A method and system for monitoring the shape of a power tower based on a laser radar

Abstract

The present application relates to the technical field of ultra-high voltage transmission line monitoring, and particularly relates to a power tower form monitoring method and system based on a laser radar. The power tower form monitoring method based on a laser radar proposed by the present application firstly compares an initial point cloud of a power tower with a real-time point cloud, determines a fitting ridge line of the power tower by using fitting of a feature plane, then finds real-time point clouds corresponding to lowest points of each base of the power tower according to the fitting ridge line, and finally fits a measured base plane by combining the lowest points of each base in the real-time point cloud to calculate the inclination, translation and settlement of the base. The power tower form monitoring method based on a laser radar proposed by the present application can measure the form of a power tower, identify the specific degree of inclination, translation and settlement, and solve the monitoring needs of a power tower of a transmission line in various actual scenarios. The method has high automation, a simple calculation process and accurate and reliable calculation results, and effectively improves the work efficiency of power inspection.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for ultra-high voltage and extra-high voltage transmission lines, and in particular to a method and system for monitoring the morphology of power towers based on lidar. Background Technology

[0002] Ultra-high voltage (UHV) transmission lines are efficient and fast energy transmission channels and optimized allocation platforms, playing a vital pivotal role in the modern energy supply system.

[0003] High-voltage power towers play a crucial role in power transmission lines. However, due to factors such as natural disasters (e.g., rain, snow, strong winds), geological mining, engineering construction, and human sabotage, tower tilting, shifting, and settling frequently occur. Severe tower displacement can affect the normal operation of ultra-high-voltage and extra-high-voltage transmission lines, and may even lead to tower collapse. Therefore, morphological monitoring of power towers to enable early detection and timely alarm of tower displacement is essential. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies that lack effective monitoring methods for real-time monitoring of high-voltage power towers, this invention proposes a power tower morphology monitoring method based on lidar. By using lidar to monitor point clouds in real time to determine the offset of the power tower, the tilt angle, settlement distance, and translation distance and angle of the power tower can be accurately calculated, thus realizing accurate monitoring and early warning of power towers.

[0005] This invention proposes a method for monitoring the morphology of power towers based on lidar, characterized by comprising the following steps:

[0006] S1. Under normal conditions, the power tower is collected by lidar, and an initial point cloud is extracted from the power tower's point cloud. This initial point cloud covers all the bases of the power tower. Based on this initial point cloud, the center point O of the power tower base plane is obtained and the non-base area is extracted.

[0007] S2. Under monitoring conditions, real-time point cloud data of the power tower is collected using lidar.

[0008] S3. Divide the real-time point cloud into multiple segments in the vertical direction, select at least two segments as fitting segments, and perform horizontal plane fitting on the laser points in the real-time point cloud located in the fitting segments. The horizontal plane obtained by fitting is used as the feature plane corresponding to the segment.

[0009] S4. Extract the corner points of each feature plane located in the same direction and perform line fitting. The fitted line is used as the fitted ridge line of the power tower in that direction.

[0010] S5, acquiring the lowest point among points that are located on the straight line where the fitted edge line lies and coincide with the non-base area as the initial ground point corresponding to the fitted edge line; acquiring the laser point that is located on the straight line where the fitted edge line lies and closest to the corresponding initial ground point in the real-time point cloud as the base endpoint of the fitted edge line;

[0011] S6, collecting the base endpoints of all fitted edge lines for plane fitting, taking the obtained fitting plane as the actually measured base plane, and acquiring the center point O' of the actually measured base plane;

[0012] S7, taking the included angle between the normal vector of the actually measured base plane and the vertical direction as the inclination angle θ of the power tower; taking the horizontal distance between the center point O of the power tower base plane and the center point O' of the actually measured base plane as the translation value, taking the vertical distance between the center point O of the power tower base plane and the center point O' of the actually measured base plane as the settlement value of the power tower, and calculating the translation direction angle λ of the power tower;

[0013]

[0014] wherein x3 is the X-axis coordinate of the center point O', y3 is the Y-axis coordinate of the center point O', and (x4, y4) is the coordinate of any two-dimensional coordinate point located on the crossarm direction of the currently measured power tower.

[0015] Preferably, a specified number of sections are selected as fitting sections according to the order from the smallest number of laser points to the largest;

[0016] Alternatively, the fitting sections are determined according to the following steps:

[0017] S311, projecting the real-time point cloud onto a vertical plane to form a two-dimensional point cloud map, and performing equidistant division on the two-dimensional point cloud map in the vertical direction to form a plurality of sections;

[0018] S312, counting the number of laser points on the two-dimensional point cloud map in each section, and establishing a variation curve;

[0019] S313, extracting the sections corresponding to wave troughs on the variation curve and the sections whose slopes are located in a set threshold interval as fitting sections;

[0020] the threshold interval is recorded as (f_min,f_max), -a1≤f_min≤0, 0<f_max≤a1, and a1 takes a value in the interval (0.05, 0.5).

