-dimensional images taken are of the same position of the pole tower, then this position is the same real-time UAV position.

[0168] According to the real-time status information of distribution network equipment, judge whether the distribution network equipment has a fault. If not, continue to monitor the status of the distribution network equipment. If so, obtain the distribution network equipment fault information and the fault location information;

[0169] For example, in the real-time data of the first 2D image, the power distribution equipment does not show deformation, displacement, wear, etc. However, in the real-time data of the second 2D image captured by a drone with the same real-time position information of the drone, the power distribution equipment shows deformation, displacement, wear, etc. In this case, it can be determined that the power distribution equipment has a fault (in this embodiment, a fault refers to deformation, displacement, wear, etc. of the power distribution equipment, and a large sag of the wire also belongs to a fault).

[0170] Based on the fault information and fault location information of the power distribution equipment, determine the operation scenario information. Then, according to the operation scenario information, dispatch maintenance personnel to repair the power distribution equipment. And according to the fault location information and the real-time position information of the drone, call the nearest drone to monitor the maintenance process of the maintenance personnel, so as to complete the intelligent management and control of the power distribution equipment.

[0171] It can be understood that according to the number of poles and towers information and the pole and tower height information, determine the pole and tower interval information and the pole and tower inspection time information. In step S1, it is recorded that "for the preset route of drone inspection, when the reference point of the preset route of drone inspection is directly opposite the position of the pole and tower or the substation, the drone needs to hover at this point, and make the equipped high-definition camera rotate up and down along the pole and tower. The hovering time and the rotation time of the high-definition camera are determined according to the height of the pole and tower. The hovering time = the height of the pole and tower × the camera shooting frequency, and the rotation time = tan -1 (the height of the pole and tower / the distance between the reference point of the preset route of drone inspection and the position of the pole and tower). The angle of the high-definition camera is defaulted to 0 degrees (directly opposite the pole and tower)". When the drone conducts inspections according to the optimized information of the drone inspection route in step S6, this process still needs to be carried out. Then the pole and tower inspection time information is the rotation time + the hovering time. The pole and tower interval information is the total number of lines between two poles and towers. For example, if there are 3 poles and towers, the pole and tower interval information is 2. Based on the pole and tower interval information, the pole and tower inspection time information and the total drone inspection time, determine the drone inspection interval time information, and then determine the drone call interval time information. The drone inspection interval time information = the total drone inspection time / the pole and tower interval × the pole and tower inspection time, and the drone call interval = max (the drone charging time, the inspection interval time - the flight time margin). In this way, it can be determined that during the power distribution inspection period (determined by the personnel in the control system center), the drones can complete the inspection tasks without interfering with each other and will not miss the power distribution equipment.

[0172] Specifically, step S6 realizes the full-process closed-loop management from inspection to repair through intelligent scheduling and real-time monitoring. Its core lies in accurate time calculation, efficient data analysis and flexible resource scheduling, and is applicable to high-complexity power distribution network operation and maintenance scenarios. During actual deployment, parameter tuning needs to be carried out in combination with specific hardware performance and network environment.

[0173] Further, referring toFigure 4 As shown in Figure 4 , an intelligent control system for power distribution non-stop operation based on 3D point cloud is proposed to implement the control method as described in any one of the above, including:

[0174] A data acquisition and preprocessing module, which is used to obtain the power distribution line distribution information and power distribution equipment information, determine the operation area information, and then determine the preset information of the UAV inspection route according to the power distribution line distribution information and the operation area information. Based on the preset information of the UAV inspection route, the UAV is called to carry a high-definition camera and a lidar to collect the initial two-dimensional image data of the operation area, and at the same time record the initial acquisition environment information, initial acquisition time information and initial acquisition location information;

[0175] A 3D modeling and angle classification module, which is used to obtain the initial UAV shooting angle according to the initial environment information, power distribution line distribution information, initial UAV position information, distance between the initial UAV and the pole top, and the shortest radial distance between the initial UAV and the pole tower, classify the two-dimensional image data, and divide the two-dimensional image data with the same initial UAV shooting angle parameters into the same two-dimensional image dataset. Based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information and initial UAV position information, a 3D point cloud model is constructed;

