A tree encroachment recognition system and method based on geometric consistency and dynamic adaptation
By constructing a KD-Tree index and a geometric consistency filtering mechanism, the problems of high false recognition rate and low processing efficiency in tree encroachment identification during power line inspection are solved, achieving efficient and accurate tree encroachment detection, and adapting to power transmission line scenarios with different geographical environments and voltage levels.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-17
- Publication Date
- 2026-07-10
Smart Images
Figure CN122368871A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of power line inspection technology, and in particular relates to a tree encroachment identification system and method based on geometric consistency and dynamic adaptation. Background Technology
[0002] The safe and stable operation of power lines is crucial to the national economy and social life. Trees growing along transmission line corridors can encroach on the safe distance of the lines, causing serious accidents such as power outages, tripping, and even fires. Therefore, regular inspections of the line corridors to promptly identify and address potential tree encroachment hazards are one of the core tasks of power grid operation and maintenance.
[0003] Traditional inspection methods rely primarily on manual patrols, which are inefficient, costly, and highly susceptible to subjective factors and complex terrain. With the widespread adoption of lidar technology, automatic detection technology based on airborne or ground-based lidar point cloud data has become a mainstream research and application direction. Existing solutions typically follow this process: first, power lines, poles, vegetation, and other features are segmented and classified from the original point cloud; then, the Euclidean distance between vegetation points and the nearest power line point is calculated; finally, the distance is compared with a preset safety threshold to determine if an intrusion has occurred. However, these existing solutions face two prominent technical challenges in practical applications, especially when processing large-scale, complex point cloud data: First, high misidentification rates in intersection areas. Power poles and power lines are physically connected, and their point clouds are mixed in intersection areas. Points in this area simultaneously possess high linearity and high-altitude characteristics. Using conventional classification methods based on geometric features and fixed thresholds, it is easy to misclassify some points at the top of the pole as power line points, or misidentify points of power lines near the pole, leading to numerous false alarms and interfering with maintenance decisions. Second, low efficiency in large-scale data processing. To extract local geometric features, a neighborhood search is required for each point. Traditional methods employ sequential traversal or brute-force search, which have high time complexity and are difficult to meet the real-time or near-real-time processing requirements of large point clouds. Based on the above reasons, this invention designs a tree encroachment recognition system and method based on geometric consistency and dynamic adaptation. Summary of the Invention
[0004] The purpose of this invention is to solve the problems in the prior art, and to propose a tree encroachment identification system and method based on geometric consistency and dynamic adaptation.
[0005] This invention first discloses a tree encroachment identification system based on geometric consistency and dynamic adaptation, comprising: The point cloud data acquisition module is used to acquire three-dimensional laser point cloud data of the power transmission line corridor; The spatial index construction module is used to construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; The feature processing module, connected to the spatial index construction module, is used to extract local geometric features from the point cloud using the KD-Tree spatial index, and classify the point cloud into at least a set of power lines and a set of vegetation points based on the extracted features. The geometric consistency filtering module, connected to the feature processing module, is used to filter the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. The filtering process includes: for each target point in the set of wire points, querying its preset number K nearest neighbor wire points, calculating the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, counting the number of nearest neighbor points whose angle is less than the angle threshold, and filtering the target point based on whether the ratio of the counted number to K is lower than the consistency ratio threshold. The encroachment detection module, connected to the geometric consistency filtering module and the feature processing module, is used to calculate the spatial distance between the power line point and the nearest vegetation point based on the filtered power line point set and the vegetation point set, and to judge and output the tree encroachment detection result according to the safe distance threshold.
[0006] In the above system, the feature processing module includes: The feature extraction unit is used to perform K-nearest neighbor query on each query point in the point cloud to obtain its neighborhood point set, and to perform principal component analysis on the neighborhood point set to calculate linearity, flatness, sphericity and principal direction vector. The classification unit is used to classify the point cloud into at least one of the following: power line points, utility pole points, vegetation points, building points, and ground points, based on the geometric features and point height information output by the feature extraction unit and according to preset classification rules.
[0007] In the above system, the classification rules executed by the classification unit include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If the linearity of a point is greater than the first linearity threshold, the angle between its main direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point. If a point's linearity is between the second and third linearity thresholds and its height falls within the vegetation height range, it is classified as a vegetation point.
