An overhead power transmission line edge positioning method, system, electronic device and medium

Abstract

The application discloses an overhead transmission line edge positioning method and system, electronic equipment and medium, and relates to the technical fields of overhead transmission line detection and image processing. The method comprises the following steps: performing Gaussian blurring on an overhead transmission line image to obtain a blurred image; performing processing on the blurred image by using an LSD line segment detection algorithm to obtain a line segment set; filtering line segments in the line segment set according to a set slope interval and slopes of the line segments in the line segment set to obtain a filtered line segment set; performing linear aggregation on the line segments in the filtered line segment set to obtain a plurality of long line segments; screening all long line segments according to slopes of the long line segments to obtain overhead transmission line edge line segments; and determining an overhead transmission line edge according to the overhead transmission line edge line segments. The application can improve robustness in a complex scene and improve the accuracy of detection results.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of overhead transmission line detection and image processing, and particularly relates to an overhead transmission line edge positioning method and system, an electronic device and a medium. BACKGROUND

[0002] Overhead transmission lines mainly refer to overhead open lines fixed on ground towers by insulators, and are important bearing media for power transmission, and have advantages such as low construction cost and high maintainability. However, compared with buried ground transmission lines, the surface of the overhead transmission line is directly exposed to the air, and is easily affected by environmental and weather factors. Especially in winter with low temperature and high humidity, ice layers are easily attached to the surface, causing accidents such as insulator flashover, hardware damage and tower collapse, and thus destroying the stability of power supply. Therefore, it is very important to periodically inspect the overhead transmission line to maintain the safety of the power grid.

[0003] In the traditional power safety inspection operation, power grid workers manually check the conditions of the overhead transmission lines in each area using a portable telescope, and timely assess and feedback the existing risks. However, due to the high altitude of the overhead transmission line erection, the wide coverage area and the fact that it is usually located in the suburbs, manual inspection is difficult to cover many safety dead angles, and the workload is huge to implement. The recent trend is to periodically collect pictures of the transmission line based on a visual terminal device mounted on a tower or a drone, and directly deploy a linear detection algorithm to locate the edge of the overhead transmission line, such as the LSD line segment detector. However, these linear detection algorithms, although perform well in idealized simulation experiments, have many problems in real engineering scenarios. On the one hand, the images collected by the visual terminal device are mostly in the background of mountains or forests, and they contain complex texture features, which causes the linear detection algorithm to face serious performance degradation problems, and has poor robustness in complex scenes. On the other hand, the edge of the transmission line is not always a straight line, and even in winter sometimes some irregular ice layers are attached, so it cannot be detected as a complete long line segment. Therefore, the result detected by the linear detection algorithm is often a set of short line segments, which contains many fragmented edge short line segments and some false line segments caused by false detection, such as Figure 1 as shown, the detection result is inaccurate.

[0004] In summary, the traditional manual inspection method cannot be well applied to the overhead transmission line inspection operation, and the existing automated method, although can preliminarily locate the edge of the overhead transmission line by directly introducing the linear detection algorithm, has poor robustness in complex scenes, and thus the final detection result is not accurate. SUMMARY

[0005] The application aims to provide an overhead power line edge positioning method, system, electronic device and medium, which can improve robustness in complex scenes and improve the accuracy of detection results.

[0006] To achieve the above-mentioned purpose, the application provides the following solutions.

[0007] An overhead power line edge positioning method comprises the following steps:

[0008] An overhead power line image is acquired;

[0009] The overhead power line image is subjected to Gaussian blur to obtain a blurred image;

[0010] An LSD (Line Segment Detector) algorithm is used to process the blurred image to obtain a line segment set;

[0011] The line segments in the line segment set are filtered according to a set slope interval and the slopes of the line segments in the line segment set to obtain a filtered line segment set;

[0012] The line segments in the filtered line segment set are subjected to linear aggregation to obtain a plurality of long line segments;

[0013] All long line segments are screened according to the slopes of the long line segments to obtain an overhead power line edge line segment;

[0014] The overhead power line edge is determined according to the overhead power line edge line segment.

[0015] Optionally, the filtering of the line segments in the line segment set according to the set slope interval and the slopes of the line segments in the line segment set to obtain a filtered line segment set specifically comprises:

[0016] The slopes of the line segments in the line segment set are determined;

[0017] The filtered line segment set is obtained according to the slopes of the line segments in the line segment set and the set slope interval; the filtered line segment set comprises all line segments with slopes within the set slope interval.

