Method and device for monitoring hidden danger of tree barrier of power transmission line

Abstract

The application discloses a kind of transmission line tree barrier hidden danger hazards monitoring method and device, based on transmission passage monitoring image and transmission passage point cloud carries out tree barrier hidden danger monitoring, the method includes the following steps: by connectivity, the trees in transmission passage point cloud is divided into multiple groups;Each group of trees in transmission passage point cloud is mapped to the monitoring image;Identify the tree area in monitoring image, and compare with the tree area that point cloud is mapped to monitoring image;According to the comparison result, the height of the tree in image is calculated, and the distance between the tree in image and the conductor of transmission line is calculated.The application can quickly determine the damage degree of tree barrier hidden danger when carrying out transmission line passage hidden danger operation, avoids the damage caused by tree growth to conductor, and provides technical support for transmission line safety guarantee.

Description

TECHNICAL FIELD

[0001] The application relates to a monitoring method and device for tree barrier hidden dangers of a power transmission line, and belongs to the technical field of intelligent operation inspection of the power transmission line. BACKGROUND

[0002] The power transmission channel scene is complex and diverse, and many trees are near the power transmission line. Because the distance between the trees and the conductor is too small, the discharge phenomenon of the conductor is easily caused. Therefore, the growth of the trees near the power transmission line brings great risk to the power transmission line.

[0003] When encountering strong wind weather, the conductor and the trees will sway, and the distance between the trees and the conductor will be further reduced, which is more likely to cause accidents. The tree barrier hidden danger is a problem that the intelligent inspection of the power transmission line focuses on. The current commonly used prevention and control means include manual inspection, visual monitoring, unmanned aerial vehicle inspection and the like. The above-mentioned inspection methods can effectively avoid the harm of the tree barrier hidden danger, but at the same time, there are also some deficiencies. The manual inspection method needs to consume a large amount of labor cost, the visual monitoring method has poor accuracy in judging the real distance, and the unmanned aerial vehicle inspection is expensive and cannot be frequently inspected.

[0004] To sum up, how to provide an efficient, accurate and low-cost monitoring method for the tree barrier hidden danger of the power transmission line and reduce the tree barrier hidden danger is one of the problems to be solved by the technical personnel in the field at present. SUMMARY

[0005] In order to solve the above-mentioned problems, the application provides a monitoring method and device for tree barrier hidden dangers of a power transmission line, which can determine the damage degree of the tree barrier hidden danger and avoid the damage of the tree growth to the conductor, thereby providing technical support for the safety protection of the power transmission line.

[0006] The technical scheme adopted by the application to solve the technical problems is as follows:

[0007] In a first aspect, the application provides a monitoring method for tree barrier hidden dangers of a power transmission line. The method is based on monitoring images of a power transmission channel and point clouds of the power transmission channel to monitor tree barrier hidden dangers. The method comprises the following steps:

[0008] The trees in the point clouds of the power transmission channel are divided into multiple groups through connectivity;

[0009] Each group of trees in the point clouds of the power transmission channel is mapped to the monitoring images;

[0010] The tree regions in the monitoring images are identified and compared with the tree regions in the monitoring images mapped by the point clouds;

[0011] The height of the trees in the images is calculated according to the comparison result, and the distance between the trees in the images and the conductor of the power transmission line is calculated.

[0012] As a possible implementation manner of the embodiment, the dividing the trees in the power transmission channel point cloud into groups through connectivity comprises:

[0013] a1: setting a segmentation threshold s, a target point p0, and a tree point set T={p0};

[0014] a2: adding all points with a distance less than s from the target point to the set T, i.e., T={p0, p1, p2,...}, and removing the points added to the set from the point cloud;

[0015] a3: setting the target point as the next point of the current target point;

[0016] a4: repeating steps a2 and a3 until the set T is no longer increased, and the set T is a divided tree region;

[0017] a5: reselecting the target point p0 in the point cloud, and repeating steps a1 to a4 until all points in the point cloud are removed;

[0018] a6: grouping all sets T with more than 100 points into a tree set Tree.

[0019] As a possible implementation manner of the embodiment, the mapping each group of trees in the power transmission channel point cloud to the monitoring image comprises:

[0020] Mapping each group of trees in the power transmission channel point cloud to the monitoring image, recording pixel points and corresponding three-dimensional points, and a mapping formula for mapping each group of trees in the power transmission channel point cloud to the monitoring image is:

[0021] temp=(p-T)·R

[0022] m=(M·temp T ) T

[0023] m=m÷m(0,2)(homogeneous transformation)

[0024] wherein p is a three-dimensional point, m is a two-dimensional point, R is a point cloud rotation matrix, T is a point cloud translation vector, and M is a camera intrinsic matrix.

[0025] As a possible implementation manner of the embodiment, the identifying a tree region in the monitoring image and comparing the tree region with a tree region mapped by the point cloud to the monitoring image comprises:

[0026] c1: a set of tree regions mapped by the point cloud to the monitoring image is recorded as A, and a set of tree regions in the monitoring image is recorded as D;

[0027] c2: find the upper bound of each tree mapping region in set A, and form set topA, find the upper bound of each tree identification region in set D, and form set topD;

[0028] c3: each upper bound in topA is recorded as ta a , find the pixel point corresponding to ta in all upper bounds in topD.