[0008] In the above system, before performing directional consistency filtering, the geometric consistency filtering module is also used to perform pole area exclusion: calculate the distance between a point in the power line point set and the nearest pole point, and if the distance is less than a preset pole influence threshold, then the point is excluded first.
[0009] In the above system, after performing directional consistency filtering, the geometric consistency filtering module is also used to perform tower false detection elimination: for the intrusion candidate point pairs initially determined by the intrusion detection module, the distance between the vegetation point and the nearest utility pole point is detected. If the distance is less than a preset threshold, the candidate point pair is determined to be a false detection and eliminated.
[0010] In the above system, after performing directional consistency filtering, the geometric consistency filtering module is also used to perform tower false detection elimination: for the intrusion candidate point pairs initially determined by the intrusion detection module, the distance between the vegetation point and the nearest utility pole point is detected. If the distance is less than a preset threshold, the candidate point pair is determined to be a false detection and eliminated.
[0011] Secondly, the present invention provides a tree encroachment identification method based on geometric consistency and dynamic adaptation, comprising: S1. Acquire 3D laser point cloud data of the power transmission line corridor; S2. Construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; S3. Use the KD-Tree spatial index to extract and classify local geometric features of the point cloud, and at least identify the power line point set and the vegetation point set; S4. Perform geometric consistency filtering on the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. This includes: for each target point in the set of wire points, query its preset number K nearest neighbor wire points, calculate the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, count the number of nearest neighbor points whose angle is less than the angle threshold, and judge and filter the target point based on whether the ratio of the counted number to K is lower than the consistency ratio threshold. S5. Based on the filtered set of power line points and the set of vegetation points, calculate the spatial distance between the power line points and the nearest vegetation points, and determine whether there is a potential risk of tree encroachment based on the safe distance threshold.
[0012] In the above method, step S3 includes: S31. For each query point in the point cloud, perform a K-nearest neighbor query using the KD-Tree to obtain its neighborhood point set; S32. Perform principal component analysis on the neighborhood point set to calculate the linearity, flatness, sphericity, and principal direction vector; S33. Based on the linearity, flatness, sphericity, principal direction vector, and point height information, the point cloud is classified into at least one of the following using preset classification rules: power line points, power pole points, vegetation points, building points, and ground points.
[0013] In the above method, the classification rules in step S33 include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If a point's linearity is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point.
[0014] In the above method, before performing directional consistency filtering, step S4 also includes a pole area exclusion step: calculating the distance between a point in the set of power line points and the nearest power pole point. If the distance is less than a preset pole influence threshold, the point is excluded from the set of power line points.
[0015] In the above method, after performing directional consistency filtering, step S4 also includes a tower false detection elimination step: for candidate point pairs that are initially determined to be encroachments based on distance calculation, it is detected whether the vegetation point is located near the utility pole. If so, the candidate point pair is determined to be a false detection and eliminated.
[0016] The beneficial effects of this invention are as follows: 1. Unlike existing methods that directly classify and judge based on distance thresholds or fixed geometric feature thresholds, this invention introduces a geometric consistency filtering mechanism. This mechanism identifies outliers by quantitatively evaluating the consistency of each wire point with other wire points in its local neighborhood along the principal direction vector of the pre-classified set of wire points. This distinguishes noise points with similar linear characteristics but discrete directional distribution within the intersection area of utility poles and wires from genuine wire points, reducing the false alarm rate of tree encroachment caused by misclassification of point clouds in intersection areas.
[0017] 2. This invention systematically integrates the KD-Tree spatial indexing strategy. This strategy optimizes the time complexity of core operations such as point cloud neighborhood search, nearest neighbor query in feature extraction, and nearest point search in intrusion detection by constructing a tree-like data structure. The effect is to improve the computational speed of large-scale point cloud data throughout the entire processing flow, enabling the algorithm to meet the real-time or near-real-time processing needs of massive point cloud data in long-distance power line corridors.
[0018] 3. Unlike solutions that use fixed thresholds and are difficult to adapt to different inspection scenarios, the angle tolerance threshold and consistency ratio threshold involved in key steps such as geometric consistency filtering in this invention are designed as configurable parameters. This allows the system to adapt to different geographical environments such as plains, mountains, and cities, as well as transmission line scenarios with different voltage levels and sag characteristics, by adjusting these parameters. This enhances the generalization ability and robustness of the entire identification system in practical applications. Attached Figure Description
[0019] Figure 1 This is a flowchart of a tree encroachment identification method based on geometric consistency and dynamic adaptation disclosed in this invention.