[0018] Optionally, the linear aggregation of the line segments in the filtered line segment set to obtain a plurality of long line segments specifically comprises:

[0019] The longest line segment in the filtered line segment set is taken as a target line segment;

[0020] A long line segment set is formed by the aggregated line segment set and the target line segment; the aggregated line segment set comprises all line segments in the filtered line segment set that meet an aggregation condition with the target line segment; the aggregation condition comprises that the difference between the slopes of two line segments is less than a set slope difference threshold or the distance between the two line segments is less than a set distance threshold.

[0021] determining whether the total length of all line segments in the long line segment set is greater than a first set threshold value, to obtain a first determination result;

[0022] if the first determination result is yes, connecting all line segments in the long line segment set as a long line segment, and deleting the line segments constituting the long line segment set from the filtered line segment set to update the filtered line segment set, and returning to the step of taking the longest line segment in the filtered line segment set as a target line segment;

[0023] if the first determination result is no, deleting the line segments constituting the long line segment set from the filtered line segment set to update the filtered line segment set, and returning to the step of taking the longest line segment in the filtered line segment set as a target line segment.

[0024] Optionally, the filtering all long line segments according to the slopes of the long line segments to obtain the overhead power line edge line segments specifically comprises:

[0025] calculating an average value of the slopes and a standard deviation of the slopes according to the slopes of all long line segments;

[0026] for any long line segment, calculating a Z-score value of the slope of the long line segment according to the average value of the slopes, the standard deviation of the slopes and the slope of the long line segment;

[0027] deleting the long line segments with the Z-score values greater than a second set threshold value to obtain the overhead power line edge line segments.

[0028] An overhead power line edge positioning system, comprising:

[0029] an acquisition module configured to acquire an overhead power line image;

[0030] a Gaussian blur module configured to perform Gaussian blur on the overhead power line image to obtain a blurred image;

[0031] an LSD line segment detector configured to process the blurred image by using an LSD (Line Segment Detector) straight line segment detection algorithm to obtain a line segment set;

[0032] a filtered line segment set determination module configured to filter the line segments in the line segment set according to a set slope interval and the slopes of the line segments in the line segment set to obtain a filtered line segment set;

[0033] a linear aggregation module configured to perform linear aggregation on the line segments in the filtered line segment set to obtain a plurality of long line segments;

[0034] The screening module is configured to screen all the long line segments according to the slopes of the long line segments to obtain the overhead transmission line edge line segments.

[0035] The overhead transmission line edge determination module is configured to determine the overhead transmission line edge according to the overhead transmission line edge line segments.

[0036] Optionally, the filtered line segment set determination module specifically comprises:

[0037] The slope calculation unit is configured to determine the slope of each line segment in the line segment set.

[0038] The filtered line segment set determination unit is configured to obtain a filtered line segment set according to the slopes of the line segments in the line segment set and the set slope interval; the filtered line segment set comprises all the line segments with the slopes located in the set slope interval.

[0039] Optionally, the linear aggregation module specifically comprises:

[0040] The target line segment determination unit is configured to determine the longest line segment in the filtered line segment set as a target line segment.

[0041] The long line segment set determination unit is configured to combine the aggregated line segment set and the target line segment to form a long line segment set; the aggregated line segment set comprises all the line segments in the filtered line segment set that meet an aggregation condition with the target line segment; the aggregation condition comprises that the difference between the slopes of two line segments is less than a set slope difference threshold or the distance between the two line segments is less than a set distance threshold.

[0042] The judgment unit is configured to determine whether the total length of all the line segments in the long line segment set is greater than a first set threshold to obtain a first determination result.

[0043] The long line segment determination and filtered line segment set updating unit is configured to, if the first determination result is yes, connect all the line segments in the long line segment set as a long line segment, delete the line segments forming the long line segment set from the filtered line segment set to update the filtered line segment set, and return to the step of determining the longest line segment in the filtered line segment set as a target line segment.

[0044] The filtered line segment set updating unit is configured to, if the first determination result is no, delete the line segments forming the long line segment set from the filtered line segment set to update the filtered line segment set, and return to the step of determining the longest line segment in the filtered line segment set as a target line segment.

