CN114821327B - Method and system for extracting and processing characteristics of power line and tower and storage medium - Google Patents

Method and system for extracting and processing characteristics of power line and tower and storage medium Download PDF

Info

Publication number
CN114821327B
CN114821327B CN202210475297.1A CN202210475297A CN114821327B CN 114821327 B CN114821327 B CN 114821327B CN 202210475297 A CN202210475297 A CN 202210475297A CN 114821327 B CN114821327 B CN 114821327B
Authority
CN
China
Prior art keywords
point cloud
point
current
tower
power line
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210475297.1A
Other languages
Chinese (zh)
Other versions
CN114821327A (en
Inventor
李志波
其他发明人请求不公开姓名
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Lvtuzhi New Technology Co ltd
Beijing Digital Green Earth Technology Co ltd
Original Assignee
Shenzhen Lvtuzhi New Technology Co ltd
Beijing Digital Green Earth Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenzhen Lvtuzhi New Technology Co ltd, Beijing Digital Green Earth Technology Co ltd filed Critical Shenzhen Lvtuzhi New Technology Co ltd
Priority to CN202210475297.1A priority Critical patent/CN114821327B/en
Publication of CN114821327A publication Critical patent/CN114821327A/en
Application granted granted Critical
Publication of CN114821327B publication Critical patent/CN114821327B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/10Terrestrial scenes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/26Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/70Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/70Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/70Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Multimedia (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Molecular Biology (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Image Analysis (AREA)

Abstract

The embodiment of the invention discloses a method, a system and a storage medium for extracting and processing characteristics of a power line and a tower, wherein the method comprises the steps of constructing a deep learning model in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud incidental category label comprises: power transmission line points, tower points and background points; acquiring input point clouds (namely single-frame point clouds) in real time by using the trained deep learning model, carrying out identification test on the input point clouds by using the deep learning model, and outputting an identification result; the output identification result is the point cloud to which the current point cloud belongs, and is accompanied by a category label and a corresponding probability; by the method, the characteristic extraction processing of the target point cloud can be rapidly realized.