[0029] As a possible implementation manner of the embodiment, the corresponding pixel point refers to having the same x coordinate in the pixel coordinate system, and the y coordinate difference is less than 20, and the y coordinate difference is greater than 20, which is considered as a matching error.

[0030] As a possible implementation manner of the embodiment, the calculating the height of the tree in the image according to the comparison result, and calculating the distance between the tree in the image and the conductor of the power transmission line, comprises:

[0031] d1: each upper bound ta in topA a is recorded as m a (u a ,v a ), the three-dimensional point corresponding to m a (u a ,v a ) is recorded as p a (x a ,y a ,z a ), the pixel point corresponding to m a (u a ,v a ) is recorded as d a (u a ,v d ), if m a (u a ,v a ) has no corresponding pixel point, skip the pixel point;

[0032] d2: calculate the pixel difference: diff a =(v d -v a );

[0033] d3: according to the camera imaging principle , calculate the corresponding real height H a of diff a , wherein f is the focal length, L is the distance between the object plane and the camera plane, h is the pixel height, and H is the real height;

[0034] d4: calculate the three-dimensional point p a (x a ,y a ,za ) corresponding three-dimensional coordinates p d (x a ,y a ,z a +H a );

[0035] d5: calculate the Euclidean distance of all guide points and three-dimensional coordinates p d (x a ,y a ,z a +H a ) to find the nearest guide point and the nearest distance;

[0036] d6: the upper boundary t a corresponding to all guide nearest point sets Near, all nearest distance sets Dis;

[0037] d7: select the minimum value in set Dis as the shortest distance of the tree area distance from the guide.

[0038] In a second aspect, the embodiments of the present application provide a transmission line tree barrier hidden danger monitoring device, which is based on transmission channel monitoring image and transmission channel point cloud to monitor tree barrier hidden danger. The device comprises:

[0039] a tree grouping module, configured to divide trees in the transmission channel point cloud into multiple groups through connectivity;

[0040] a point cloud mapping module, configured to map each group of trees in the transmission channel point cloud to the monitoring image;

[0041] a tree area comparison module, configured to identify a tree area in the monitoring image and compare the tree area with a tree area mapped to the monitoring image by the point cloud;

[0042] a distance calculation module, configured to calculate a height of a tree in the image according to a comparison result, and calculate a distance between the tree in the image and a guide wire of the transmission line.

[0043] As a possible implementation manner of the embodiments, the specific process that the tree grouping module divides the trees in the transmission channel point cloud into multiple groups through connectivity is as follows:

[0044] a1: set a segmentation threshold s, a target point p0, and a tree point set T={p0};

[0045] a2: add all points with a distance less than s from the target point to the set T, i.e. T={p0, p1, p2,...}, and remove the points added to the set from the point cloud;

[0046] a3: set the target point as the next point of the current target point;

[0047] a4: repeat step a2 and a3 until the set T is no longer increased, and the set T is a divided tree region;

[0048] a5: reselect the target point p0 in the point cloud, and repeat steps a1 to a4 until all points in the point cloud are removed;

[0049] a6: group all sets T with more than 100 points into a tree set Tree.

[0050] As a possible implementation manner of the present embodiment, the mapping formula of the point cloud mapping module for mapping each group of trees in the power transmission channel point cloud to the monitoring image is:

[0051] temp = (p-T)·R

[0052] m = (M·temp T ) T

[0053] m = m ÷ m (0,2) (homogeneous transformation)

[0054] Wherein, p is a three-dimensional point, m is a two-dimensional point, R is a point cloud rotation matrix, T is a point cloud translation vector, and M is a camera intrinsic matrix.

[0055] As a possible implementation manner of the present embodiment, the specific process of the tree region comparison module for identifying the tree region in the monitoring image and comparing it with the tree region mapped by the point cloud on the monitoring image is:

[0056] c1: the tree region set mapped by the point cloud on the monitoring image is recorded as A, and the tree region set identified in the monitoring image is recorded as D;

[0057] c2: find the upper boundary of each tree mapping region in set A to form set topA, and find the upper boundary of each tree identification region in set D to form set topD;

[0058] c3: each upper boundary in topA is recorded as ta a , and find the pixel point corresponding to ta in all upper boundaries in topD.

[0059] As a possible implementation manner of the present embodiment, the corresponding pixel point refers to having the same x coordinate in the pixel coordinate system, and the y coordinate difference is less than 20, and the y coordinate difference greater than 20 is considered as a matching error.

[0060] As a possible implementation manner of the present embodiment, the specific process of the distance calculation module for calculating the height of the tree in the image according to the comparison result, and calculating the distance between the tree in the image and the power line conductor is:

[0061] d1: Upper boundary t a Each pixel in the diagram is denoted as m. a (u a ,v a ), m a (u a ,v a The corresponding three-dimensional point is denoted as p. a (x a ,y a ,z a ), m a (u a ,v a The corresponding pixel is denoted as d. a (u a ,v d If m a (u a ,v a If there is no corresponding pixel, skip that pixel.