[0045] Optionally, the screening module specifically comprises:

[0046] a mean value and a standard deviation of the slopes are determined from the slopes of all long line segments;

[0047] a Z-score value is calculated for each long line segment from the mean value of the slopes, the standard deviation of the slopes and the slope of the long line segment;

[0048] an overhead power line edge line segment is calculated by removing long line segments with a Z-score value greater than a second set threshold value.

[0049] An electronic device comprising:

[0050] a memory for storing a computer program and a processor for running the computer program to cause the electronic device to perform the overhead power line edge positioning method according to the above.

[0051] A computer readable storage medium storing a computer program which, when executed by a processor, implements the overhead power line edge positioning method according to the above.

[0052] According to the embodiments of the present application, the following technical effects are achieved: the Gaussian blur of the overhead power line image reduces the performance degradation caused by image noise, and improves the robustness in complex scenes; the slopes of the line segments in the line segment set obtained by processing the blurred image according to the set slope interval and the LSD line segment detection algorithm are used to filter the line segments in the line segment set, and a filtered line segment set is obtained; the line segments in the filtered line segment set are linearly aggregated to obtain a plurality of long line segments; all long line segments are screened according to the slopes of the long line segments, and an overhead power line edge line segment is obtained, thereby improving the accuracy of the detection result. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0054] Figure 1 a schematic diagram of the problems existing in the use of the existing linear detection algorithm;

[0055] Figure 2 a flowchart of the overhead power line edge positioning method provided by the embodiments of the present application;

[0056] Figure 3A schematic diagram of an overhead power transmission line edge positioning method provided by an embodiment of the present application is shown in the figure.

[0057] Figure 4 A schematic diagram of an overhead power transmission line image taken by a UAV is shown in the figure.

[0058] Figure 5 A schematic diagram of a line segment aggregation process is shown in the figure.

[0059] Figure 6 A detection flowchart of an LSD line segment detection algorithm is shown in the figure. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0061] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0062] As shown in Figure 2 and Figure 3 , an overhead power transmission line edge positioning method is provided by an embodiment of the present application, characterized in that comprising:

[0063] Step 101: obtaining an overhead power transmission line image. As shown in Figure 4 , the overhead power transmission line image is an overhead power transmission line picture taken by a UAV camera.

[0064] Step 102: performing Gaussian blur on the overhead power transmission line image to obtain a blurred image. Because the foreground texture of the transmission line taken in the working scene is rich and the background is mixed with various image noises, it is necessary to perform Gaussian blur preprocessing on the image.

[0065] Step 103: processing the blurred image by using an LSD line segment detection algorithm to obtain a line segment set. The line segment set is a set composed of line segments with linear characteristics.

[0066] Step 104: filtering the line segments in the line segment set according to a set slope interval and the slopes of the line segments in the line segment set to obtain a filtered line segment set.

[0067] Step 105: performing linear aggregation on the line segments in the filtered line segment set to obtain a plurality of long line segments. As shown in Figure 5 .

[0068] Step 106: screening all long line segments according to the slope of each long line segment to obtain overhead power line edge line segments.

[0069] Step 107: determining the overhead power line edge according to the overhead power line edge line segments.

[0070] In practical application, Gaussian blur is performed on the overhead power line image to obtain a blurred image, and the specific process includes:

[0071] A two-dimensional convolution kernel with a size of N*N (a scale of 3*3 and a standard deviation of 1) is set to traverse the pixels in the form of a sliding window on the original image, the texture information in the entire picture is processed, the image noise in the foreground and the background is faded, and the performance degradation problem caused by the image noise in the foreground and the background in complex situations is reduced. The parameter value of each position on the convolution kernel is the weighted average value of the Gaussian function of the difference between the horizontal and vertical coordinates of the position and the horizontal and vertical coordinates of the center position of the convolution kernel. Specifically, as shown in the following formula (1), wherein x and y represent the horizontal and vertical coordinates of a position on the convolution kernel; Δx and Δy represent the difference between the horizontal and vertical coordinates of the position and the horizontal and vertical coordinates of the center of the convolution kernel; Gaussian(x, y) represents the Gaussian value of the position; and σ represents the standard deviation, which is a hyperparameter.