Description

Method and system for extracting and processing characteristics of power line and tower and storage medium
Technical Field
The invention relates to the technical field of power transmission line inspection, in particular to a method and a system for extracting and processing characteristics of a power line and a tower and a storage medium.
Background
The power transmission line inspection is a basic measure for ensuring the stable operation of a power system, and the development of a stable and reliable power transmission line inspection system has great significance. At present, a large amount of data point clouds can be obtained after a routing inspection process adopting onboard LiDAR, however, how to rapidly realize the feature extraction processing of target point clouds is a technical problem needed by technicians in the field.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method, a system, and a storage medium for extracting and processing characteristics of a power line and a tower, so as to solve the problems indicated in the background art.
An embodiment of the present invention provides a method for extracting and processing characteristics of a power line and a tower, including:
step S100, a deep learning model is constructed in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud accompanying category labels include: power transmission line points, tower points and background points;
s200, acquiring input point clouds in real time by using the trained deep learning model, carrying out identification test on the input point clouds by using the deep learning model, and outputting an identification result;
and outputting the identification result which is the point cloud attached with the category label and the corresponding probability to which the current point cloud belongs.
Preferably, as one possible embodiment; the pre-constructed deep learning model comprises the following steps:
step S101, manually acquiring point cloud data on power transmission lines in different scenes in a main network and a distribution network in advance by adopting a UNet convolutional neural network, and then labeling a power line, a pole tower and a background in the point cloud data to construct an initial training set;
step S102, acquiring an actual labeling result of point cloud data of an initial training set; outputting a classification result of the point cloud data in the initial training set;
and S103, continuously carrying out classification error back propagation iteration on the classification result and the actual labeling result of the initial training set, carrying out intensive training to obtain a final target training set, and carrying out training according to the target training set to finally obtain a deep learning model.
Preferably, as one possible embodiment; in the step S200, acquiring the input point cloud in real time by using the trained deep learning model, performing recognition test on the input point cloud by using the deep learning model, and outputting a recognition result, which specifically includes:
step S201, acquiring input point cloud in real time, extracting target features of the currently input point cloud, outputting the probability that the current point cloud belongs to a transmission line point as a, outputting the probability that the current point cloud belongs to a tower point as b and outputting the probability that the current point cloud belongs to a background point as c by the deep learning model according to the extracted target features, wherein the requirements that a + b + c =1 are met;
step S202, constructing a kdtree for the currently input point cloud, wherein the kdtree is used for performing neighbor neighborhood retrieval on the feature vector of the current point cloud and quickly indexing other point clouds approximate to the feature vector of the current point cloud in the current space, and taking other point cloud sets of the current point cloud and the feature vector as target point clouds;
step S203, presetting a target point cloud comprehensive probability threshold value p, carrying out probability judgment on the current point cloud, and if the current point cloud is a point which meets the condition that the probability a that the current point cloud belongs to a power transmission line point and the probability b that the current point cloud belongs to a tower point are greater than p, judging that the current point cloud is a target point of a required power line point and the tower point;
step S204, presetting a lowest standard probability threshold value of a power line point as Kb and a lowest standard probability threshold value of a pole tower point as Ka; carrying out probability judgment on the current point cloud, and judging the current point cloud to be a power line point if the current point cloud meets the condition that the probability a of the current point cloud belonging to the power line point is greater than Kb; if the current point cloud meets the condition that the probability b of the current point cloud belonging to the pole tower point is greater than Ka, judging that the current point cloud is the pole tower point;
step S205, an uncertain point cloud set is constructed for the point clouds which do not meet the condition of being more than p, the point clouds which do not meet the condition of being more than Kb and the point clouds which do not meet the condition of being more than Ka in the current point clouds.
Preferably, as one possible embodiment; in the execution process of the step S200, further identifying the current point cloud in the uncertain point cloud set;
step S206, the kdtree in the step S204 is called, and all point clouds with the current point cloud neighborhood radius within d in the uncertain point cloud set are obtained;
step S207, calculating the point cloud density in the point cloud which is similar to all the feature vectors in the current point cloud neighborhood radius d, namely the point cloud density of the target point cloud;
step S208, obtaining a power line point density standard threshold pl, comparing and judging the power line point density standard threshold pl according to the point cloud density of the target point cloud with the current point cloud neighborhood radius within d, and if the difference between the point cloud density of the target point cloud with the current point cloud neighborhood radius within d and the pl is judged to be within the standard range, judging that the target point cloud in the current target point cloud set is the power line point.
Preferably, as one possible embodiment; after the step S207, after the current point cloud is identified, the method further includes storing an identification result of the current point cloud.
Preferably, as one possible embodiment; storing the recognition result of the current point cloud, comprising the following operations:
step S209, storing the classified identification result of the current point cloud into the kdtree, and outputting the identification result of the current next frame point cloud when the next frame point cloud is obtained; the kdtree is used for the classified identification result of the current point cloud;
step S210, when the identification result of the current next frame of point cloud is output, identifying the classification result of the point cloud retrieved from the adjacent neighborhood with approximate vector characteristics, and if the point cloud in the adjacent neighborhood is identified as a tower point, identifying the current next frame of point cloud as the tower point; and if the point cloud in the neighborhood is identified as the background point, the current next frame of point cloud is determined as the background point.
The invention provides a processing system comprising: the device comprises a construction unit and an identification processing unit;
the building unit is used for building a deep learning model in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud accompanying category labels include: power transmission line points, tower points and background points;
the recognition processing unit is used for acquiring the input point cloud in real time by using the trained deep learning model, and the deep learning model carries out recognition test on the input point cloud and outputs a recognition result; and outputting the identification result which is the point cloud attached with the category label and the corresponding probability to which the current point cloud belongs.
The present invention provides a computer-readable storage medium comprising: which stores a computer program that, when executed, implements a method of feature extraction processing from the one power line and tower.
Compared with the prior art, the invention has the following technical effects:
the method provided by the embodiment of the invention; a deep learning model is constructed in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud accompanying category labels include: power transmission line points, tower points and background points; acquiring input point clouds (namely single-frame point clouds) in real time by using the trained deep learning model, carrying out identification test on the input point clouds by using the deep learning model, and outputting an identification result; the output identification result is the point cloud to which the current point cloud belongs, and is accompanied by a category label and a corresponding probability; by the method, the characteristic extraction processing of the target point cloud can be rapidly realized.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings required to be used in the embodiments will be briefly described below, and it should be understood that the following drawings only illustrate some embodiments of the present invention, and therefore should not be considered as limiting the scope of the present invention. Like components are numbered similarly in the various figures.
Fig. 1 shows a main flowchart of a method for extracting and processing characteristics of a power line and a tower according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a processing system according to a second embodiment of the present invention.
Description of the main element symbols: a building unit 10 and an identification processing unit 20.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the following description, suffixes such as "module", "component", or "unit" used to denote elements are used only for the convenience of description of the present application, and have no specific meaning in themselves. Thus, "module" and "component" may be used in a mixture.