[0021] Preferably, let the vertex of the currently measured power tower be marked as A2, the vertices of two power towers adjacent to the currently measured power tower be marked as A1 and A3 respectively, and the angular bisector BB' of the included angle ∠A1A2A3 is taken; in S311, the real-time point cloud is projected in the vertical plane where the angular bisector BB' lies to form a two-dimensional point cloud map.

[0022] Preferably, the method for fitting the real-time point cloud in the fitting segment to a feature plane includes the following steps:

[0023] S321. Traverse each laser point D in the fitted section. n Find the other laser points D in the fitted region. m With the laser point D n The horizontal plane M(D) n The distance dist(D) m ,M(D n ));

[0024] S322. Determine the laser point D that satisfies the optimization objective. n Horizontal plane M(D) n ) as the best-fit plane;

[0025] The optimization objective is:

[0026]

[0027] Where W represents the weight, L is the height of the fitted segment, and E is the set of laser points in the fitted segment;

[0028] S323. Calculate the outermost contour point cloud located on the best-fit plane in the real-time point cloud, and extract the plane enclosed by the outermost contour point cloud as the feature plane.

[0029] Preferably, in S323, the outermost contour point cloud on the best-fit plane is obtained using the 2D Point Set Surfaces algorithm.

[0030] Preferably, step S7 further includes: denoting the center point O' of the currently calculated measured base plane as O. t Let O' be the center point O' of the measured base plane obtained in the previous calculation. t-1 '; based on vector vector Move the non-base area to update its spatial coordinates, then return to step S1.

[0031] Preferably, the acquisition of the initial point cloud includes the following steps:

[0032] S11. Under normal conditions of the power tower, use lidar to collect point cloud data of the power tower as normal point cloud, and designate a target area S that covers the four bases of the power tower. This target area S covers the four bases in all directions.

[0033] S12. Extract the point cloud located in the target region S from the normal point cloud and denote it as the base normal point cloud. Obtain the voxel grid of the base normal point cloud.

[0034] S13. Cluster the normal point cloud of the base, identify the base region in the voxel grid, and use the point cloud in the base region of the normal point cloud of the base as the initial screening point cloud; divide the target region S into the base region and the non-base region.

[0035] S14. Filter the initial point cloud to remove noise and outliers. The filtered initial point cloud is the initial point cloud.

[0036] Preferably, when determining the center point O of the power tower base plane, firstly, plane fitting is performed on the laser point located at the bottom layer of each base area to obtain a closed polygon with the projection of the laser point at the bottom layer of each base area on the fitted plane as the vertex. The center point of this closed polygon is the center point O of the power tower base plane.

[0037] The present invention proposes a power tower morphology monitoring system based on lidar, which stores a computer program. When the computer program is executed, it is used to implement the power tower morphology monitoring method based on lidar.

[0038] The power tower morphology monitoring system based on lidar includes:

[0039] The initial module is used to designate the area covering the four bases of the power tower as the target area S from the normal point cloud of the power tower under normal conditions collected by lidar.

[0040] The preprocessing module is used to collect real-time point clouds of power towers and perform preprocessing.

[0041] The projection module is used to project real-time point clouds onto a vertical plane to form a two-dimensional point cloud map.

[0042] The extraction module is used to divide the two-dimensional point cloud map into segments, count the number of laser points in each segment, establish a change curve, and extract the fitted segment based on the trend of the change curve.

[0043] The fitting module is used to fit the real-time point cloud in the fitting segment, obtain the feature plane, and extract the corner points of each feature plane.

[0044] The search module is used to calculate the straight line equation based on the corner point of each feature plane in the same direction as the fitting ridge line, and search for the non-base area in the direction of the ground according to the straight line corresponding to each fitting ridge line to obtain the initial ground point on the fitting ridge line. The search module searches for the point in the real-time point cloud that is located on the fitting ridge line and is closest to the corresponding initial ground point as the base endpoint on the fitting ridge line.

[0045] The real-time shape recognition module is used to fit the measured base plane of the power tower being measured by combining all base endpoints.

[0046] Preferred options also include:

[0047] The data display module is used to calculate the center point O of the power tower base plane under normal conditions and the center point O' of the measured base plane. It combines the center point O and the center point O' to calculate the power tower tilt angle, settlement value, translation value, and translation direction angle, and then displays the data.