[0176] An optimal path planning module, which is used to determine the preferred UAV shooting angle parameters based on the two-dimensional image dataset, determine the corresponding initial UAV position information based on the preferred UAV shooting angle parameters to obtain the preferred UAV position information, and then determine the preferred information of the UAV inspection route through the preferred UAV position information and the 3D point cloud model;

[0177] A real-time monitoring and fault response module, which is used to call the UAV to collect real-time two-dimensional image data according to the preferred information of the UAV inspection route, determine the real-time point cloud data, monitor the state of the power distribution equipment to obtain the real-time state information of the power distribution equipment, and then perform intelligent control on the power distribution equipment based on the real-time two-dimensional image data and the real-time state information of the power distribution equipment.

[0178] Furthermore, the data acquisition and preprocessing module includes:

[0179] A geographic information processing unit, which is used to obtain the power distribution line distribution information and power distribution equipment information, determine the operation area information, and then determine the preset information of the UAV inspection route according to the power distribution line distribution information and the operation area information;

[0180] The UAV acquisition unit is used to call the high-definition camera and lidar carried by the UAV to collect the initial two-dimensional image data of the operation area based on the preset information of the UAV inspection route, and record the initial acquisition environment information, initial acquisition time information, and initial acquisition location information at the same time.

[0181] Further, the 3D modeling and angle classification module includes:

[0182] The calculation unit is used to obtain the initial shooting angle of the UAV according to the initial environment information, distribution information of the power distribution lines, initial position information of the UAV for acquisition, distance information between the initial UAV for acquisition and the top of the pole tower, and the shortest radial distance information between the initial UAV for acquisition and the pole tower, classify the two-dimensional image data, and divide the two-dimensional image data with the same initial shooting angle parameters of the UAV into the same two-dimensional image dataset;

[0183] The construction unit is used to construct a 3D point cloud model based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information, and initial position information of the UAV for acquisition.

[0184] Further, the optimal path planning module includes:

[0185] The evaluation unit is used to determine the preferred UAV shooting angle parameters based on the two-dimensional image dataset;

[0186] The route generation unit is used to determine the corresponding initial position information of the UAV for acquisition based on the preferred UAV shooting angle parameters to obtain the preferred UAV position information, and then determine the preferred information of the UAV inspection route through the preferred UAV position information and the 3D point cloud model.

[0187] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud, characterized in that: include: S1. Obtain distribution network line distribution information and distribution network equipment information, and determine the operation area information, and then determine the preset information of the drone inspection route based on the distribution network line distribution information and the operation area information; S2. Based on the preset information of the drone inspection route, the drone is called to carry a high-definition camera and a laser radar to collect the initial two-dimensional image data of the operation area, and the initial collection environment information, initial collection time information and initial collection location information are recorded at the same time; The initial collection position information includes the initial collection drone position information, the initial collection drone and the tower top distance information, and the initial collection drone and the tower shortest radial distance information; S3. According to the initial environment information, the distribution network line distribution information, the location information of the initial acquisition drone, the distance between the initial acquisition drone and the pole tower top, and the shortest radial distance between the initial acquisition drone and the pole tower, the shooting angle of the initial acquisition drone is obtained, and the two-dimensional image data is classified, and the two-dimensional image data with the same shooting angle parameters of the initial acquisition drone are divided into the same two-dimensional image data set; S4, constructing a three-dimensional point cloud model based on the two-dimensional image data, the initial acquisition time information, the initial acquisition environment information and the initial acquisition UAV position information; S5. Determine the preferred UAV shooting angle parameters based on the two-dimensional image data set, and determine the corresponding initial UAV position information based on the preferred UAV shooting angle parameters to obtain the preferred UAV position information, and then determine the preferred UAV inspection route information through the preferred UAV position information and the three-dimensional point cloud model; S6. Based on the optimization information of the drone inspection route, call the drone to collect real-time data of two-dimensional images and determine the real-time data of point clouds to monitor the status of the distribution network equipment and obtain the real-time status information of the distribution network equipment. Then, based on the real-time data of two-dimensional images and the real-time status information of the distribution network equipment, the distribution network equipment is intelligently managed and controlled.