[0062] d2: Calculate pixel difference: diff a =(v d -v a );

[0063] d3: Based on the camera imaging principle Calculate diff a Corresponding actual height H a Where f is the focal length, L is the distance between the object plane and the camera plane, h is the pixel height, and H is the actual height;

[0064] d4: Calculate the 3D point p of the tree a (x a ,y a ,z a The corresponding three-dimensional coordinates p d (x a ,y a ,z a +H a );

[0065] d5: Calculate the coordinates of all traverse points and the three-dimensional coordinates p. d (x a ,y a ,z a +H a Find the nearest traverse point and its closest distance using the Euclidean distance from the traverse point.

[0066] d6: Upper boundary t a The set Near consists of all the nearest points of the corresponding wires, and the set Dis consists of all the nearest distances.

[0067] d7: select the minimum value in the set Dis as the shortest distance of the tree region to the conductor.

[0068] The technical scheme of the embodiment of the application can have the following beneficial effects:

[0069] The embodiment of the application adopts tree connected domain division, image and point cloud mapping, region comparison and other technologies, realizes automatic division of trees in the power transmission line channel, calculates the shortest distance of each tree region to the conductor, and infers the tree growth in the image according to the tree recognition result, can divide the risk level according to the tree distance measurement result, discovers and handles the tree barrier hidden danger that can cause harm to the power transmission channel line in time, strengthens the management and control of the power transmission channel, and the operation and inspection personnel can more accurately manage and control the tree height, and reduces accidents caused by tree barrier hidden dangers.

[0070] The application is based on monitoring images and three-dimensional point clouds, compared with artificial inspection or visual inspection method, greatly improves the accuracy and control of tree barrier hidden danger distance measurement, and the inspection is more convenient.

[0071] The application can collect point clouds only once, and then measure the tree barrier hidden danger distance through subsequent monitoring images for a long time, and if the scene changes greatly, update the point clouds again, compared with the unmanned aerial vehicle inspection or point cloud analysis method, greatly reduces the inspection cost.

[0072] The application calculates the shortest distance between the tree and the conductor, can divide the hidden danger level according to different distances, and better controls the power transmission line.

[0073] The application can quickly determine the harm degree of tree barrier hidden danger when inspecting the hidden danger of the power transmission line channel, avoids the damage of tree growth to the conductor, and provides technical support for the safety protection of the power transmission line. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 It is a flow chart of a power transmission line tree barrier hidden danger monitoring method according to an example embodiment;

[0075] Figure 2 It is a schematic diagram of a power transmission line tree barrier hidden danger monitoring device according to an example embodiment;

[0076] Figure 3 It is a monitoring image of a power transmission channel according to an example embodiment;

[0077] Figure 4 It is a power transmission line point cloud diagram in a power transmission channel according to an example embodiment;

[0078] Figure 5A point cloud diagram of trees in a power transmission channel according to an exemplary embodiment is shown.

[0079] Figure 6 A schematic diagram of a qualified tree area is obtained according to an exemplary embodiment is shown.

[0080] Figure 7 A mapping effect diagram of trees in a power transmission channel point cloud mapped onto a monitoring image according to an exemplary embodiment is shown.

[0081] Figure 8 A tree area recognition effect diagram in a monitoring image according to an exemplary embodiment is shown.

[0082] Figure 9 A ranging result diagram of trees and power transmission line conductors according to an exemplary embodiment is shown. DETAILED DESCRIPTION

[0083] The application is further described below in conjunction with the accompanying drawings and embodiments:

[0084] In order to clearly illustrate the technical features of the present application, the following detailed description will be given in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples of implementing the present application. In order to simplify the disclosure of the present application, the components and settings of specific examples are described below. In addition, the present application can repeatedly refer to numerals and / or letters in different examples. Such repetition is for the purpose of simplification and clarity, and does not in itself indicate the relationship between the various embodiments and / or settings being discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. The present application omits the description of well-known components and processing techniques and processes to avoid unnecessary limitation of the present application.

[0085] As shown in Figure 1 The monitoring method for tree barrier hazards of a power transmission line provided by the embodiment of the present application is based on a monitoring image of a power transmission channel and a point cloud of the power transmission channel to monitor tree barrier hazards. The method comprises the following steps:

[0086] The trees in the point cloud of the power transmission channel are divided into multiple groups through connectivity;

[0087] Each group of trees in the point cloud of the power transmission channel is mapped onto the monitoring image;

[0088] The tree area in the monitoring image is identified and compared with the tree area mapped onto the monitoring image by the point cloud;

[0089] The height of the trees in the image is calculated according to the comparison result, and the distance between the trees in the image and the conductors of the power transmission line is calculated.