[0048] The early warning module is used to execute the set alarm strategy when the tilt angle of the power tower is greater than the set tilt threshold, or the translation value is greater than the set translation threshold, or the settlement value is greater than the set settlement threshold, or the translation direction angle is greater than the set offset threshold.

[0049] The advantages of this invention are:

[0050] (1) The present invention proposes a power tower morphology monitoring method based on lidar, which can measure the morphology of power towers in various practical scenarios and identify the specific degree of tilting, translation, and settlement, thus solving the monitoring needs of power transmission line towers. This method has a high degree of automation, a simple calculation process, and accurate and reliable calculation results, effectively improving the work efficiency of power inspection.

[0051] (2) The present invention uses the base rather than the ground as the basis for judgment to perform calculations, which avoids the influence of random changes in the ground and can further improve the accuracy of the calculation.

[0052] (3) By selecting fitting segments, this invention facilitates the construction of feature planes in towers with simple structures. This allows for the identification of the tower's ridgeline by combining the corner points of the feature planes, thereby ensuring the accuracy of the final measured base plane. In this invention, a two-dimensional point cloud map is formed by projecting the angle bisector BB' onto the vertical plane. Then, segment division and fitting segment selection are performed, which avoids the laser point height deviation caused by projection and ensures the consistency of segment division.

[0053] (4) In this invention, when extracting the initial point cloud from the normal point cloud, the discrete point set is divided into cubic voxels of the same size, which facilitates the establishment of a three-dimensional index data structure, which helps to speed up data processing and improve computational efficiency.

[0054] (5) The present invention proposes a power tower morphology monitoring system based on lidar, which realizes continuous monitoring and alarm of various factors such as early displacement and slight tilt of the tower by a highly integrated edge device, thereby realizing continuous monitoring, early judgment and timely alarm of power tower, thus avoiding serious tower collapse accidents that may occur later. Attached Figure Description

[0055] Figure 1This is a flowchart of a power tower morphology monitoring method based on lidar proposed in this invention;

[0056] Figure 2 Flowchart of the initial point cloud acquisition method;

[0057] Figure 3 A flowchart illustrating the method for obtaining the fitting segment and feature plane;

[0058] Figure 4 The diagram shown is a schematic of the power tower and target area S in Example 1;

[0059] Figure 4 (a) is Figure 4 The front view of the power tower shown;

[0060] Figure 4 (b) is a schematic diagram of the fitted ridgeline on the power tower in an inclined state;

[0061] Figure 5 The diagram shown is a schematic of the current measurement of the crossarm direction of the power tower;

[0062] Figure 6 The image shown is a schematic diagram of a two-dimensional point cloud map after real-time point cloud projection.

[0063] Figure 7 The image shown is a schematic diagram of a two-dimensional point cloud after it has been divided into grids.

[0064] Figure 8 As shown Figure 7 Curves showing the variation in the number of laser points in each segment of a 2D point cloud image;

[0065] Figure 9 The diagram shown is a schematic diagram of the corner points of the feature plane of the fitting section in Example 1;

[0066] Figure 10 The figure shown is a schematic diagram of the measured point cloud data of the base in Example 1;

[0067] Figure 11 The diagram shown is a schematic of the translation state of the power tower in Example 1. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] This embodiment proposes a method for monitoring the morphology of power towers based on lidar, which includes the following steps S1-S7.

[0070] S1. Under normal conditions, the power tower is collected by lidar, and an initial point cloud is extracted from the power tower's point cloud. This initial point cloud covers all the bases of the power tower. Based on this initial point cloud, the center point O of the power tower base plane is obtained, and the non-base areas are extracted.

[0071] Specifically, the normal state of a power tower refers to its condition when it has not tilted, shifted, or settled. In this step, after acquiring the laser point cloud of the power tower in its normal state using lidar, filtering is first performed. This can be done using smoothing filtering, voxel filtering, radius filtering, or a combination of these methods. A target region S is then specified, which is the coordinate space completely encompassing the four bases of the power tower from all directions. Figure 4 As shown in (a), for the filtered laser point cloud, the laser points located in the target region S can be extracted as the initial point cloud.

[0072] Reference Figure 2 In specific implementation, the following steps S11-S14 can be used to obtain the initial point cloud.

[0073] S11. Under normal conditions of the power tower, use lidar to collect point cloud data of the power tower as normal point cloud, and designate a target area S that covers the four bases of the power tower. This target area S covers the four bases in all directions.

[0074] In practice, S11 should maintain a unified field of view for at least 5 minutes to ensure that the normal point cloud is the overlap of multiple frames of data with the same field of view, so that the point cloud in the target area is dense enough.

[0075] S12. Extract the point cloud located in the target region S from the normal point cloud and denote it as the base normal point cloud. Obtain the voxel grid of the base normal point cloud.