2. The intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 1 is characterized in that: Step S1 specifically includes: Obtain basic data of distribution network lines from the power company's geographic information system. The basic data of distribution network lines includes distribution network voltage level information, line direction information, tower location information, substation location information and line connection relationship information; The line direction information, tower location information, substation location information and line connection relationship information are processed, and the data formats are unified in turn to obtain the distribution network line distribution information; Based on the distribution information of the distribution network lines, determine the maximum impact reference range information, and obtain the natural protection area information and densely populated area information within the maximum impact reference range in the geographic information system based on the maximum impact reference range information; Determine the prohibited areas for drone flights based on the information of nature reserves and densely populated areas within the maximum impact reference range; Remove the drone flight prohibited area from the maximum impact reference range, and use the remaining area as the candidate operation area information; Determine the UAV flight interference area information based on the distribution network line distribution information and distribution network voltage level information; Based on the information of the operation area to be selected and the information of the UAV flight interference area, the UAV flight interference area is removed from the information of the operation area to be selected, and the remaining area is used as the operation area information; Based on the operation area information and distribution network line distribution information, the location information of all points closest to the distribution network line in the operation area is traversed to obtain the reference points of the preset route for drone inspection; According to the tower location information and the preset route reference of the drone inspection, the preset route of the drone inspection is obtained, so as to determine the preset route information of the drone inspection.

3. The intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 1 is characterized in that: Step S3 specifically includes: According to the distribution network line distribution information, the location information of the initial acquisition drone and the shortest radial distance information of the tower, the location information of the nearest tower is determined. The location information of the nearest tower is the location information of the tower that is closest to the drone and corresponds to the location information of the initial acquisition drone in the distribution network line distribution information. Based on the location information of the initial acquisition drone and the location information of the nearest tower, the shooting angle is constructed to determine the right triangle with the location of the initial acquisition drone and the location of the nearest tower as the two points of the hypotenuse of the right triangle; A spatial coordinate system determined by the shooting angle is constructed by taking the vertex of the right angle of the right triangle determined by the shooting angle as the origin; According to the location information of the initial acquisition drone and the location information of the nearest tower, the location coordinates (x0, y0, z0) of the initial acquisition drone and the location coordinates (x1, y1, z1) of the nearest tower are determined in the spatial coordinate system determined by the shooting angle; According to the Pythagorean theorem, the difference between the projection coordinates of the tower on the horizontal plane and the projection coordinates of the drone on the horizontal plane is obtained, Δx=x1-x0 and Δy=y1-y0, so as to determine the projection distance between the drone and the tower on the horizontal plane. pass Get the initial drone shooting azimuth angle θ; Get the vertical distance between the drone and the top of the tower Δz=z1-z0; pass Get the pitch angle of the initial drone shot; Integrate the azimuth angle and pitch angle of the initial drone shooting to get the initial drone shooting angle Among them, θ i is the initial UAV shooting azimuth angle of the initial UAV position corresponding to the i-th two-dimensional image data, is the initial acquisition drone shooting pitch angle of the initial acquisition drone position corresponding to the i-th two-dimensional image data; Initialize multiple empty two-dimensional image data sets, and each two-dimensional image data set corresponds to a shooting angle range; Traverse each 2D image data and its corresponding initial drone shooting angle Checks whether each 2D image dataset contains If so, the two-dimensional image data is divided into the corresponding two-dimensional image data set; if not, a new data set is initialized and the two-dimensional image data is divided into it.

