Power transmission line segmentation method based on laser radar point cloud

Abstract

The invention relates to the technical field of information technology and environmental perception, and discloses a power transmission line segmentation method based on a laser radar point cloud, which comprises the following steps: collecting power transmission line point cloud data by using a laser radar, and carrying out preprocessing such as filtering and downsampling; calculating a local vector feature of each point in the point cloud according to a laser radar point position relationship, and extracting points conforming to the features of the power transmission line; extracting point cloud clusters of the segmented power transmission line by using a clustering method, and obtaining boundary points of each point cloud cluster; and combining the segmented power transmission lines by means of boundary points and boundary vectors of the segmented power transmission lines in a region search mode. The method can provide support for environment perception related application fields such as obstacle avoidance and aerial target monitoring, and has a good market application prospect.

Description

Technical Field

[0001] The present invention relates to the fields of information technology and environmental perception technology, and in particular to a transmission line segmentation method based on laser radar point cloud. Background Art

[0002] With the rapid development of the power industry and the advancement of smart grids, the safety monitoring and maintenance of transmission lines has become an important part of the stable operation of the power grid. The traditional inspection method of transmission lines mainly relies on manual inspection, which is not only inefficient but also difficult to adapt to complex and changing environmental and weather conditions. Therefore, the use of advanced technical means to achieve automated and intelligent monitoring of transmission lines has become an important development trend in the current power industry.

[0003] As an advanced measurement technology, LiDAR (Light Detection and Ranging) has been widely used in the field of power transmission line monitoring due to its advantages of high precision, high efficiency, and all-weather capability. By emitting laser beams and receiving reflected signals, LiDAR can obtain the spatial position and shape information of the target object, and then generate high-precision point cloud data. These point cloud data not only contain the geometric information of the transmission line, but also the environmental information around it, providing rich data support for the safety monitoring of the transmission line.

[0004] The current automatic segmentation and extraction algorithms for transmission lines can be divided into two categories. The first type of algorithm combines elevation threshold segmentation and spatial domain segmentation algorithms to perform rough extraction of transmission lines, and then uses Hough transform or RANSAC to fit the 2D projection of the transmission lines. However, this method is more complicated to extract and obtain the projection plane of the transmission lines, and both require the extraction of a large range of features in the point cloud, and the extraction speed is slow. The second type of algorithm uses a neural network to identify transmission lines in point cloud data, but this type of algorithm requires a large amount of training data and the calibration labels are more complex, and requires high computing power. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that global features need to be obtained when extracting power lines based on lidar point clouds in the past, and propose a method for extracting power lines based on local vector features.

[0006] The technical solution of the present invention is achieved in this way:

[0007] A method for segmenting a power line based on a laser radar point cloud comprises the following steps:

[0008] Step 1: Use LiDAR to collect transmission line point cloud data and perform preprocessing such as filtering and downsampling;

[0009] Step 2: According to the position relationship of the LiDAR points, the local vector features of each point in the point cloud are calculated, and the points that meet the characteristics of the transmission line are extracted;

[0010] Step 3: Use the clustering method to extract the point cloud clusters of the segmented transmission line and obtain the boundary points of each point cloud cluster;

[0011] Step 4: Use the regional search method to merge the segmented transmission lines with the help of their boundary points and boundary vectors.

[0012] As a further solution of the present invention: the step 1 specifically includes the following contents:

[0013] Use LiDAR to collect environmental information and perform data preprocessing to enhance local vector features of transmission lines;

[0014] Filter out the noise near the sensor in the point cloud data based on the distance information and set the area of ​​interest;

[0015] Downsample the point cloud data to reduce the number of point clouds, and speed up the subsequent processing of the point cloud without excessively eliminating the details of the point cloud.