[0090] As a possible implementation manner of the embodiment, the step of dividing the trees in the power transmission channel point cloud into groups through connectivity comprises the following steps:

[0091] a1: setting a segmentation threshold s, a target point p0, and a tree point set T={p0};

[0092] a2: adding all points with a distance less than s from the target point to the set T, i.e., T={p0, p1, p2,...}, and removing the points added to the set from the point cloud;

[0093] a3: setting the target point as the next point of the current target point;

[0094] a4: repeating steps a2 and a3 until the set T is no longer increased, and the set T is a divided tree region;

[0095] a5: reselecting the target point p0 in the point cloud, and repeating steps a1 to a4 until all points in the point cloud are removed;

[0096] a6: grouping all sets T with more than 100 points into a tree set Tree.

[0097] As a possible implementation manner of the embodiment, the step of mapping each group of trees in the power transmission channel point cloud to the monitoring image comprises the following steps:

[0098] Mapping each group of trees in the power transmission channel point cloud to the monitoring image, recording the pixel points and their corresponding three-dimensional points, and a mapping formula for mapping each group of trees in the power transmission channel point cloud to the monitoring image is:

[0099] temp=(p-T)·R

[0100] m=(M·temp T ) T

[0101] m=m÷m(0,2)(homogeneous transformation)

[0102] wherein p is a three-dimensional point, m is a two-dimensional point, R is a point cloud rotation matrix, T is a point cloud translation vector, and M is a camera intrinsic matrix.

[0103] As a possible implementation manner of the embodiment, the step of identifying the tree regions in the monitoring image and comparing the tree regions with the tree regions mapped by the point cloud to the monitoring image comprises the following steps:

[0104] c1: the set of tree regions mapped by the point cloud to the monitoring image is denoted as A, and the set of tree regions in the monitoring image is denoted as D;

[0105] c2: find the upper boundary of each tree mapping region in set A, and form set topA, find the upper boundary of each tree identification region in set D, and form set topD;

[0106] c3: each upper boundary in topA is recorded as ta a , find the pixel point corresponding to ta in all upper boundaries in topD.

[0107] As a possible implementation manner of the embodiment, the corresponding pixel point refers to having the same x coordinate in the pixel coordinate system, and the y coordinate difference is less than 20, and the y coordinate difference is greater than 20, which is considered as a matching error.

[0108] As a possible implementation manner of the embodiment, the calculating the height of the tree in the image according to the comparison result, and calculating the distance between the tree in the image and the conductor of the power transmission line comprises:

[0109] d1: each upper boundary ta in topA a is recorded as m a (u a ,v a ), the three-dimensional point corresponding to m a (u a ,v a ) is recorded as p a (x a ,y a ,z a ), the pixel point corresponding to m a (u a ,v a ) is recorded as d a (u a ,v d ), if m a (u a ,v a ) has no corresponding pixel point, skip the pixel point;

[0110] d2: calculate the pixel difference: diff a =(v d -v a );

[0111] d3: according to the camera imaging principle , calculate the real height H a corresponding to diff a , wherein f is the focal length, L is the distance from the object plane to the camera plane, h is the pixel height, and H is the real height;

[0112] d4: calculate the three-dimensional point p a (x a ,y a ,za ) corresponding three-dimensional coordinates p d (x a ,y a ,z a +H a );

[0113] d5: calculate the Euclidean distance of all guide points and three-dimensional coordinates p d (x a ,y a ,z a +H a ) to find the nearest guide point and the nearest distance;

[0114] d6: the upper boundary t a corresponding to all guide points form a set Near, and all nearest distances form a set Dis;

[0115] d7: select the minimum value in the set Dis as the shortest distance of the tree area to the guide wire.

[0116] As Figure 2 shown, the monitoring device for tree barrier hidden danger of power transmission line provided by the embodiment of the application carries out tree barrier hidden danger monitoring based on power transmission channel monitoring image and power transmission channel point cloud, and the device comprises:

[0117] a tree grouping module, configured to divide trees in the power transmission channel point cloud into multiple groups through connectivity;

[0118] a point cloud mapping module, configured to map each group of trees in the power transmission channel point cloud to the monitoring image;

[0119] a tree area comparison module, configured to identify a tree area in the monitoring image and compare the tree area with a tree area mapped to the monitoring image by the point cloud;

[0120] a distance calculation module, configured to calculate the height of the tree in the image according to the comparison result, and calculate the distance between the tree in the image and the guide wire of the power transmission line.

[0121] The process of carrying out tree barrier hidden danger monitoring based on the power transmission channel monitoring image and the power transmission channel point cloud by using the monitoring device for tree barrier hidden danger of power transmission line provided by the embodiment of the application is as follows.

[0122] First, an image and a point cloud file of a certain section of power transmission line channel are acquired, camera intrinsic parameters, a rotation matrix, and a translation vector are acquired, the distance between tree barrier hidden danger and the guide wire in the power transmission channel is calculated, and through field measurement, the error of the tree barrier hidden danger distance measurement result is about 1 m. The device image and the point cloud image are as shown in Figure 3 and Figure 4 , wherein the point cloud of the tree is as shown in Figure 5 .