4. The intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 1 is characterized in that: Step S4 specifically includes: Based on the initial acquisition time information and the initial environment information, the light intensity value, the color temperature value, the atmospheric transparency and the air humidity corresponding to the two-dimensional image data are obtained; Obtain the lighting compensation factor according to the lighting intensity value and the color temperature value; Obtain atmospheric scattering compensation factor according to atmospheric transparency and air humidity; Based on the illumination compensation factor and the atmospheric scattering compensation factor, the two-dimensional image data is compensated to obtain the two-dimensional image standard data; The SIFT algorithm is used to extract feature points from the two-dimensional image standard data, and the feature points between different two-dimensional image standard data are violently matched to obtain image matching feature point pairs; Based on the initial drone location information, the control center location information is obtained, and a three-dimensional point cloud coordinate system is constructed with the control center location as the origin; Based on the two-dimensional image standard data, the initial UAV shooting angle parameters and the image matching feature point pairs, the three-dimensional coordinates of the feature points of the image matching feature point pairs in the three-dimensional point cloud coordinate system are obtained; The three-dimensional coordinates of the feature points are combined to form initial three-dimensional point cloud data, and then the initial three-dimensional point cloud data is optimized to obtain initial three-dimensional point cloud standard data; The initial 3D point cloud standard data is fused to obtain a 3D point cloud model.

5. The intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 1 is characterized in that: Step S5 specifically includes: Based on the two-dimensional image data set and the distribution network equipment information, determine the distribution network equipment area values ​​and the two-dimensional image area values ​​in all the two-dimensional images in the two-dimensional image data set; According to the distribution network equipment information, the wire type information is determined, and according to the wire type information, the wire area values ​​in all the two-dimensional images are obtained; Obtaining an average value of the wire area values ​​in all two-dimensional images, and taking the average value of the wire area values ​​as a threshold value of a valid reference index of the two-dimensional image; The ratio of the area value of the distribution network equipment to the area value of the two-dimensional image is used as the effective reference index of the two-dimensional image; Acquire the two-dimensional image effective reference index corresponding to all the two-dimensional image data in the two-dimensional image data set, mark the two-dimensional image data whose two-dimensional image effective reference index is greater than or equal to the two-dimensional image effective reference index threshold, and obtain the two-dimensional image marked data; The importance index of the drone shooting angle is obtained by taking the proportion of the two-dimensional image marker data in the corresponding two-dimensional image data set; A drone shooting angle importance index threshold is set, and a two-dimensional image dataset whose drone shooting angle importance index is greater than or equal to the drone shooting angle importance index threshold is used as a two-dimensional image preferred dataset; The initial UAV shooting angle parameters corresponding to the two-dimensional image data in the two-dimensional image optimization data set are used as the optimal UAV shooting angle parameters; Based on the preferred UAV shooting angle parameters, the corresponding initial UAV position information is determined, thereby obtaining the preferred UAV position information; Based on the preferred UAV position information and the three-dimensional point cloud model, determining the preferred UAV position coordinates in the three-dimensional point cloud model; The optimal UAV position coordinates in the three-dimensional point cloud model are combined to obtain the optimal UAV inspection route information.

6. The intelligent control method for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 1 is characterized in that: Step S6 specifically includes: Based on the drone inspection route optimization information and the corresponding initial collection time information, the total drone inspection time is obtained; Based on the optimal route information of drone inspection, the corresponding distribution network line distribution information is determined, thereby determining the number and height of towers; According to the information of pole tower quantity and pole tower height, determine the pole tower interval information and pole tower inspection time information; Based on the tower interval information, the tower inspection time information and the total inspection time of the drone, the drone inspection interval time information is determined, thereby determining the drone call interval time information; The drones are called in sequence to collect real-time data of two-dimensional images and point clouds according to the interval time information of the drone calls; Based on the real-time data of the two-dimensional image, the real-time location information of the corresponding drone is determined, and the real-time data of the two-dimensional image and the real-time data of the point cloud corresponding to the same real-time location information of the drone are compared to obtain the real-time status information of the distribution network equipment; According to the real-time status information of the distribution network equipment, determine whether the distribution network equipment has a fault. If not, continue to monitor the status of the distribution network equipment. If so, obtain the fault information and fault location information of the distribution network equipment; Based on the fault information and fault location information of the distribution network equipment, the operation scenario information is determined. Then, based on the operation scenario information, maintenance personnel are dispatched to repair the distribution network equipment. Based on the fault location information and the real-time location information of the drone, the nearest drone is called to monitor the maintenance process of the maintenance personnel, thereby completing the intelligent management and control of the distribution network equipment.