[0016] As a further solution of the present invention: the step 2 specifically includes the following contents:

[0017] Calculate the local vector features of each point in the point cloud, count and extract the points that meet the local vector features of the transmission line;

[0018] For any point in the point cloud, search for the three nearest neighboring points of point; set a neighborhood distance threshold neighThresh, and count the number n of local neighboring points whose distance to point is less than neighThresh among the three neighboring points:

[0019] If n = 0, point is an outlier and its local space vector cannot be calculated;

[0020] If n=1, point has the local vector feature of the transmission line boundary, that is, point may be the boundary point of a certain section of the transmission line; calculate the vector vector=point1-point formed by point and the local neighboring point point1, then boundVector=-vector is its boundary vector, that is, the direction to be searched when looking for points belonging to the same section of the transmission line in the subsequent search;

[0021] If n = 2, calculate the vectors vector1 = poini1-point and vector2 = point2-point formed by point and the two local neighboring points point1 and point2, and calculate the angle θ between the two vectors v, the formula is as follows:

[0022]

[0023] The vector angle θ is obtained v Convert to straight line angle θ l , the formula is as follows:

[0024]

[0025] If the straight line angle θ l If the angle is less than the angle threshold angleThresh, then point has the local vector feature of the transmission line, that is, point may be a point in the middle of a certain transmission line; and if θ v <angleThresh, then point may be a boundary point, and its boundary vector boundVector=-(vector1+vector2);

[0026] If n = 3, calculate the vectors formed by point and the three local neighborhood points point1, point2, point3: vector1 = point1-point, vector2 = point2-point, vector3 = point2-point, and calculate the angle between any two of these three vectors: The formula is as follows:

[0027]

[0028] The resulting vector angle Convert to straight line angle The formula is as follows

[0029]

[0030] If the straight line angle are all less than the angle threshold angleThresh, then point has the local vector feature of the transmission line, and if Then point has the local vector characteristics of the transmission line boundary, and its boundary vector boundVector=-(vector1+vector2+vector3);

[0031] All points with local vector features of transmission lines and local vector features of boundaries are extracted.

[0032] As a further solution of the present invention: Step 3 specifically includes the following contents:

[0033] After the processing in step 2, some points in the point cloud will be mistakenly identified as points on the transmission line and retained as outliers, or because of the disappearance of neighboring points, the points in the middle of the transmission line will become boundary points. Therefore, it is necessary to repeat the same processing in step 2 for the point cloud processed in step 2, filter out the misidentified points, and find new boundary points;

[0034] The clustering method is used to merge two points whose middle distance in the point cloud is less than neighThresh into a cluster of point clouds, thereby obtaining several clusters of point clouds that may be transmission lines.

[0035] As a further solution of the present invention: the step 4 specifically includes the following contents:

[0036] The point cloud clusters belonging to the same transmission line are searched along the boundary vector direction of the boundary points in the point cloud cluster by using the regional search method;

[0037] For any point cloud cluster in the merged result of step 3, find two boundary points and two boundary vectors in each point cloud cluster and perform recursive processing;

[0038] After the recursion of all clusters is completed, the number of points in all clusters found is determined. If the number is greater than the number threshold minPointNum, all the points found are recorded as a complete transmission line.

[0039] As a further solution of the present invention: the recursive processing specifically includes the following contents:

[0040] Generate k position coordinates with interval interval from the boundary point position to the direction of the boundary vector, and search whether there is a point pointNearset closest to a certain position coordinate in the point cloud:

[0041] If pointNearset exists, find the cluster clusterFound where pointNearset is located, find the boundary vector boundvectorNext in clusterFound that has an acute angle with the boundary vector of cluster, and perform the above recursive operation on this boundary vector boundvectorNext;

[0042] If pointNearset does not exist, the recursion ends.

[0043] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0044] A computer storage medium stores a computer program, which implements the steps of the above method when executed by a processor.

[0045] Beneficial effects of the present invention:

[0046] 1. The present invention regards each point in the lidar point cloud and its three nearest neighboring points as a local space, calculates and statistically determines whether the vector features inside each local space conform to the local vector features of the transmission line, and then uses a clustering method to obtain the segmented transmission line. Finally, recursively search for other transmission lines at the end of the segmented transmission line in its direction to extract the entire transmission line.

[0047] 2. The present invention can provide support for environmental perception-related application fields such as obstacle avoidance and aerial target monitoring, and has good market application prospects.