[0076] S13. Cluster the normal point cloud of the base and identify the base region in the voxel grid. Use the point cloud in the base region of the normal point cloud of the base as the initial screening point cloud. The target region S consists of the base region and the non-base region. By identifying the base region, the non-base region is determined accordingly.

[0077] In S13, when clustering the normal point cloud of the base, one laser point in each base can be designated as the root node first, and then a tree structure can be used to expand the clustering of all laser points in the target area to identify the base area in the voxel grid; for example, for a four-corner power tower, four base areas can be identified, corresponding to the four bases of the power tower.

[0078] S14. Filter the initial point cloud to remove noise and outliers. The filtered initial point cloud is the initial point cloud.

[0079] When determining the center point O of the power tower base plane, firstly, plane fitting is performed by combining the laser point at the bottom layer of each base area. Then, a closed polygon with the projection of the laser point at the bottom layer of each base area on the fitted plane as the vertex is obtained. The center point of this closed polygon is the center point O of the power tower base plane.

[0080] Taking a four-corner power tower as an example, the center point O of the power tower base plane is determined as follows: select the lowest laser point in each base area, a total of 4 points, perform plane fitting based on the 4 laser points, project the 4 laser points onto the fitting plane to obtain 4 projection points, and take the center point of the quadrilateral with the 4 projection points as vertices as the center point O of the power tower base plane.

[0081] S2. In monitoring mode, real-time point cloud data of power towers is collected using lidar.

[0082] Specifically, in step S2, the point cloud of the power tower collected by the lidar is filtered and used as the real-time point cloud.

[0083] In practice, the point cloud data of power towers collected by lidar can first be subjected to voxel filtering to achieve downsampling and remove redundant data; then the downsampled point cloud data can be subjected to radius filtering to remove point clouds outside a specific range to remove noise; and finally, the point cloud data after radius filtering can be subjected to statistical filtering to remove outliers.

[0084] S3. Divide the real-time point cloud into multiple segments in the vertical direction, select at least two segments as fitting segments, and perform horizontal plane fitting on the laser points in the real-time point cloud located in the fitting segments. The horizontal plane obtained by fitting is used as the feature plane corresponding to the segment.

[0085] Specifically, in this embodiment, the fitting segment can be selected based on the number of laser points in the segment. For example, the segment with a smaller laser point density can be selected as the fitting segment to ensure that the feature plane is extracted at the tower body of the power tower, so as to avoid interference from the outer extension areas such as crossarms and insulators with dense laser points on the tower body morphology monitoring.

[0086] In order to ensure the number of laser points in each section and to ensure that the feature plane tends to the cross-section of the power tower, in this embodiment, the height L of the section can be taken in the range of [0.5 meters, 5.5 meters].

[0087] Reference Figure 3 In this embodiment, the selection of the fitting segment specifically includes the following sub-steps S311-S313.

[0088] S311, projecting real-time point cloud onto a vertical plane to form a two-dimensional point cloud map, and equally dividing the two-dimensional point cloud map in the vertical direction to form a plurality of sections;

[0089] S312, counting the number of laser points on the two-dimensional point cloud map in each section, and establishing a variation curve;

[0090] S313, extracting the sections corresponding to wave troughs on the variation curve and the sections with slopes within a set threshold interval as fitting sections.

[0091] The threshold interval is recorded as (f_min,f_max), -a1≤f_min≤0, 0<f_max≤a1, a1 can specifically take a value in the interval (0.05, 0.5), for example, 0.3.

[0092] In the above step S311, in order to avoid the height deviation of laser points caused by projection, let the vertex of the currently measured power pylon be A2, and the vertices of the two power pylons adjacent to the currently measured power pylon be A1 and A3 respectively, and the angle bisector BB' of the included angle ∠A1A2A3 is obtained, which is specifically as Figure 5 shown, wherein θ1=θ2. The real-time point cloud is projected in the vertical plane where the angle bisector BB' is located to form a two-dimensional point cloud map.

[0093] In this embodiment, any existing plane fitting method or tool can be used to fit the real-time point cloud in the fitting sections into a feature plane. The feature plane is located in the real-time point cloud area of the fitting sections and tends to the cross-sectional shape of the power pylon.

[0094] S4, extracting corner points of each feature plane located at the same orientation for straight line fitting, and taking the straight line obtained by fitting as the fitted edge line of the power pylon at this orientation.

[0095] Specifically, in this embodiment, the Pipe algorithm can be used to calculate the corner points of the feature plane, and the RANSAC spatial straight line fitting method is used to obtain the fitted edge line.