7. An intelligent management and control system for non-stop power supply operation of distribution network based on three-dimensional point cloud, used to implement the management and control method according to any one of claims 1 to 6, characterized in that: include: A data acquisition and preprocessing module, which is used to obtain the distribution information of the distribution network lines and the distribution network equipment information, and determine the operation area information, and then determine the preset information of the drone inspection route according to the distribution information of the distribution network lines and the operation area information, and based on the preset information of the drone inspection route, call the drone to carry a high-definition camera and a laser radar to collect the initial two-dimensional image data of the operation area, and record the initial collection environment information, initial collection time information and initial collection location information; A three-dimensional modeling and angle classification module, which is used to obtain the shooting angle of the initial acquisition drone according to the initial environment information, the distribution network line distribution information, the initial acquisition drone position information, the initial acquisition drone and the tower top distance information and the shortest radial distance information between the initial acquisition drone and the tower, and classify the two-dimensional image data, divide the two-dimensional image data with the same initial acquisition drone shooting angle parameters into the same two-dimensional image data set, and construct a three-dimensional point cloud model based on the two-dimensional image data, the initial acquisition time information, the initial acquisition environment information and the initial acquisition drone position information; An optimal path planning module, which is used to determine the preferred UAV shooting angle parameters based on the two-dimensional image data set, determine the corresponding initial UAV position information based on the preferred UAV shooting angle parameters, obtain the preferred UAV position information, and then determine the preferred UAV inspection route information through the preferred UAV position information and the three-dimensional point cloud model; The real-time monitoring and fault response module is used to call the drone to collect two-dimensional image real-time data and determine the point cloud real-time data according to the drone inspection route optimization information, monitor the status of the distribution network equipment, obtain the real-time status information of the distribution network equipment, and then perform intelligent management and control of the distribution network equipment based on the two-dimensional image real-time data and the real-time status information of the distribution network equipment.

8. The intelligent management and control system for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 7 is characterized in that: The data acquisition and preprocessing module includes: A geographic information processing unit, the geographic information processing unit is used to obtain distribution network line distribution information and distribution network equipment information, and determine the operation area information, and then determine the drone inspection route preset information according to the distribution network line distribution information and the operation area information; The drone collection unit is used to call the drone equipped with a high-definition camera and a laser radar to collect initial two-dimensional image data of the operation area based on the preset information of the drone inspection route, and at the same time record the initial collection environment information, initial collection time information and initial collection location information.

9. The intelligent management and control system for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 7 is characterized in that: The three-dimensional modeling and angle classification module includes: A calculation unit, the calculation unit is used to obtain the shooting angle of the initial acquisition drone according to the initial environment information, the distribution network line distribution information, the location information of the initial acquisition drone, the distance between the initial acquisition drone and the pole tower top, and the shortest radial distance between the initial acquisition drone and the pole tower, and classify the two-dimensional image data, and divide the two-dimensional image data with the same shooting angle parameters of the initial acquisition drone into the same two-dimensional image data set; A construction unit is used to construct a three-dimensional point cloud model based on the two-dimensional image data, initial acquisition time information, initial acquisition environment information and initial acquisition UAV position information.

10. The intelligent management and control system for non-stop power supply operation of distribution network based on three-dimensional point cloud according to claim 7, characterized in that: The optimal path planning module includes: An evaluation unit, the evaluation unit being used to determine a preferred drone shooting angle parameter based on the two-dimensional image data set; The route generation unit is used to determine the corresponding initial acquisition drone position information based on the preferred drone shooting angle parameters, obtain the preferred drone position information, and then determine the drone inspection route preferred information through the preferred drone position information and the three-dimensional point cloud model.

Images