[0072] In practical application, before the blurred image is processed by using the LSD line segment detection algorithm to obtain a line segment set, the specific process includes: performing gray-scale preprocessing on the blurred image according to formula (2) Gray=(R+G+B) / 3, wherein R / G / B represents the color value of each pixel in the red, green and blue channels of the RGB three channels, and Gray represents the single-channel gray value obtained by averaging the three channels.

[0073] In practical application, the blurred image is processed by using the LSD line segment detection algorithm, as shown in the following formula (3), and the specific process includes: Figure 6

[0074] All pixels in the blurred image are traversed, and it is assumed that the coordinates of every four adjacent pixels are as shown in the following formula (4), wherein i(x, y) represents the image gray value of pixel (x, y), and then the gradient value in the horizontal and vertical directions can be calculated according to the gray values of adjacent pixels according to the following formula (5). wherein g x (x, y) and g y (x, y) represent the gradient values of the pixel with coordinates (x, y) in the x and y axis directions, respectively. Further, the image gradient value of the point can be calculated according to the following formula (6). ​​​The horizontal line angle value and the gradient amplitude value are calculated, where θ(x, y) represents the horizontal line angle value (which can also be directly abbreviated as θ), and the horizontal line points to the extension direction of the edge in the image, which is perpendicular to the direction of color change from black to white between different pixels; and G(x, y) represents the gradient amplitude value, which is the summary of the gradient values in the horizontal and vertical directions of the pixel, and can be used to describe the speed of color change. For example, when G(x, y) is larger, the contrast between two pixels will be higher.

[0075] Next, region growing is performed, starting from the pixel with the highest gradient amplitude value in the image, and traversing the adjacent pixels, adding the pixels with a difference between the horizontal line angle and the region line angle θ region of the region set, and updating the region line angle of the corresponding region set according to j where j represents a pixel number in the region set, and θ region represents the horizontal line angle value of the jth pixel. The initial value of θ region is equal to the horizontal line angle of the starting pixel point, and then it is updated to a new value according to formula (6) after the region set adds adjacent pixels each time. When the region set has no adjacent pixels that can be traversed, one iteration is completed, and the pixel points with a lower gradient amplitude value are traversed according to the same rule.

[0076] Each region after iteration is composed of its initial pixel and adjacent pixels that meet the conditions, and a minimum circumscribed rectangle can be generated for each region. Finally, there are many region sets, and the initial pixels of the region sets are retrieved in turn according to the gradient amplitude value. A straight line is generated with the center point of the minimum circumscribed rectangle as the center point, and the direction of the straight line is the same as the direction of the region line. The straight line has two intersection points with the rectangle, and the line segment connected by the two intersection points is one of the elements in the line segment set generated by the LSD line segment detector (hereinafter, (x1, y1) and (x2, y2) represent the coordinates of the two end points of a line segment, and y=kx+b represents the linear equation of a line segment).

[0077] In practical applications, the region growing is as follows:

[0078] 1 2 3 4 5 6 7 8 9

[0079] (1) It is assumed that the 5th pixel is the pixel with the highest gradient amplitude value, and its adjacent pixels are 12346789.

[0080] (2) The 5th pixel is added to the region set, and the initial value of the region line angle is set to θ region = θ 5号If the difference between the horizontal line angle of the 123th pixel and the region line angle is lower than the threshold τ, the 123th pixel is added to the region set and the region line angle is updated according to the formula (6). At this time, the region set is {1, 2, 3, 5}, wherein the 5th pixel in the set is in the traversed state and the 123th pixel is in the untraversed state.

[0081] (3) The first untraversed element 1st pixel in the region set is taken out and marked as the traversed state, and the adjacent pixels of the 1st pixel are traversed. The adjacent pixels meeting the condition are added to the region set and the region line angle is updated.

[0082] The above operation is repeated until there is no pixel in the untraversed state in the region set. At this time, all the elements in the region set are the final result.

[0083] In actual application, the line segment set obtained in step 103 includes not only the correct power line edge line segments but also some noise line segments caused by image noise. According to the rule in prior art 1, the tilt angles of the detected correct edge line segments are concentrated between -15° and 15°. Therefore, most of the noise line segments can be screened out by judging the slope of the line segment. The line segments in the line segment set are filtered according to the set slope interval and the slope of each line segment in the line segment set to obtain a filtered line segment set, which specifically includes:

[0084] The slope of each line segment in the line segment set is determined.