[0096] S5, acquiring a non-base area in an initial point cloud, acquiring the lowest point among points located on the straight line where the fitted edge line is located and overlapping with the non-base area as the initial ground point corresponding to the fitted edge line; acquiring the laser point located on the straight line where the fitted edge line is located and closest to the corresponding initial ground point in the real-time point cloud as the base end point of the fitted edge line.

[0097] During specific implementation, in step S13, all spatial coordinate points located outside the base area in the target area can be further taken as the non-base area.

[0098] As Figure 4 shown in the embodiment of a power pylon, Figure 4(a) is a front view of the bottom of the power tower in its normal state, showing that the target area S completely covers the four bases; Figure 4 (b) is a bottom front view of the power tower when it is tilted, with straight lines Z1 and Z2 being two fitted ridges. (Combined with...) Figure 4 (b) It can be seen that the lowest point among the points on line Z1 that coincide with the non-base region is point n1, and the nearest point in the real-time point cloud found upwards from point n1 along line Z1 tends towards point n2; the lowest point among the points on line Z2 that coincide with the non-base region is point n3, and the nearest point in the real-time point cloud found upwards from point n3 along line Z2 tends towards point n4. Therefore, in this embodiment, the base endpoints found based on the fitted ridgeline tend towards the center position of the base bottom of the power tower in that orientation under real-time conditions, and the fitted plane formed by the base endpoints tends towards the bottom plane of the real-time power tower.

[0099] It is worth noting that the target area S is a spatial coordinate area manually specified based on the power tower under normal conditions, and the non-base area is the spatial area of ​​the target area S excluding all the bases of the power tower under normal conditions. Therefore, the spatial coordinates of the non-base area are fixed, and the initial ground points of each fitted ridgeline can be directly determined through spatial positioning.

[0100] S6. Perform plane fitting on the base endpoints of all fitted ridges, and use the fitted plane as the measured base plane. Obtain the center point O' of the measured base plane.

[0101] The center point O' of the measured base plane is the center point of the closed polygon whose vertices are the projections of each base endpoint onto the measured base plane.

[0102] S7. Obtain the angle between the normal vector of the measured base plane and the vertical direction as the tilt angle θ of the power tower; obtain the horizontal distance between the center point O of the power tower base plane and the center point O' of the measured base plane as the translation value, and obtain the vertical distance between the center point O and the center point O' as the settlement value of the power tower; the calculation formula for the translation direction angle λ of the power tower is as follows:

[0103]

[0104] Where x3 is the X-axis coordinate of the center point O', y3 is the Y-axis coordinate of the center point O', that is, (x3, y3) is the two-dimensional coordinate of the center point O'; (x4, y4) is the coordinate of any two-dimensional coordinate point located in the crossarm direction of the current measuring power tower.

[0105] θ can also be calculated directly using the following formula:

[0106]

[0107] Where (x1,y1,z1) and (x2,y2,z2) are the three-dimensional coordinates of two arbitrary points on the normal vector of the measured base plane, and π is the value of pi.

[0108] The above-mentioned method for monitoring the morphology of power towers based on lidar is verified below with specific embodiments.

[0109] Reference Figure 3 In the following embodiments, the method for fitting the real-time point cloud in the fitting segment to a feature plane includes the following steps S321-S323.

[0110] S321. Traverse each laser point D in the fitted section. n Find the other laser points D in the fitted region. m With the laser point D n The horizontal plane M(D) n The distance dist(D) m ,M(D n ));

[0111]

[0112] in, This represents the normal vector in the vertical direction, i.e., dist(D) m ,M(D n )) represents laser point D m With the laser point D n The distance in the vertical direction.

[0113] S322. Determine the laser point D that satisfies the optimization objective. n Horizontal plane M(D) n ) as the best-fit plane;

[0114] The optimization objective is:

[0115]

[0116] Where W represents the weight, L is the height of the fitted segment, and E is the set of laser points in the fitted segment.

[0117] S323. Calculate the outermost contour point cloud located on the best-fit plane in the real-time point cloud, and extract the plane enclosed by the outermost contour point cloud as the feature plane.

[0118] In practice, the 2DPoint Set Surfaces algorithm can be used to obtain the outermost contour point cloud on the best-fit plane.

[0119] Example 1

[0120] by Figure 4The four-corner power tower shown is an example.

[0121] In this embodiment, when collecting the initial point cloud, the lidar continuously collects data for 5 minutes in each field of view to ensure the initial point cloud density.

[0122] In this embodiment, the power tower is projected onto the vertical plane where BB' is located to obtain a two-dimensional point cloud image, as shown below. Figure 6 As shown;

[0123] like Figure 7 As shown, a 5m × 5m grid is used to divide the two-dimensional point cloud map. The grid height is the height of the segment, that is, the segment height L is 5 meters. In this embodiment, the power tower is divided into 16 segments.