[0048] The present invention will be further described below in conjunction with the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of a flow chart of a method for segmenting a power line based on a laser radar point cloud according to the present invention;

[0050] Figure 2 This is a physical picture of the sensor of the present invention;

[0051] Figure 3 This is a schematic diagram of the point being an outlier point when the number of neighborhood points n=0 in the present invention;

[0052] Figure 4 This is a schematic diagram of the present invention where point is a boundary point when the number of neighborhood points n=1;

[0053] Figure 5 Schematic diagram of judging point category when the number of neighborhood points n=2 in the present invention, wherein (a) point is an on-line point, (b) point is a boundary point, and (c) point is a non-on-line point;

[0054] Figure 6 Schematic diagram of judging point category when the number of neighborhood points n=3 in the present invention, wherein (a) point is an on-line point, (b) point is a boundary point, and (c) point is a non-on-line point;

[0055] Figure 7 The point cloud clusters of the present invention are merged into a complete transmission line schematic diagram;

[0056] Figure 8 This is a schematic diagram of the original point cloud data of the present invention;

[0057] Fig. 9 This is a schematic diagram of the transmission line extracted by the present invention. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0059] First, see Appendix Figure 1-9 The present invention discloses a transmission line segmentation method based on laser radar point cloud, which comprises the following steps:

[0060] Step 1: Use LiDAR to collect transmission line point cloud data and perform preprocessing such as filtering and downsampling;

[0061] Step 2: According to the position relationship of the LiDAR points, the local vector features of each point in the point cloud are calculated, and the points that meet the characteristics of the transmission line are extracted;

[0062] Step 3: Use the clustering method to extract the point cloud clusters of the segmented transmission line and obtain the boundary points of each point cloud cluster;

[0063] Step 4: Use the regional search method to merge the segmented transmission lines with the help of their boundary points and boundary vectors.

[0064] In the specific implementation process of the present invention, a 3D laser radar sensor with 400 lines per frame is used to generate a 3D point cloud at a frequency of 10 Hz, and the scanning range in the vertical direction is 120° (±60°), the scanning range in the vertical direction is 25° (±12.5°), the horizontal angle resolution is 0.0648°, and the vertical angle resolution is 0.05°. Figure 1 The device is used to collect radar point cloud data and extract a frame of point cloud from it, as shown in Figure 2 shown.

[0065] Step 1: Use LiDAR to collect point cloud data and perform data preprocessing to enhance the local vector features of the transmission line;

[0066] Use a straight-through filter to select the point cloud within the range of 15m<y<150m as the region of interest (there is noise within 15m, and the power lines beyond 150m cannot be detected), and remove the noise near the sensor in the point cloud data;

[0067] The point cloud data is downsampled using a voxel filter, and the voxel size is selected as (0.25m, 0.25m, 0.25m). The number of point clouds is reduced, and the speed of subsequent point cloud processing is accelerated without excessively eliminating the details of the point clouds. After voxel filtering, each point in the point cloud belonging to the transmission line has a very simple local vector feature that is different from other points: it is only linear within a certain spatial range, which is conducive to the subsequent segmentation and extraction of the transmission line.

[0068] Step 2: Calculate the local vector features of each point in the point cloud, count and extract the points that meet the local vector features of the transmission line;

[0069] For any point in the point cloud, search for the three nearest neighboring points of point. Set a neighborhood distance threshold neighThresh = 0.4m, and count the number of local neighboring points n whose distance to point is less than neighThresh among the three neighboring points:

[0070] If n = 0, point is an outlier and its local space vector cannot be calculated;

[0071] If n=1, point has the local vector feature of the transmission line boundary, that is, point may be the boundary point of a certain section of the transmission line. Calculate the vector vector formed by point and the local neighboring point point1, vector=point1-point, then boundVector=-vectoc is its boundary vector, that is, the direction to be searched when looking for points belonging to the same section of the transmission line in the future;