[0085] The filtered line segment set is obtained according to the slope of each line segment in the line segment set and the set slope interval to screen out obviously incorrect line segments; the filtered line segment set includes all the line segments with the slope in the set slope interval, and the set slope interval is determined according to the tilt angle of the power line photographed by the unmanned aerial vehicle in the lateral flight attitude.

[0086] In actual application, the slope of each line segment in the line segment set is determined as follows: The slope k is calculated, wherein (x1, y1) and (x2, y2) represent the coordinates of the two end points of the line segment. Since there is a tangent conversion relationship between the slope and the tilt angle of the line segment, only the line segment with the slope k in the range of [arctan(-15°), arctan15°] is reserved in this step, and other filtered line segments are obviously incorrect misdetected line segments.

[0087] In practical application, although the line segment obtained in step 103 is easy to be fragmented into multiple short line segments due to the influence of noise pixels, the noise pixels cause fragmentation, and further cause robustness problems, as shown in the rule of the second prior art, the fragmented line segments belonging to the same edge are different, but overall will be distributed in the direction of the same straight line, and the straight line equation is also approximate. On the other hand, according to statistics, it is found that the longer the length of the line segment output by the LSD, the higher the detection accuracy, so starting from the long line segment can quickly determine the line segment group belonging to the same edge, so the linear aggregation is performed on the filtered line segment set, the fragmented short line segments corresponding to the same edge are repaired into complete long line segments, and other discrete false detection line segments are screened out, so the linear aggregation is performed on the line segments in the filtered line segment set to obtain multiple long line segments, and the linear aggregation specifically includes:

[0088] The longest line segment in the filtered line segment set is taken as a target line segment.

[0089] The aggregated line segment set and the target line segment form a long line segment set; the aggregated line segment set includes all line segments in the filtered line segment set that meet an aggregation condition with the target line segment; the aggregation condition includes that the difference between the slopes of two line segments is less than a set slope difference threshold thed_k or the distance between the two line segments is less than a set distance threshold thed_d, and the meeting of the aggregation condition proves that the line segments are approximately distributed in the direction of the same straight line.

[0090] A first determination result is obtained by determining whether the total length of all line segments in the long line segment set is greater than a first set threshold thed_dis.

[0091] If the first determination result is yes, all line segments in the long line segment set are connected into a long line segment, the line segments constituting the long line segment set are deleted from the filtered line segment set to update the filtered line segment set, and the step of taking the longest line segment in the filtered line segment set as a target line segment is returned.

[0092] If the first determination result is no, the line segments constituting the long line segment set are deleted from the filtered line segment set to update the filtered line segment set, and the step of taking the longest line segment in the filtered line segment set as a target line segment is returned.

[0093] In practical application, the determination process of the aggregated line segment set is as follows:

[0094] The elements of the filtered line segment set obtained through step 104 are reordered into a new set Q according to their line segment lengths, and then the longest line segment l is taken from set Q without replacement each time, which is compared with other elements in the set to determine the aggregability between each other. If aggregable, the compared line segment is also taken out of set Q for use. After one traversal, if the total length of all taken line segments exceeds theed_dis=100 pixels (hyperparameter), the average of k and b is taken to generate a complete long line segment output, and all taken line segments are deleted from the filtered line segment set for the next traversal. If it is not exceeded, all taken line segments are deleted from the filtered line segment set for the next traversal.

[0095] There are two criteria for determining whether two line segments can be aggregated:

[0096] (1) Assuming there are two line segments A and B, their linear equations are respectively The slope difference Δk = |k1-k2| is obtained by subtracting their slopes k1, k2, and if Δk is lower than the threshold value thed_k=0.02, it means that they are approximately oriented, and the two line segments can be aggregated; otherwise, it is considered that the two line segments cannot be aggregated.

[0097] (2) The slopes of line segments A and B are set to the average value By default, they are parallel, and the distance d between the two parallel lines can be calculated: If d is less than the threshold value thed_d=4, it is considered that the two line segments can be aggregated, otherwise it is considered that the two line segments cannot be aggregated.