[0020] Figure 2 This is a diagram illustrating the working effect of a tree encroachment identification system based on geometric consistency and dynamic adaptation disclosed in this invention. Detailed Implementation
[0021] To facilitate understanding of this application and to make the aforementioned objectives, features, and advantages of this application more apparent, a detailed description of specific embodiments of this application is provided below in conjunction with the accompanying drawings. Numerous specific details are set forth in the following description to provide a thorough understanding of this application, and preferred embodiments are shown in the accompanying drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application. This application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified. In the description of this application, "several" means at least one, such as one, two, etc., unless otherwise explicitly specified. It should be noted that when an element is referred to as being "fixed to" another element, it can be directly attached to the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementations. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is only for describing particular implementations and is not intended to limit the scope of this application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0022] Reference Figure 1-2 A tree encroachment identification system based on geometric consistency and dynamic adaptation includes: The point cloud data acquisition module is used to acquire three-dimensional laser point cloud data of the power transmission line corridor; The spatial index construction module is used to construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; A feature processing module, connected to the spatial index construction module, is used to extract local geometric features from the point cloud using the KD-Tree spatial index, and classify the point cloud into at least a set of power line points and a set of vegetation points based on the extracted features. The feature processing module includes: The feature extraction unit is used to perform K-nearest neighbor query on each query point in the point cloud to obtain its neighborhood point set, and to perform principal component analysis on the neighborhood point set to calculate linearity, flatness, sphericity and principal direction vector. The classification unit is used to classify the point cloud into at least one of the following: power line points, utility pole points, vegetation points, building points, and ground points, based on the geometric features and point height information output by the feature extraction unit and according to preset classification rules.
[0023] The classification rules executed by the classification unit include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If the linearity of a point is greater than the first linearity threshold, the angle between its main direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point. If a point's linearity is between the second and third linearity thresholds and its height falls within the vegetation height range, it is classified as a vegetation point.
[0024] The geometric consistency filtering module, connected to the feature processing module, is used to filter the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. Its filtering logic includes: for each target point in the set of wire points, querying its preset number K nearest neighbor wire points, calculating the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, counting the number of nearest neighbor points whose angle is less than the angle threshold, and filtering the target point based on whether the ratio of the counted number to K is lower than the consistency ratio threshold. The encroachment detection module, connected to the geometric consistency filtering module and the feature processing module, is used to calculate the spatial distance between the power line point and the nearest vegetation point based on the filtered power line point set and the vegetation point set, and to judge and output the tree encroachment detection result according to the safe distance threshold.
[0025] Secondly, the present invention provides a tree encroachment identification method based on geometric consistency and dynamic adaptation, comprising the following steps: S1. Acquire 3D laser point cloud data of the power transmission line corridor; S2. Construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; S3. Using the KD-Tree spatial index, perform local geometric feature extraction and classification on the point cloud, identifying at least the power line point set and the vegetation point set, specifically including the following steps: S31. For each query point in the point cloud, perform a K-nearest neighbor query using the KD-Tree to obtain its neighborhood point set; S32. Perform principal component analysis on the neighborhood point set to calculate the linearity, flatness, sphericity, and principal direction vector; S33. Based on the linearity, flatness, sphericity, principal direction vector and point height information, the point cloud is classified into at least one of power line points, power pole points, vegetation points, building points and ground points according to the preset classification rules; The classification rules in step S33 include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If a point's linearity is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point.
[0026] S4. Perform geometric consistency filtering on the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. This includes: for each target point in the set of wire points, query its preset number K nearest neighbor wire points, calculate the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, count the number of nearest neighbor points whose angle is less than the angle threshold, and judge and filter the target point based on whether the ratio of the counted number to K is lower than the consistency ratio threshold. Before performing directional consistency filtering, step S4 also includes a pole area exclusion step: calculating the distance between a point in the set of power line points and the nearest power pole point. If the distance is less than a preset pole influence threshold, the point is excluded from the set of power line points.
[0027] After performing directional consistency filtering in step S4, a tower false detection elimination step is also included: for candidate point pairs that are initially determined to be encroachments based on distance calculation, it is checked whether the vegetation points are located near the utility pole. If so, the candidate point pair is determined to be a false detection and eliminated.