[0123] a. Divide the trees in the point cloud into multiple groups by connectivity to form a tree set Tree;

[0124] a1: Set the tree segmentation threshold s to 1, the target point p0 to (-25.33, 122.19, 9.97), and the tree point set T = {[-25.33, 122.19, 9.97]};

[0125] a2: Add all points with a distance less than s to the target point to the set T;

[0126] T = {

[0127] [-25.33, 122.19, 9.97]

[0128] [-25.38, 122.32, 9.92]

[0129] [-25.11, 122.54, 10.37]

[0130] [-25.07, 122.55, 9.96]

[0131] [-24.94, 122.55, 9.94]

[0132] [-24.77, 122.54, 10.11]

[0133] [-25.07, 122.58, 8.63]

[0134] [-25.01, 122.43, 9.82]

[0135] [-24.92, 122.75, 10.18]

[0136] [-24.76, 122.75, 10.28]

[0137] ...

[0138] }

[0139] a3: Set the target point to the next point of the current target point, i.e. (-25.38, 122.32, 9.92);

[0140] a4: Repeat steps a2 and a3 until the set T no longer increases, and the set T is a tree region; T = {

[0141] [-25.33, 122.19, 9.97]

[0142] [-25.38, 122.32, 9.92]

[0143] [-25.11, 122.54, 10.37]

[0144] [-25.07, 122.55, 9.96]

[0145] [-24.94, 122.55, 9.94]

[0146] [-24.77, 122.54, 10.11]

[0147] [-25.07, 122.58, 8.63]

[0148] [-25.01, 122.43, 9.82]

[0149] [-24.92, 122.75, 10.18]

[0150] [-24.76, 122.75, 10.28]

[0151] [-25.36, 122.3, 10.75]

[0152] [-25.11, 122.3, 10.56]

[0153] [-25.54, 122.17, 10.88]

[0154] [-25.44, 122.18, 10.78]

[0155] [-25.31, 122.18, 10.72]

[0156] [-25.47, 122.06, 10.85]

[0157] [-25.42, 123.15, 14.28]

[0158] [-25.29, 123.15, 14.3]

[0159] [-25.46, 123.05, 14.34]

[0160] [-25.31, 123.04, 14.32]

[0161] …

[0162] }

[0163] a5: repeating steps a1 to a4 until all points in the point cloud are removed;

[0164] a6: filter out the set of less than 100 points, and group all sets of more than 100 points T into a tree set Tree; finally, 15 tree regions meeting the conditions are obtained, as shown in Figure 6

[0165] b. Map each set of trees in the point cloud to the monitoring image;

[0166] Input

[0167] M = [[2.23722737e+03 0.00000000e+00 1.31100000e+03]

[0168] [0.00000000e+00 2.24053809e+03 9.84000000e+02]

[0169] [0.00000000e+00 0.00000000e+00 1.00000000e+00]]

[0170] R = [[0.99564296 -0.01460996 -0.09209719] [0.09235461 0.01804687 0.99556243]

[0172] [-0.01288306 -0.99973053 0.01931754]]

[0173] T = [-8.60000134 0.15.20000362]

[0174] The mapping coordinates are calculated as:

[0175] img = {

[0176] [923, 1111]

[0177] [925, 1115]

[0178] [926, 1117]

[0179] [923, 1151]

[0180] [926, 1150]

[0181] [932, 1138]

[0182] [934, 1139]

[0183] [923, 1150]

[0184] [925, 1151]​

[0185] [929,1143]

[0186] [932,1143]

[0187] [923,1112]

[0188] [925,1115]

[0189] [926,1119]

[0190] [924,1151]

[0191] [927,1146]

[0192] [932,1139]

[0193] [934,1139]

[0194] [935,1146]

[0195] [937,1147]

[0196] …

[0197] }

[0198] Mapping effect as shown in Figure 7 .

[0199] c, identify the tree area in the image and compare with the tree area of point cloud mapping;

[0200] Tree identification effect as shown in Figure 8 .

[0201] c1: the area set of tree mapping is recorded as A, and the area set of tree identification is recorded as D; A={ [

[0203] [923,1111]

[0204] [925,1115]

[0205] [926,1117]

[0206] [923,1151]

[0207] [926,1150]

[0208] [932,1138]

[0209] [934,1139]

[0210] [923,1150]

[0211] [925,1151]

[0212] [929,1143]

[0213] [932,1143]

[0214] [923,1112]

[0215] [925,1115]

[0216] [926,1119]

[0217] [924,1151]

[0218] [927,1146]

[0219] [932,1139]

[0220] [934,1139]

[0221] [935,1146]

[0222] [937,1147]

[0223] ……

[0224] ],

[0225] ……

[0226] }

[0227] D＝{ [

[0229] [2440,979]

[0230] [2441,979]

[0231] [2442,979]

[0232] [2443,979]

[0233] [2444,979]

[0234] [2445,979]

[0235] [2446,979]

[0236] [2447,979]

[0237] [2448,979]

[0238] [2435,980]