[0098] In practical applications, there are also noise line segments accidentally aggregated together in the long line segments aggregated in step 105. Since the orientations of the power transmission lines in each picture are mostly the same, these noise line segments generally have an abnormal slope when aggregated into long line segments, and therefore, the third priori rule can be used to solve this problem by clustering the slopes of all long line segments to find those with abnormal slopes and filter them out. Specifically, the Z-score of each long line segment is calculated based on the mean and standard deviation of the slopes of all long line segments, and if it is higher than thed_out, it is directly filtered out as an error detection result. Therefore, the long line segments are screened according to the slopes of the long line segments to obtain the edge line segments of the overhead power transmission line, specifically including:

[0099] The mean and standard deviation of the slopes of all long line segments are calculated.

[0100] For any long line segment, the Z-score value of the slope of the long line segment is calculated based on the mean of the slopes, the standard deviation of the slopes, and the slope of the long line segment.

[0101] The long line segments with the Z-score value greater than the second set threshold thed_out are deleted to obtain the overhead power transmission line edge line segments, that is, the line segments with the slope not in the group are deleted.

[0102] In actual application, the average value of the slope and the standard deviation of the slope are calculated according to the slopes of all the long line segments, specifically including:

[0103] Suppose the slopes of all the long line segments are k1, k2,..., kn respectively. n Then the average value k and the standard deviation σ(k) of the slopes can be calculated according to , wherein k i is the slope of the i-th long line segment, and n represents the total number of all the long line segments.

[0104] In actual application, the Z-score value of the slope of the long line segment is calculated according to the average value of the slope, the standard deviation of the slope and the slope of the long line segment, specifically including:

[0105] According to , wherein i represents the i-th long line segment, and Z i is the Z-score of the i-th long line segment. If the Z-score of a line segment is greater than the second set threshold thed_out = 1.8, it is considered as an outlier line segment and needs to be deleted; otherwise, it is reserved as the result output.

[0106] For the above method, an embodiment of the present application provides an overhead power transmission line edge positioning system, comprising:

[0107] An acquisition module is configured to acquire an overhead power transmission line image.

[0108] A Gaussian blur module is configured to perform Gaussian blur on the overhead power transmission line image to obtain a blurred image.

[0109] An LSD line segment detector is configured to perform processing on the blurred image by using an LSD (Line Segment Detector) straight line segment detection algorithm to obtain a line segment set.

[0110] A filtered line segment set determination module is configured to filter the line segments in the line segment set according to a set slope interval and the slopes of the line segments in the line segment set to obtain a filtered line segment set.

[0111] A linear aggregation module is configured to perform linear aggregation on the line segments in the filtered line segment set to obtain a plurality of long line segments.

[0112] A screening module is configured to screen all the long line segments according to the slopes of the long line segments to obtain overhead power transmission line edge line segments.

[0113] An overhead power transmission line edge determination module is configured to determine the overhead power transmission line edge according to the overhead power transmission line edge line segments.

[0114] As an optional implementation, the filtered line segment set determination module specifically includes:

[0115] A slope calculation unit is configured to determine the slope of each line segment in the line segment set.

[0116] A filtered line segment set determination unit is configured to obtain a filtered line segment set according to the slope of each line segment in the line segment set and the set slope interval; the filtered line segment set includes all line segments with a slope within the set slope interval.

[0117] As an optional implementation, the linear aggregation module specifically includes:

[0118] A target line segment determination unit is configured to determine the longest line segment in the filtered line segment set as a target line segment.

[0119] A long line segment set determination unit is configured to form a long line segment set by aggregating the target line segment and a line segment aggregation set; the line segment aggregation set includes all line segments in the filtered line segment set that meet an aggregation condition with the target line segment; the aggregation condition includes that the difference between the slopes of two line segments is less than a set slope difference threshold or the distance between the two line segments is less than a set distance threshold.

[0120] A judgment unit is configured to determine whether the total length of all line segments in the long line segment set is greater than a first set threshold to obtain a first judgment result.

[0121] A long line segment determination and filtered line segment set update unit is configured to, if the first judgment result is yes, connect all line segments in the long line segment set as a long line segment, delete the line segments forming the long line segment set from the filtered line segment set to update the filtered line segment set, and return to the step of determining the longest line segment in the filtered line segment set as a target line segment.