[0124] like Figure 8 As shown, the curves of the change in the number of laser points on the two-dimensional point cloud map in each segment are plotted. The extreme value segment corresponding to the trough and the smooth segment with 0 < slope < 0.3 are taken as the fitting segment. That is, the fitting segment includes a total of 9 segments: segment L1, L5-L11, and L15.

[0125] In this embodiment, after obtaining the best fitting plane for each fitting segment, the 2DPointSetSurfaces algorithm is used to obtain the outermost contour point cloud of the best fitting plane as the feature plane, and the Pipe algorithm is used to calculate the four corner points in the outermost contour point cloud. In this embodiment, taking segments L9 and L8 as examples, the four corner points calculated for segment L9 are D91, D92, D93, and D94, and the four corner points calculated for segment L8 are D81, D82, D83, and D84; the spatial distribution of D91, D92, D93, D94, D81, D82, D83, and D84 is as follows. Figure 9 As shown.

[0126] In this embodiment, the four fitted ridges Z1, Z2, Z3, and Z4 obtained using the RANSAC spatial straight line fitting method are as follows: Figure 10 As shown.

[0127] In this embodiment, based on the KD-tree algorithm, starting from the corner point in the highest fitting segment, the search is performed downwards along the fitting ridgeline until the point D0, which is the lowest point on the fitting ridgeline in the non-base region, is reached. The last point in the real-time point cloud on the fitting ridgeline before reaching D0 is taken as the endpoint of the measured base, i.e., the base endpoint corresponding to the fitting ridgeline.

[0128] like Figure 10 In the equation, after the line containing Z1 touches the corresponding point D0, it will touch the point D before D0. Z1 As the base endpoint corresponding to the fitted ridge Z1.

[0129] After the line containing Z2 touches the corresponding point D0, it will touch the point D before D0. Z2 As the base endpoint corresponding to the fitted ridge Z2.

[0130] After the line containing Z3 touches the corresponding point D0, it will touch the point D before D0. Z3 As the base endpoint corresponding to the fitted ridge Z3.

[0131] After the line containing Z4 touches the corresponding point D0, it will touch the point D before D0. Z4 As the base endpoint corresponding to the fitted ridge Z4.

[0132] It is worth noting that in this embodiment, all coordinates within the target area other than the base area are considered as non-base areas, thus maximizing the range of non-base areas. This ensures that searching downwards along the line corresponding to the linear equation will inevitably reach point D0 in the non-base area, thereby stopping the search and improving search accuracy.

[0133] like Figure 11 As shown, in this embodiment, point D is obtained. Z1 D Z2 D Z3 D Z4 The fitted plane is used as the actual base plane, and the center point O' of the actual base plane is obtained. The two-dimensional coordinates of the center point O' on the horizontal plane are marked as (x3, y3). A point O1(x4, y4) is randomly selected on the crossarm direction B1 of the current power pole. Then, O'(x3, y3) and O1(x4, y4) can be substituted into the above calculation formula for the translation direction angle λ to calculate the translation direction angle λ of the current power pole. Figure 11 In the middle, direction B2 is perpendicular to direction B1.

[0134] In this embodiment, the calculated offset data of the power tower is compared with the measured value in Table 1 below.

[0135] Table 1: Measured and Actual Values ​​of Power Tower Offset in Example 1

[0136] Measured values 0.6869° 0.59° 1.98 0 Measured value 0.6722° 0.54° 2.94 0.2 error 0.0147° 0.05° 0.96 0.2

[0137] Example 2

[0138] In this embodiment, the same parameter settings as in Embodiment 2 were used to measure another power tower. The measured values ​​and actual values ​​of the offset data of the power tower are compared as shown in Table 2 below.

[0139] Table 2: Measured and Actual Values ​​of Power Tower Offset in Example 2

[0140] Measured values 0.3871° 0.45° 3.15 1.5 Measured value 0.3992° 0.51° 1.94 1.35 error 0.0121° 0.06° 1.21 0.15

[0141] Table 3: Measurement errors in Examples 1 and 2

[0142] Example 1 Error 0.0147° 0.05° 0.96 0.2 Example 2 Error 0.0121° 0.06° 1.21 0.15 average error 0.0134° 0.055° 1.085 0.175

[0143] As can be seen from Table 3, the power tower offset value calculated by the LiDAR-based power tower morphology monitoring method provided by the present invention is extremely close to the measured power tower offset value, which proves that the LiDAR-based power tower morphology monitoring method provided by the present invention has high calculation accuracy, can realize effective monitoring of power towers, and has high reliability.