[0239] [2436,980]

[0240] [2437,980]

[0241] [2438,980]

[0242] [2439,980]

[0243] [2440,980]

[0244] [2441,980]

[0245] [2442,980]

[0246] [2443,980]

[0247] [2444,980]

[0248] [2445,980]

[0249] [476,1071]

[0250] [477,1071]

[0251] [478,1071]

[0252] [479,1071]

[0253] [480,1071]

[0254] [481,1071]

[0255] [482,1071]

[0256] [483,1071]

[0257] [484,1071]

[0258] [485,1071]

[0259] [486,1071]

[0260] [487,1071]

[0261] [488,1071]

[0262] [489,1071]

[0263] [490,1071]

[0264] [467,1072]

[0265] [468,1072]

[0266] [469,1072]

[0267] [470,1072]

[0268] [471,1072]

[0269] …

[0270] ],

[0271] …

[0272] }

[0273] c2: find the upper bound of each tree mapping region in set A, form set topA, find the upper bound of each tree identification region in set D, form set topD; topA={ [

[0275] [922.0, 1114.0] [923.0, 1111.0] [924.0, 1099.0] [925.0, 1113.0] [926.0, 1098.0] [927.0, 1101.0] [928.0, 1090.0] [929.0, 1095.0] [930.0, 1089.0] [931.0, 1090.0] [932.0, 1087.0] [933.0, 1088.0] [934.0, 1083.0] [935.0, 1088.0] [936.0, 1092.0] [937.0, 1090.0] [938.0, 1094.0] [939.0, 1083.0] [940.0, 1091.0] [941.0, 1094.0] …

[0276] ],

[0277] …

[0278] }

[0279] topD={ [

[0281] [1312, 1131]

[0282] [1313, 1126]

[0283] [1314,1118]

[0284] [1315,1113]

[0285] [1316,1111]

[0286] [1317,1110]

[0287] [1318,1109]

[0288] [1319,1108]

[0289] [1320,1107]

[0290] [1321,1106]

[0291] [1322,1105]

[0292] [1323,1104]

[0293] [1324,1103]

[0294] [1325,1102]

[0295] [1326,1102]

[0296] [1327,1101]

[0297] [1328,1100]

[0298] [1329,1100]

[0299] [1330,1099]

[0300] [1331,1099]

[0301] …

[0302] ],

[0303] …

[0304] }

[0305] c3: topA each of the upper boundary in the record t a , in all the upper boundary in the topD find the corresponding ta each pixel point of the pixel point.

[0306] Calculate the corresponding pixel points, for example, a certain area:

[0307] [1887.0,1040.0--1887.0,1021.0]

[0308] [1888.0,1038.0--1888.0,1021.0]

[0309] [1889.0,1037.0--1889.0,1021.0]

[0310] [1890.0,1025.0--1890.0,1021.0]

[0311] [1891.0,1020.0--1891.0,1020.0]

[0312] [1892.0,1019.0--1892.0,1020.0]

[0313] [1893.0,1027.0--1893.0,1020.0]

[0314] [1894.0,1025.0--1894.0,1019.0]

[0315] [1895.0,1023.0--1895.0,1019.0]

[0316] [1896.0,1021.0--1896.0,1019.0]

[0317] [1897.0,1021.0--1897.0,1018.0]

[0318] [1898.0,1024.0--1898.0,1018.0]

[0319] [1899.0,1020.0--1899.0,1018.0]

[0320] [1900.0,1025.0--1900.0,1017.0]

[0321] [1901.0,1022.0--1901.0,1017.0]

[0322] [1902.0,1025.0--1902.0,1017.0]

[0323] [1903.0,1023.0--1903.0,1016.0]

[0324] [1904.0,1021.0--1904.0,1016.0]

[0325] [1905.0,1023.0--1905.0,1016.0]

[0326] [1906.0,1019.0--1906.0,1015.0]

[0327] d. Estimate the height of the trees in the image and calculate the distance between the trees and the guide wires. d1: Calculate the pixel difference (diff). a =(v d -v a );

[0328] diff a ={ [ 19.0 17.0 16.0 4.0 0.0

[0335] -1.0 7.0 6.0 4.0 2.0 3.0 6.0 2.0 8.0 5.0 8.0 7.0 5.0 7.0 4.0

[0350] ...

[0351] ],

[0352] ...

[0353] }

[0354] d2: Calculate diff a Corresponding actual height H a H a ={ [ 1.3341273588935565 1.1914322374349566 1.1228954561405977 0.2807293782645598 0.0

[0361] -0.07026752087335532 0.4922052273603588 0.42116230979914093 0.28078449024776175 0.14099327325780173 0.21150660605547192 0.4217128725003192 0.140506851335485 0.5617674111632006 0.35129759896940277 0.561422246826717 0.4909542744842129 0.35091494695242587 0.4913209231743029 0.28235064992427994

[0376] ...

[0377] ],

[0378] ...