[0028] S5. Based on the filtered set of power line points and the set of vegetation points, calculate the spatial distance between the power line points and the nearest vegetation points, and determine whether there is a potential risk of tree encroachment based on the safe distance threshold.
[0029] In specific use of this invention, the working process of the geometric consistency filtering module includes: First, perform region filtering, specifically including... 1. Filtering out areas near utility poles: For each utility pole point, calculate its distance to the nearest utility pole. If the distance is less than a preset threshold (e.g., 3 meters), the point is considered to be within the influence zone of the utility pole and is excluded.
[0030] 2. Tower detection for target vegetation points: For each candidate point pair for encroachment (power line point - vegetation point), detect whether the vegetation point is located near the power pole. If the distance between the vegetation point and the nearest power pole is less than a threshold, it is considered a misclassified tower structure, and the encroachment pair is excluded.
[0031] 3. Geometric Consistency Verification: For each wire point, query its 20 nearest neighbor wire points, calculate the angle between the main direction vector of the point and the main direction vectors of each neighbor point, count the number of neighbors whose angle is less than the angle threshold, calculate the consistency ratio, and if the consistency ratio is lower than the set threshold, the point is considered to be in the structural inconsistency area and is filtered out.
[0032] The formula for calculating the angle between two principal direction vectors is:
[0033] in, Let this be the principal direction vector at that point. Let be the direction vector of the neighboring point; if If they are in the same direction, then it is assumed that they are in the same direction. The angle tolerance threshold is used to determine whether the angle between two direction vectors is within the same range. Its function is to control the sensitivity of attitude consistency verification, and it can be modified according to the actual situation. Then, geometric features are extracted. Principal component analysis is used to extract the geometric features of the local neighborhood of the point cloud, including linearity, flatness, sphericity, and principal orientation angle, providing feature input for subsequent classification. At the same time, a KD-Tree is used to create an efficient index to facilitate the search for point P and its K nearest neighbors. The process includes the following steps: For point P and its K nearest neighbors
[0034] Calculate its centroid C: ; Calculate the neighborhood covariance matrix and perform eigenvalue decomposition to obtain the eigenvalues and principal direction vector. The formula is as follows: This lays the foundation for subsequent work.
[0035] Calculate linearity, flatness, and sphericity based on eigenvalues: Linearity = Flatness = linearity = The principal direction angle is the angle between the main direction and the vertical direction. ,in It is a vector in the vertical direction.
[0036] The feature classification module's workflow includes: Based on geometric features and height thresholds, point cloud data is automatically classified into categories such as power lines, utility poles, vegetation, buildings, and ground. The system classifies targets based on the following geometric features: Wires: High linearity + high altitude + horizontal linear distribution; Utility poles: high linearity + high altitude + vertical current distribution; Vegetation: linear and irregular spherical distribution of low, medium and high vegetation types; Buildings: Flatness + height block distribution.
[0037] In practice, the classification rules can be set as follows: The power lines have a linearity greater than 0.6, a main direction angle greater than 60°, and a height exceeding 10 meters. The utility poles must have a linearity greater than 0.6, a main direction angle less than 30°, and a height exceeding 3 meters. The vegetation is divided into three types: low, medium, and high, with a height range of 2 to 15 meters. The building has a flatness greater than 0.5 and a height exceeding 2 meters.
[0038] In the encroachment detection module, the Euclidean distance between the power line point and the nearest vegetation point is calculated to determine if it is an encroachment point. If the distance is less than a safety threshold, it is determined to be an encroachment point. The safety threshold can be set to 0.6 meters for two three-dimensional points. and The calculation formula is as follows:
[0039] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative and not exhaustive. All modifications within the scope of this invention or its equivalents are included in this invention.
Claims
1. A tree encroachment identification system based on geometric consistency and dynamic adaptation, characterized in that, include: The point cloud data acquisition module is used to acquire three-dimensional laser point cloud data of the power transmission line corridor; The spatial index construction module is used to construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; The feature processing module, connected to the spatial index construction module, is used to extract local geometric features from the point cloud using the KD-Tree spatial index, and classify the point cloud into at least a set of power lines and a set of vegetation points based on the extracted features. The geometric consistency filtering module, connected to the feature processing module, is used to filter the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. The filtering includes: for each target point in the set of wire points, querying its preset number K nearest neighbor wire points, calculating the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, counting the number of nearest neighbor points whose angle is less than an angle threshold, and filtering the target point based on whether the ratio of the counted number to K is lower than a consistency ratio threshold. The encroachment detection module, connected to the geometric consistency filtering module and the feature processing module, is used to calculate the spatial distance between the power line point and the nearest vegetation point based on the filtered power line point set and the vegetation point set, and to judge and output the tree encroachment detection result according to the safe distance threshold.