Example one
Referring to fig. 1, a method for extracting and processing characteristics of a power line and a tower provided in an embodiment of the present invention includes:
step S100, a deep learning model is constructed in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud incidental category label comprises: power transmission line points, tower points and background points;
step S200, acquiring input point clouds (namely single-frame point clouds) in real time by using the trained deep learning model, carrying out identification test on the input point clouds by using the deep learning model, and outputting identification results;
and outputting the identification result which is the point cloud attached with the category label and the corresponding probability to which the current point cloud belongs.
It should be noted that (i.e. the probability and the label of the category (power line, tower, background point) to which the current point cloud belongs), because the model may have a certain misjudgment, after the result is calculated by the model, the statistical analysis is additionally performed on the output result of the model, assuming that each point is classified and then divided into a power system, the categories of the tower and the background are a, b, and c, respectively, and satisfy a + b + c =1, and the calculation process is as follows:
preferably, as one possible embodiment; the pre-constructed deep learning model comprises the following steps:
step S101, manually acquiring point cloud data on power transmission lines in different scenes in a main network and a distribution network in advance by adopting a UNet convolutional neural network, and then labeling a power line, a pole tower and a background in the point cloud data to construct an initial training set;
step S102, acquiring an actual labeling result of point cloud data of an initial training set; outputting a classification result of the point cloud data in the initial training set;
and S103, continuously carrying out classification error back propagation iteration on the classification result and the actual labeling result of the initial training set, carrying out intensive training to obtain a final target training set, and carrying out training according to the target training set to finally obtain a deep learning model.
It should be noted that continuously performing classification error back propagation iteration (belonging to the known technology and not described herein) on the classification result and the actual labeling result of the initial training set can strengthen the training to obtain the final target training set; and then training according to the target training set to finally obtain the deep learning model.
Preferably, as one possible embodiment; in the execution process of step S200, the trained deep learning model is used to obtain the input point cloud (i.e. single frame point cloud) in real time, and the deep learning model performs recognition test on the input point cloud and outputs a recognition result, which specifically includes:
step S201, acquiring input point cloud in real time, extracting target features of the currently input point cloud, outputting the probability that the current point cloud belongs to a transmission line point as a, outputting the probability that the current point cloud belongs to a tower point as b and outputting the probability that the current point cloud belongs to a background point as c by the deep learning model according to the extracted target features, wherein the requirements that a + b + c =1 are met;
step S202, constructing a kdtree for the currently input point cloud, wherein the kdtree is used for performing neighbor neighborhood retrieval on the feature vector of the current point cloud and quickly indexing other point clouds approximate to the feature vector of the current point cloud in the current space, and taking other point cloud sets of the current point cloud and the feature vector as target point clouds;
step S203, presetting a target point cloud comprehensive probability threshold value p, carrying out probability judgment on the current point cloud, and if the current point cloud is a point which meets the condition that the probability a that the current point cloud belongs to a power transmission line point and the probability b > p that the current point cloud belongs to a tower point, judging that the current point cloud is a target point of a required power line point and the tower point (at the moment, the background point is filtered, which is equivalent to the condition that the current point cloud is not a background point);
step S204, presetting a lowest standard probability threshold value of a power line point as Kb and a lowest standard probability threshold value of a pole tower point as Ka; carrying out probability judgment on the current point cloud, and judging the current point cloud to be a power line point if the current point cloud meets the condition that the probability a of the current point cloud belonging to the power line point is greater than Kb; if the current point cloud meets the condition that the probability b of the current point cloud belonging to the tower point is greater than Ka, judging that the current point cloud is the tower point;
step S205, an uncertain point cloud set is constructed for the point clouds which do not meet the condition of being more than p, the point clouds which do not meet the condition of being more than Kb and the point clouds which do not meet the condition of being more than Ka in the current point clouds.
Preferably, as one possible embodiment; in the execution process of the step S200, further identifying the current point cloud in the uncertain point cloud set;
step S206, the kdtree in the step S204 is called, and all point clouds with the current point cloud neighborhood radius within d in the uncertain point cloud set are obtained;
step S207, calculating the point cloud density in the point cloud with the approximate feature vectors of all the feature vectors with the current point cloud neighborhood radius of d, namely the point cloud density of the target point cloud;
step S208, obtaining a power line point density standard threshold pl, comparing and judging the power line point density standard threshold pl according to the point cloud density of the target point cloud with the current point cloud neighborhood radius within d, and if the difference between the point cloud density of the target point cloud with the current point cloud neighborhood radius within d and the pl is judged to be within the standard range, judging that the target point cloud in the current target point cloud set is the power line point.
Preferably, as one possible embodiment; after the step S207, after the current point cloud is identified, the method further includes storing an identification result of the current point cloud.
Preferably, as one possible embodiment; storing the recognition result of the current point cloud, comprising the following operations:
step S209, storing the classified identification result of the current point cloud into the kdtree, and outputting the identification result of the current next frame point cloud when the next frame point cloud is obtained; the kdtree is used for the classified identification result of the current point cloud;
step S210, when the identification result of the current next frame of point cloud is output, identifying the classification result of the point cloud retrieved from the neighbor neighborhood with approximate vector characteristics, and if the point cloud in the neighbor neighborhood is identified as a pole tower point, identifying the current next frame of point cloud as the pole tower point; and if the point cloud in the neighborhood is identified as the background point, the current next frame of point cloud is identified as the background point.
Note that, at the remaining uncertainty points in step S205, the density of points within the domain radius d is searched and calculated in step S202, and if the power line point density threshold is set to pl, the point having a point density smaller than pl is the power line point. The density of the tower points is difficult to distinguish from the density of the background points, therefore, in the previous frame of results, whether points in a certain range in the neighborhood are tower points or not is searched, if the points are the tower points, the tower points are set, and if the points are the background points, the tower points are set.
Example two
An embodiment of the present invention provides a node management test system, including: a construction unit 10, an identification processing unit 20;
the building unit 10 is used for building a deep learning model in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud accompanying category labels include: power transmission line points, tower points and background points;
the recognition processing unit 20 is configured to obtain an input point cloud in real time by using the trained deep learning model, perform recognition test on the input point cloud by using the deep learning model, and output a recognition result; and outputting the identification result which is the point cloud attached with the category label and the corresponding probability to which the current point cloud belongs.
It is to be noted that; collecting original point cloud data, and labeling and classifying the data; preprocessing the classified data, including point cloud data normalization, space position quantification, voxelization, cutting and other operations; inputting the preprocessed point cloud into a network for feature learning and extraction; inputting the learned characteristics into a full-connection layer and extracting the probability of the category to which each point belongs to achieve a classification effect; calculating the back propagation iteration of the classification error according to the classification result and the actual labeling result to realize a training network;
the present invention provides a computer-readable storage medium comprising: which stores a computer program that, when executed, implements a method of performing feature extraction processing from the power line and tower.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention.