[0072] If n = 2, calculate the vectors vector1 = point1-point and vector2 = point2-point formed by point and the two local neighboring points point1 and point2, and calculate the angle θ between the two vectors v , the formula is as follows

[0073]

[0074] The vector angle θ is obtained v Convert to straight line angle θ l , the formula is as follows

[0075]

[0076] If the straight line angle θ l If the angle is less than an angle threshold angleThresh = 30°, then point has the local vector feature of the transmission line, that is, point may be a point in the middle of a certain transmission line. v <angleThresh, then point may be a boundary point, and its boundary vector boundVectoc=-(vector1+vector2);

[0077] If n = 3, calculate the vectors formed by point and the three local neighboring points point1, point2, point3: vactor1 = point1-point, vector2 = point2-point, vector3 = point3-point, and calculate the angle between any two of these three vectors. The formula is as follows

[0078]

[0079] The resulting vector angle Convert to straight line angle θ l 1 θ l 2 ,θ l 3 , the formula is as follows

[0080]

[0081] If the straight line angle are all less than the angle threshold angleThresh, then point has the local vector feature of the transmission line, and if Then point has the local vector characteristics of the transmission line boundary, and its boundary vector boundVector=-(vector1+vector2+vector3);

[0082] All points with local vector features of transmission lines and local vector features of boundaries are extracted.

[0083] Step 3: Repeat the process of step 2 to filter out the misidentified points, find new boundary points, and then use the clustering method to obtain the point cloud of the transmission line segment;

[0084] After processing in step 2, some points in the point cloud will be mistakenly identified as points on the transmission line and retained as outliers, or because of the disappearance of neighboring points, the points in the middle of the transmission line will become boundary points. Therefore, it is necessary to repeat the same processing in step 2 for the point cloud processed in step 2, filter out the misidentified points, and find new boundary points.

[0085] The clustering method is used to merge two points whose middle distance in the point cloud is less than neighThresh into a cluster of point clouds, thereby obtaining several clusters of point clouds that may be transmission lines.

[0086] Step 4: Use the regional search method to search for point cloud clusters belonging to the same transmission line along the boundary vector direction of the boundary points in the point cloud cluster;

[0087] For any point cloud cluster in the merged result of step 3, find two boundary points and two boundary vectors in each cluster point cloud, and recursively perform the following processing respectively:

[0088] Generate k = 10 position coordinates with an interval of interval = 0.2 from the boundary point position to the direction of the boundary vector, and search whether there is a point pointNearset closest to a certain position coordinate in the point cloud:

[0089] If pointNearset exists, find the cluster clusterFound where pointNearset is located, find the boundary vector boundvectorNext in clusterFound that has an acute angle with the boundary vector of cluster, and perform the above recursive operation on this boundary vector boundvectorNext.

[0090] If pointNearset does not exist, the recursion ends.

[0091] After the recursion of all clusters is completed, the number of points in all clusters found is determined. If the number is greater than the number threshold minPointNum=20, all the points found are recorded as a complete transmission line.

[0092] The frequency of the transmission line can reach 9 Hz using this method.

[0093] In a second aspect, the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0094] In a third aspect, the present invention provides a computer storage medium having a computer program stored thereon, which implements the steps of the above method when executed by a processor.

[0095] So far, the purpose of the present invention has been achieved.

[0096] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for segmenting a power line based on a laser radar point cloud, characterized in that: The steps include: Step 1: Use LiDAR to collect transmission line point cloud data and perform preprocessing such as filtering and downsampling; Step 2: According to the position relationship of the LiDAR points, the local vector features of each point in the point cloud are calculated, and the points that meet the characteristics of the transmission line are extracted; Step 3: Use the clustering method to extract the point cloud clusters of the segmented transmission line and obtain the boundary points of each point cloud cluster; Step 4: Use the regional search method to merge the segmented transmission lines with the help of their boundary points and boundary vectors.