2. The system according to claim 1, characterized in that, The feature processing module includes: The feature extraction unit is used to perform K-nearest neighbor query on each query point in the point cloud to obtain its neighborhood point set, and to perform principal component analysis on the neighborhood point set to calculate linearity, flatness, sphericity and principal direction vector. The classification unit is used to classify the point cloud into at least one of the following: power line points, utility pole points, vegetation points, building points, and ground points, based on the geometric features and point height information output by the feature extraction unit and according to preset classification rules.
3. The system according to claim 2, characterized in that, The classification rules executed by the classification unit include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If the linearity of a point is greater than the first linearity threshold, the angle between its main direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point. If a point's linearity is between the second and third linearity thresholds and its height falls within the vegetation height range, it is classified as a vegetation point.
4. The system according to claim 1, characterized in that, Before performing directional consistency filtering, the geometric consistency filtering module is also used to perform pole area exclusion: calculate the distance between a point in the power line point set and the nearest pole point. If the distance is less than a preset pole influence threshold, then the point is excluded first.
5. The system according to claim 1 or 4, characterized in that, After performing directional consistency filtering, the geometric consistency filtering module is also used to perform tower false detection elimination: for the intrusion candidate point pairs initially determined by the intrusion detection module, the distance between the vegetation point and the nearest utility pole point is detected. If the distance is less than a preset threshold, the candidate point pair is determined to be a false detection and eliminated.
6. A tree encroachment identification method based on geometric consistency and dynamic adaptation, characterized in that, include: S1. Acquire 3D laser point cloud data of the power transmission line corridor; S2. Construct a KD-Tree spatial index structure based on the three-dimensional laser point cloud data; S3. Use the KD-Tree spatial index to extract and classify local geometric features of the point cloud, and at least identify the power line point set and the vegetation point set; S4. Perform geometric consistency filtering on the classified set of wire points to exclude misidentified points caused by the intersection of utility poles and wires. This includes: for each target point in the set of wire points, query its preset number K nearest neighbor wire points, calculate the angle between the main direction vector of the target point and the main direction vector of each nearest neighbor point, count the number of nearest neighbor points whose angle is less than the angle threshold, and judge and filter the target point based on whether the ratio of the counted number to K is lower than the consistency ratio threshold. S5. Based on the filtered set of power line points and the set of vegetation points, calculate the spatial distance between the power line points and the nearest vegetation points, and determine whether there is a potential risk of tree encroachment based on the safe distance threshold.
7. The method according to claim 7, characterized in that, Step S3 includes: S31. For each query point in the point cloud, perform a K-nearest neighbor query using the KD-Tree to obtain its neighborhood point set; S32. Perform principal component analysis on the neighborhood point set to calculate the linearity, flatness, sphericity, and principal direction vector; S33. Based on the linearity, flatness, sphericity, principal direction vector, and point height information, the point cloud is classified into at least one of the following using preset classification rules: power line points, power pole points, vegetation points, building points, and ground points.
8. The method according to claim 8, characterized in that, The classification rules in step S33 include: If the linearity of a point is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is greater than the first angle threshold, and its height is greater than the first height threshold, then it is classified as a wire point. If a point's linearity is greater than the first linearity threshold, the angle between its principal direction vector and the vertical direction is less than the second angle threshold, and its height is greater than the second height threshold, then it is classified as a utility pole point.
9. The method according to claim 7, characterized in that, Before performing directional consistency filtering, step S4 also includes a pole area exclusion step: calculating the distance between a point in the set of power line points and the nearest power pole point. If the distance is less than a preset pole influence threshold, the point is excluded from the set of power line points.
10. The method according to claim 7, characterized in that, After performing directional consistency filtering, step S4 also includes a tower false detection elimination step: for candidate point pairs that are initially determined to be encroachments based on distance calculation, it is detected whether the vegetation points are located near the utility poles. If so, the candidate point pair is determined to be a false detection and eliminated.