[0122] A filtered line segment set update unit is configured to, if the first judgment result is no, delete the line segments forming the long line segment set from the filtered line segment set to update the filtered line segment set, and return to the step of determining the longest line segment in the filtered line segment set as a target line segment.

[0123] As an optional implementation, the screening module specifically includes:

[0124] A mean value and standard deviation determination unit is configured to calculate the mean value and standard deviation of the slope according to the slope of all long line segments.

[0125] A Z-score value calculation unit is configured to calculate a Z-score value of the slope of the long line segment according to the average value of the slope, the standard deviation of the slope and the slope of the long line segment.

[0126] An overhead transmission line edge line segment calculation unit is configured to delete the long line segment with the Z-score value greater than the second set threshold to obtain the overhead transmission line edge line segment.

[0127] The embodiment of the present application further provides an electronic device, which comprises:

[0128] The electronic device further comprises a memory and a processor, wherein the memory is configured to store a computer program, and the processor is configured to execute the computer program to enable the electronic device to perform the overhead transmission line edge positioning method according to the above embodiment.

[0129] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the overhead transmission line edge positioning method according to the above embodiment.

[0130] The embodiment of the present application has the following technical effects:

[0131] The embodiment of the present application enables the inspection work of the overhead transmission line to be no longer realized by means of artificial inspection, and the power grid safety maintenance personnel can directly utilize the unmanned aerial vehicle to automatically survey the situation of the overhead transmission line in the power grid system, so that the risk assessment and early warning can be realized in time.

[0132] The embodiment of the present application analyzes the overhead transmission line image data collected by the camera device mounted on the unmanned aerial vehicle, reduces the performance degradation caused by image noise in the preprocessing stage by means of the Gaussian blur technology, and innovatively introduces three statistical priors when processing the line segment result set of the LSD line segment detector: Prior one: in a real scene, the unmanned aerial vehicle mostly performs image collection of the overhead transmission line in a lateral flight attitude, at this time, the tilt angle of the transmission line in the image is concentrated in the interval of [-15°, 15°], and the line segment detected on the edge thereof is also the same. Prior two: the line segments belonging to the same edge are close to the same straight line, have an approximate straight line equation, and the property of the LSD line segment detector makes the longer the line segment, the greater the possibility of correct prediction. Prior three: the orientations of the transmission lines appearing in the same image are approximate, and therefore the line segments detected on the edges thereof also have similar slopes. Processing the line segments in the LSD result set based on the three priors can filter the misdetected line segments in the LSD result set, and repair the fragmented short line segments into ideal edge long line segments, and has stronger robustness in a complex scene, effectively inhibits the influence of image noise in the foreground and background on the detection result, and even if misdetection occurs, can be adaptively corrected.

[0133] The various embodiments described in this specification are presented for the purpose of illustrating the principles of the present application and its best mode of operation. Each of the embodiments described in this specification has been provided for the purpose of illustration only and the various embodiments are not intended to limit the present application in any way unless otherwise specifically indicated. The same parts and / or features of the various embodiments described in this specification can be referenced using the same reference numerals for the ease of understanding of the present application.

[0134] The principles and implementations of the present application have been described in the above embodiments, which are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation and application range of the present application can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for locating the edge of an overhead transmission line, characterized in that, include: Acquire images of overhead power transmission lines; Gaussian blur is applied to the image of the overhead transmission line to obtain the blurred image; The blurred image is processed using the LSD line segment detection algorithm to obtain a set of line segments; The line segments in the line segment set are filtered according to the set slope range and the slope of each line segment in the line segment set to obtain a filtered line segment set. Linear aggregation is performed on the line segments in the filtered line segment set to obtain multiple long line segments; specifically, the longest line segment in the filtered line segment set is taken as the target line segment. The aggregated line segment set and the target line segment are combined to form a long line segment set; the aggregated line segment set includes all line segments in the filtered line segment set that meet the aggregation conditions with the target line segment; the aggregation conditions include the difference in slope between two line segments being less than a set slope difference threshold or the distance between two line segments being less than a set distance threshold; Determine whether the total length of all line segments in the set of long line segments is greater than a first preset threshold, and obtain a first determination result; If the first judgment result is yes, then all the line segments in the long line segment set are connected into a long line segment, and the line segments that make up the long line segment set are deleted from the filter line segment set to update the filter line segment set, and the step of "taking the longest line segment in the filter line segment set as the target line segment" is returned. If the first judgment result is negative, the line segments that make up the long line segment set are deleted from the filter line segment set to update the filter line segment set, and the step of "taking the longest line segment in the filter line segment set as the target line segment" is returned. All long line segments are filtered according to their slopes to obtain the edge line segments of the overhead transmission line. Specifically, this involves calculating the average slope and the standard deviation of the slope based on the slopes of all long line segments. For any long line segment, calculate the Z-socre value of the slope of the long line segment based on the average slope, the standard deviation of the slope, and the slope of the long line segment; Delete long line segments with Z-socre values ​​greater than the second set threshold to obtain the edge line segments of overhead transmission lines; The edge of the overhead power line is determined based on the edge segment of the overhead power line.