[0379] }

[0380] d3: Calculate tree three-dimensional coordinates p d ;

[0381] p d ={ [

[0383] [19.09,172.13,15.044127358893558]

[0384] [19.09,171.8,15.021432237434956]

[0385] [19.18,172.03,15.062895456140597]

[0386] [19.3,172.02,15.060729378264561]

[0387] [19.37,172.21,15.22]

[0388] [19.490000000000002,172.21,15.169732479126646] [19.56, 172.32, 15.122205227360359]

[0389] [19.55, 172.02, 15.231162309799142]

[0390] [19.69, 172.01, 15.200784490247761]

[0391] [19.83, 172.74, 15.240993273257802] [19.95, 172.74, 15.281506606055473]

[0392] [19.88, 172.21, 15.30171287250032]

[0393] [19.97, 172.12, 15.310506851335484]

[0394] [20.05, 172.03, 15.371767411163201]

[0395] [20.1, 172.12, 15.401297598969403]

[0396] [20.16, 171.91, 15.391422246826718]

[0397] [20.22, 171.8, 15.410954274484213]

[0398] [20.27, 171.91, 15.420914946952426]

[0399] [20.39, 171.91, 15.401320923174303]

[0400] [20.65, 172.87, 15.522350649924281] …

[0401] ],

[0402] …

[0403] }

[0404] d5: Calculate the Euclidean distance of all guide points to all three-dimensional coordinates, find the nearest guide point and the nearest distance;

[0405] The shortest distance of each tree area is calculated, and the final result is as follows:

[0406] [{'distance': 61.6, 'line': [[1911, 1013], [1768, 215]]}, {'distance': 60.46, 'line': [[1943, 1005], [1860, -83]]}, {'distance': 60.75, 'line': [[1924, 1008], [1804, 99]]}, {'distance': 67.18, 'line': [[1334, 1096], [1468, 183]]}, {'distance': 69.48, 'line': [[2256, 1017], [1780, 178]]}, {'distance': 64.02, 'line': [[2121, 1021], [1811, 76]]}, {'distance': 62.91, 'line': [[1693, 1041], [1789, 119]]}]

[0407] The final distance measurement result is shown in Figure 9

[0408] The tree line distance height measured by the embodiment is estimated by the actual measurement of the tree line distance, and the error is about 1 m.

[0409] The present application adopts tree division, image and point cloud mapping amount area comparison and other technologies, realizes automatic division of trees in the power transmission line channel, calculates the shortest distance of each tree area to the conductor, and according to the tree recognition result, infers the tree growth in the image, and can divide the risk level according to the tree distance measurement result, discovers and handles the tree barrier hidden danger that may cause harm to the power transmission channel line in time, strengthens the management and control strength of the power transmission channel, and the operation and inspection personnel can more accurately control the tree height, reduces the accidents caused by tree barrier hidden danger.

[0410] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit it, although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that: the specific embodiments of the present application can be modified or replaced, without departing from the spirit and scope of the present application, any modification or equivalent replacement, which should be covered in the protection scope of the claims of the present application.​

Claims

1. A method for monitoring the hazards of tree obstructions along power transmission lines, characterized in that, The method for monitoring tree obstruction hazards based on power transmission corridor surveillance images and power transmission corridor point clouds includes the following steps: Trees in the power transmission channel point cloud are divided into multiple groups based on connectivity; Map each group of trees in the point cloud of the power transmission channel onto the captured image; Identify tree regions in the surveillance images and compare them with the tree regions mapped from the point cloud onto the surveillance images; Calculate the height of the trees in the image based on the comparison results, and calculate the distance between the trees and the power transmission line conductors in the image; The process of identifying tree regions in the surveillance image and comparing them with tree regions mapped from the point cloud onto the surveillance image includes: c1: The set of tree regions mapped from the point cloud onto the surveillance image is denoted as c1. A The set of tree regions identified in the surveillance image is denoted as... D ; c2: Find the set A The upper boundary of each tree mapping region forms a set. topA Finding sets D The upper boundary of each tree identification region forms a set. topD ; c3: topA Each upper boundary in the middle is denoted as t a ,exist topD Find the corresponding one among all upper boundaries. t a Each pixel is a pixel; The calculation of the tree height in the image based on the comparison results, and the calculation of the distance between the tree in the image and the power transmission line conductor, include: d1: Upper boundary t a Each pixel in the data is denoted as . m a , m a The corresponding three-dimensional point is denoted as p a , m a The corresponding pixel is denoted as d a ,if m a If there is no corresponding pixel, skip that pixel. d2: Calculate pixel difference: ; d3: Based on the camera imaging principle ,calculate diff a Corresponding actual height H a ,in f It's the focal length. L It is the distance between the object's plane and the camera plane. h It is the pixel height. H It is the actual height; d4: Calculate the 3D points of the tree p a Corresponding three-dimensional coordinates p d ; d5: Calculate the coordinates of all traverse points in three dimensions. p d Using the Euclidean distance, find the nearest traverse point and its closest distance; d6: Upper boundary t a The set of all the nearest points of the corresponding conductors Near The set of all nearest neighbors Dis ; d7: Select set Dis The minimum value in the range is taken as the shortest distance from the guide wire to the area of ​​trees.