[0144] In another embodiment, the power tower morphology monitoring method based on lidar proposed in this invention further includes step S7: denoting the currently calculated center point O' as O t Let the center point O' obtained in the previous calculation be denoted as O. t-1 '; when O t When 'O' is the center point of the first calculation, O t-1 'That is, the center point O of the power tower base plane obtained based on the initial point cloud; according to the vector vector Move the target region S, extract the non-base region from the updated target region S, and then return to step S1.

[0145] As can be seen, in this embodiment, steps S1-S7 are executed cyclically. In each calculation of the tilt angle θ, translation value, translation direction angle λ, and settlement value, the center point O is always the initial base center point. However, the non-base area used in each calculation is the non-base area updated based on the previous calculation. That is, the target area S is adaptively changed based on the tilt angle θ, translation value, translation direction angle λ, and settlement value calculated in S7 of the current cycle to obtain the updated target area S'. In other words, the non-base area of ​​the target area S' is used as the new non-base area and substituted into the calculation of the initial ground point for the next fitting ridge search.

[0146] It is worth noting that the initial target area S has been divided into a base area and a non-base area. Therefore, the update of the non-base area is essentially the updated non-base area being tilted relative to the unupdated non-base area by the tilt angle added in this round of calculation, translated by the translation value and translation direction angle added in this round of calculation, and settled by the settlement value added in this round of calculation.

[0147] By continuously refining the specific shape of the target area, the selection of the target area, especially the non-base area, follows the current spatial shape of the power tower. This ensures that the calculation in each cycle remains consistent with the environment, resulting in a stable error rate and avoiding cumulative errors caused by the mismatch between the power tower position and the initially set target area after extreme tilting, translation, or settlement.

[0148] Of course, those skilled in the art will recognize that the present invention is not limited to the details of the exemplary embodiments described above, but also includes the same or similar structures that can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0149] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0150] The technologies, shapes, and structures not described in detail in this invention are all known technologies.

Claims

1. A method for monitoring the morphology of power towers based on lidar, characterized in that, Comprising the following steps: S1, when the power pylon is in a normal state, collecting point clouds of the power pylon by means of a lidar, and extracting an initial point cloud from the point clouds of the power pylon, wherein the initial point cloud covers all bases of the power pylon; acquiring a center point O of the base plane of the power pylon based on the initial point cloud and extracting a non-base area; S2, in a monitoring state, collecting real-time point clouds of the power pylon by means of a lidar; S3, dividing the real-time point cloud into a plurality of sections in a vertical direction, selecting at least two sections as fitting sections, performing horizontal plane fitting on laser points located in the fitting sections in the real-time point cloud, and taking the obtained horizontal plane from fitting as a feature plane corresponding to the section; S4, extracting corner points, located at the same orientation, of each feature plane to perform straight line fitting, and taking the obtained straight line from fitting as a fitted edge line of the power pylon at the orientation; S5, acquiring the lowest point among points that are located on the straight line where the fitted edge line is located and coincide with the non-base area as an initial ground point corresponding to the fitted edge line; acquiring the laser point that is located on the straight line where the fitted edge line is located in the real-time point cloud and closest to the corresponding initial ground point as a base end point of the fitted edge line; S6, collecting base end points of all fitted edge lines to perform plane fitting, taking the obtained fitting plane as an actually measured base plane, and acquiring a center point O' of the actually measured base plane; S7, acquiring an included angle between a normal vector of the actually measured base plane and the vertical direction as the inclination angle θ of the power pylon; acquiring the horizontal distance between the center point O of the base plane of the power pylon and the center point O' of the actually measured base plane as a translation value, acquiring the vertical distance between the center point O of the base plane of the power pylon and the center point O' of the actually measured base plane as the settlement value of the power pylon, and calculating the translation direction angle λ of the power pylon; wherein x3 is the X-axis coordinate of the center point O', y3 is the Y-axis coordinate of the center point O', and (x4, y4) is the coordinate of any two-dimensional coordinate point located in the cross arm direction of the currently measured power pylon.

2. The power tower morphology monitoring method based on lidar as described in claim 1, characterized in that, selecting a specified number of sections as fitting sections according to the order of the number of laser points from less to more; alternatively, the fitting sections are determined according to the following steps: S311, projecting the real-time point cloud onto a vertical plane to form a two-dimensional point cloud map, and equally dividing the two-dimensional point cloud map in the vertical direction to form a plurality of sections; S312, counting the number of laser points on the two-dimensional point cloud map in each section, and establishing a change curve; S313, extracting the section corresponding to a wave trough on the change curve and the section with a slope located in a set threshold interval as fitting sections; the threshold interval is recorded as (f_min,f_max), -a1≤f_min≤0, 0<f_max≤a1, and a1 takes a value on the interval (0.05,0.5).