2. The method for segmenting a power line based on a laser radar point cloud according to claim 1, characterized in that: The step 1 specifically includes the following contents: Use LiDAR to collect environmental information and perform data preprocessing to enhance local vector features of transmission lines; Filter out the noise near the sensor in the point cloud data based on the distance information and set the area of ​​interest; Downsample the point cloud data to reduce the number of point clouds, and speed up the subsequent processing of the point cloud without excessively eliminating the details of the point cloud.

3. The method for segmenting a power line based on a laser radar point cloud according to claim 2, characterized in that: The step 2 specifically includes the following contents: Calculate the local vector features of each point in the point cloud, count and extract the points that meet the local vector features of the transmission line; For any point in the point cloud, search for the three nearest neighboring points of point; set a neighborhood distance threshold neighThresh, and count the number n of local neighboring points whose distance to point is less than neighThresh among the three neighboring points: If n = 0, point is an outlier and its local space vector cannot be calculated; If n=1, point has the local vector feature of the transmission line boundary, that is, point may be the boundary point of a certain section of the transmission line; calculate the vector vector=point1-point formed by point and the local neighboring point point1, then boundVector=-vector is its boundary vector, that is, the direction to be searched when looking for points belonging to the same section of the transmission line in the subsequent search; If n = 2, calculate the vectors vector1 = point1-point and vector2 = point2-point formed by point and the two local neighboring points poini1 and point2, and calculate the angle θ between the two vectors v , the formula is as follows: The vector angle θ is obtained v Convert to straight line angle θ l , the formula is as follows: If the straight line angle θ l If the angle is less than the angle threshold angleThresh, then point has the local vector feature of the transmission line, that is, point may be a point in the middle of a certain transmission line; and if θ v <angleThresh, then point may be a boundary point, and its boundary vector boundVector=-(vector1+vector2); If n = 3, calculate the vectors formed by point and the three local neighboring points point1, point2, point3: vector1 = point1-point, vector2 = point2-point, vector3 = point3-point, and calculate the angle between any two of these three vectors The formula is as follows: The resulting vector angle Convert to straight line angle The formula is as follows If the straight line angle are all less than the angle threshold angleThresh, then point has the local vector feature of the transmission line, and if Then point has the local vector characteristics of the transmission line boundary, and its boundary vector boundVector=-(vector1+vector2+vector3); All points with local vector features of transmission lines and local vector features of boundaries are extracted.

4. The method for segmenting a power line based on a laser radar point cloud according to claim 3, characterized in that: The step 3 specifically includes the following contents: After the processing in step 2, some points in the point cloud will be mistakenly identified as points on the transmission line and retained as outliers, or because of the disappearance of neighboring points, the points in the middle of the transmission line will become boundary points. Therefore, it is necessary to repeat the same processing in step 2 for the point cloud processed in step 2, filter out the misidentified points, and find new boundary points; The clustering method is used to merge two points whose middle distance in the point cloud is less than neighThresh into a cluster of point clouds, thereby obtaining several clusters of point clouds that may be transmission lines.

5. The method for segmenting a power line based on a laser radar point cloud according to claim 4, characterized in that: The step 4 specifically includes the following contents: The point cloud clusters belonging to the same transmission line are searched along the boundary vector direction of the boundary points in the point cloud cluster by using the regional search method; For any point cloud cluster in the merged result of step 3, find two boundary points and two boundary vectors in each point cloud cluster and perform recursive processing; After the recursion of all clusters is completed, the number of points in all clusters found is determined. If the number is greater than the number threshold minPointNum, all the points found are recorded as a complete transmission line.

6. The method for segmenting a power line based on a laser radar point cloud according to claim 5, characterized in that: The recursive processing specifically includes the following contents: Generate k position coordinates with interval interval from the boundary point position to the direction of the boundary vector, and search whether there is a point pointNearset closest to a certain position coordinate in the point cloud: If pointNearset exists, find the cluster clusterFound where pointNearset is located, find the boundary vector boundvectorNext in clusterFound that has an acute angle with the boundary vector of cluster, and perform the above recursive operation on this boundary vector boundvectorNext; If pointNearset does not exist, the recursion ends.

7. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the methods of claims 1 to 6 when executing the computer program.

8. A computer storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are implemented.

Images