2. The monitoring method for the hidden dangers of tree obstruction on transmission lines according to claim 1, characterized in that, The method of dividing the trees in the power transmission channel point cloud into multiple groups based on connectivity includes: a1: Set the segmentation threshold s Target point p 0, Tree point set ; a2: All points whose distance from the target point is less than... s Add the point to the set T tree ,Right now And remove the points added to the set from the point cloud; a3: Set T tree The next point after the current target point is set as the target point; a4: Repeat steps a2 and a3 until the set is complete. T tree No more additions, set T tree This refers to a designated area of ​​trees; a5: Reselect the target point in the point cloud p 0. Repeat steps a1 to a4 until all points in the point cloud have been removed; a6: The set of all points greater than 100 T tree Composition of tree collection Tree .

3. The monitoring method for the hidden dangers of tree obstruction on transmission lines according to claim 1, characterized in that, The process of mapping each group of trees in the power transmission channel point cloud onto the monitored image includes: Each group of trees in the power transmission channel point cloud is mapped onto the monitoring image, and the pixels and their corresponding 3D points are recorded. The mapping formula for mapping each group of trees in the power transmission channel point cloud onto the monitoring image is as follows: in, p It is a three-dimensional point. m It is a two-dimensional point. R It is a point cloud rotation matrix. T It is a point cloud translation vector. M It is the camera intrinsic parameter matrix.

4. A monitoring device for the hidden dangers of tree obstruction on power transmission lines, characterized in that, The device for monitoring tree obstruction hazards based on power transmission corridor surveillance images and power transmission corridor point clouds includes: The tree grouping module is used to divide trees in the power transmission channel point cloud into multiple groups based on connectivity. The point cloud mapping module is used to map each group of trees in the point cloud of the power transmission channel onto the monitoring image; The tree region comparison module is used to identify tree regions in the surveillance image and compare them with the tree regions mapped from the point cloud onto the surveillance image. The distance calculation module is used to calculate the height of trees in the image based on the comparison results, and to calculate the distance between the trees in the image and the power transmission line conductors; The specific process by which the tree region comparison module identifies tree regions in the surveillance image and compares them with the tree regions mapped from the point cloud onto the surveillance image is as follows: c1: The set of tree regions mapped from the point cloud onto the surveillance image is denoted as c1. A The set of tree regions identified in the surveillance image is denoted as... D ; c2: Find the set A The upper boundary of each tree mapping region forms a set. topA Finding sets D The upper boundary of each tree identification region forms a set. topD ; c3: topA Each upper boundary in the middle is denoted as t a ,exist topD Find the corresponding one among all upper boundaries. t a Each pixel is a pixel; The specific process by which the distance calculation module calculates the height of the trees in the image based on the comparison results, and calculates the distance between the trees in the image and the power transmission line conductors, is as follows: d1: Upper boundary t a Each pixel in the data is denoted as . m a , m a The corresponding three-dimensional point is denoted as p a , m a The corresponding pixel is denoted as d a ,if m a If there is no corresponding pixel, skip that pixel. d2: Calculate pixel difference: ; d3: Based on the camera imaging principle ,calculate diff a Corresponding actual height H a ,in f It's the focal length. L It is the distance between the object's plane and the camera plane. h It is the pixel height. H It is the actual height; d4: Calculate the 3D points of the tree p a Corresponding three-dimensional coordinates p d ; d5: Calculate the coordinates of all traverse points in three dimensions. p d Using the Euclidean distance, find the nearest traverse point and its closest distance; d6: Upper boundary t a The set of all the nearest points of the corresponding conductors Near The set of all nearest neighbors Dis ; d7: Select set Dis The minimum value in the range is taken as the shortest distance from the guide wire to the area of ​​trees.

5. The monitoring device for the hidden dangers of tree obstruction on power transmission lines according to claim 4, characterized in that, The specific process by which the tree grouping module divides the trees in the power transmission channel point cloud into multiple groups based on connectivity is as follows: a1: Set the segmentation threshold s Target point p 0, Tree point set ; a2: All points whose distance from the target point is less than... s Add the point to the set T tree ,Right now And remove the points added to the set from the point cloud; a3: Set the next point in set T after the current target point as the target point; a4: Repeat steps a2 and a3 until the set is complete. T tree No more additions, set T tree This refers to a designated area of ​​trees; a5: Reselect the target point in the point cloud p 0. Repeat steps a1 to a4 until all points in the point cloud have been removed; a6: The set of all points greater than 100 T tree Composition of tree collection Tree .

6. The monitoring device for the hidden dangers of tree obstruction on transmission lines according to claim 4, characterized in that, The mapping formula used by the point cloud mapping module to map each group of trees in the power transmission channel point cloud to the monitored image is as follows: in, p It is a three-dimensional point. m It is a two-dimensional point. R It is a point cloud rotation matrix. T It is a point cloud translation vector. M It is the camera intrinsic parameter matrix.

Images