3. The method for monitoring the morphology of power towers based on lidar as described in claim 2, characterized in that, letting the vertex of the currently measured power pylon be A2, the vertices of two power pylons adjacent to the currently measured power pylon be A1 and A3 respectively, and taking the angular bisector BB' of the included angle ∠A1A2A3; in S311, projecting the real-time point cloud in the vertical plane where the angular bisector BB' is located to form a two-dimensional point cloud map.

4. The power tower morphology monitoring method based on lidar as described in claim 1, characterized in that, the method for fitting the real-time point cloud in the fitting sections into a feature plane comprises the following steps: S321. Traverse each laser point D in the fitted section. n Find the other laser points D in the fitted region. m With the laser point D n The horizontal plane M(D) n The distance dist(D) m ,M(D n )); S322. Determine the laser point D that satisfies the optimization objective. n Horizontal plane M(D) n ) as the best-fit plane; The optimization objective is: Where W represents the weight, L is the height of the fitted segment, and E is the set of laser points in the fitted segment; S323. Calculate the outermost contour point cloud located on the best-fit plane in the real-time point cloud, and extract the plane enclosed by the outermost contour point cloud as the feature plane.

5. The power tower morphology monitoring method based on lidar as described in claim 4, characterized in that, In S323, the 2D Point Set Surfaces algorithm is used to obtain the outermost contour point cloud on the best-fit plane.

6. The method for monitoring the morphology of power towers based on lidar as described in claim 1, characterized in that, S7 also includes: Let the center point O' of the currently calculated measured base plane be denoted as O. t Let O' be the center point O' of the measured base plane obtained in the previous calculation. t-1 '; based on vector vector Move the non-base area to update its spatial coordinates, then return to step S1.

7. The method for monitoring the morphology of power towers based on lidar as described in claim 1, characterized in that, Obtaining the initial point cloud includes the following steps: S11. Under normal conditions of the power tower, use lidar to collect point cloud data of the power tower as normal point cloud, and designate a target area S that covers the four bases of the power tower. This target area S covers the four bases in all directions. S12. Extract the point cloud located in the target region S from the normal point cloud and denote it as the base normal point cloud. Obtain the voxel grid of the base normal point cloud. S13. Cluster the normal point cloud of the base, identify the base region in the voxel grid, and use the point cloud in the base region of the normal point cloud of the base as the initial screening point cloud; divide the target region S into the base region and the non-base region. S14. Filter the initial point cloud to remove noise and outliers. The filtered initial point cloud is the initial point cloud.

8. The method for monitoring the morphology of power towers based on lidar as described in claim 1, characterized in that, When determining the center point O of the power tower base plane, firstly, perform plane fitting by combining the laser point at the bottom layer of each base area to obtain a closed polygon with the projection of the laser point at the bottom layer of each base area on the fitted plane as the vertex. The center point of this closed polygon is the center point O of the power tower base plane.

9. A power tower morphology monitoring system based on lidar, characterized in that, The device contains a computer program, which, when executed, is used to implement the power tower morphology monitoring method based on lidar as described in any one of claims 1-8. The power tower morphology monitoring system based on lidar includes: The initial module is used to designate the area covering the four bases of the power tower as the target area S from the normal point cloud of the power tower under normal conditions collected by lidar. The preprocessing module is used to collect real-time point clouds of power towers and perform preprocessing. The projection module is used to project real-time point clouds onto a vertical plane to form a two-dimensional point cloud map. The extraction module is used to divide the two-dimensional point cloud map into segments, count the number of laser points in each segment, establish a change curve, and extract the fitted segment based on the trend of the change curve. The fitting module is used to fit the real-time point cloud in the fitting segment, obtain the feature plane, and extract the corner points of each feature plane. The search module is used to calculate the straight line equation based on the corner point of each feature plane in the same direction as the fitting ridge line, and search for the non-base area in the direction of the ground according to the straight line corresponding to each fitting ridge line to obtain the initial ground point on the fitting ridge line. The search module searches for the point in the real-time point cloud that is located on the fitting ridge line and is closest to the corresponding initial ground point as the base endpoint on the fitting ridge line. The real-time shape recognition module is used to fit the measured base plane of the power tower being measured by combining all base endpoints.

10. The power tower morphology monitoring system based on lidar as described in claim 9, characterized in that, Also includes: The data display module is used to calculate the center point O of the power tower base plane under normal conditions and the center point O' of the measured base plane. It combines the center point O and the center point O' to calculate the power tower tilt angle, settlement value, translation value, and translation direction angle, and then displays the data. The early warning module is used to execute the set alarm strategy when the tilt angle of the power tower is greater than the set tilt threshold, or the translation value is greater than the set translation threshold, or the settlement value is greater than the set settlement threshold, or the translation direction angle is greater than the set offset threshold.

Images