2. The method for locating the edge of an overhead transmission line according to claim 1, characterized in that, The step of filtering the line segments in the line segment set according to the set slope range and the slope of each line segment in the line segment set to obtain a filtered line segment set specifically includes: Determine the slope of each line segment in the set of line segments; A filtered line segment set is obtained based on the slope of each line segment in the line segment set and the set slope interval; the filtered line segment set includes all line segments whose slope is within the set slope interval.

3. An edge positioning system for overhead transmission lines, characterized in that, include: The acquisition module is used to acquire images of overhead transmission lines; A Gaussian blur module is used to apply Gaussian blur to the image of the overhead transmission line to obtain a blurred image. The LSD line segment detector is used to process the blurred image using the LSD line segment detection algorithm to obtain a set of line segments; The filter segment set determination module is used to filter the line segments in the line segment set according to a set slope range and the slope of each line segment in the line segment set, so as to obtain a filter segment set. The linear aggregation module is used to linearly aggregate the line segments in the filtered line segment set to obtain multiple long line segments; Specifically, it includes: The target line segment determination unit is used to select the longest line segment in the filtered line segment set as the target line segment. A long line segment set determination unit is used to combine an aggregated line segment set and the target line segment into a long line segment set; the aggregated line segment set includes all line segments in the filtered line segment set that meet the aggregation conditions with the target line segment; the aggregation conditions include the difference in slope between two line segments being less than a set slope difference threshold or the distance between two line segments being less than a set distance threshold; The judgment unit is used to determine whether the total length of all line segments in the set of long line segments is greater than a first preset threshold, and to obtain a first judgment result. The long line segment determination and filter line segment set update unit is used to connect all the line segments in the long line segment set into a long line segment if the first judgment result is yes, and delete the line segments that make up the long line segment set from the filter line segment set to update the filter line segment set, and return to the step of "taking the longest line segment in the filter line segment set as the target line segment". The filter segment set update unit is used to delete the line segments that make up the long line segment set from the filter segment set to update the filter segment set if the first judgment result is negative, and return to the step of "taking the longest line segment in the filter segment set as the target line segment". The filtering module is used to filter all long line segments according to their slopes to obtain the edge segments of overhead transmission lines; specifically, it includes: The mean and standard deviation determination unit is used to calculate the mean and standard deviation of the slope based on the slope of all long line segments; The Z-socre value calculation unit is used to calculate the Z-socre value of the slope of any long line segment based on the average slope, the standard deviation of the slope, and the slope of the long line segment. The overhead transmission line edge segment calculation unit is used to delete long segments with Z-socre values ​​greater than a second set threshold to obtain the overhead transmission line edge segments. An overhead transmission line edge determination module is used to determine the edge of the overhead transmission line based on the edge segment of the overhead transmission line.

4. The overhead transmission line edge positioning system according to claim 3, characterized in that, The filter segment set determination module specifically includes: A slope calculation unit is used to determine the slope of each line segment in the set of line segments; A filter segment set determination unit is used to obtain a filter segment set based on the slope of each segment in the segment set and the set slope interval; the filter segment set includes all segment whose slope is within the set slope interval.

5. An electronic device, characterized in that, include: A memory and a processor, the memory being used to store a computer program, the processor running the computer program to cause the electronic device to perform the overhead transmission line edge positioning method according to any one of claims 1 to 2.

6. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the overhead transmission line edge positioning method as described in any one of claims 1 to 2.

Images