Claims (4)

1. A method for extracting and processing characteristics of a power line and a tower is characterized by comprising the following steps:
step S100, a deep learning model is constructed in advance; the deep learning model is used for dividing the point cloud into different categories and attaching category labels to the point cloud of each category; the point cloud accompanying category labels include: power transmission line points, tower points and background points;
s200, acquiring input point clouds in real time by using the trained deep learning model, carrying out identification test on the input point clouds by using the deep learning model, and outputting an identification result;
the output identification result is the point cloud to which the current point cloud belongs, and is accompanied by a category label and a corresponding probability;
the pre-constructed deep learning model comprises the following steps:
step S101, manually acquiring point cloud data on power transmission lines in different scenes in a main network and a distribution network in advance by adopting a UNet convolutional neural network, and then labeling a power line, a pole tower and a background in the point cloud data to construct an initial training set;
step S102, acquiring an actual marking result of point cloud data of an initial training set; outputting a classification result of the point cloud data in the initial training set;
step S103, continuously carrying out classification error back propagation iteration on the classification result of the initial training set and the actual labeling result, carrying out intensive training to obtain a final target training set, and carrying out training according to the target training set to finally obtain a deep learning model;
in the step S200, acquiring the input point cloud in real time by using the trained deep learning model, performing recognition test on the input point cloud by using the deep learning model, and outputting a recognition result, which specifically includes:
step S201, acquiring input point cloud in real time, extracting target features of the currently input point cloud, outputting the probability that the current point cloud belongs to a transmission line point as a, outputting the probability that the current point cloud belongs to a tower point as b and outputting the probability that the current point cloud belongs to a background point as c by the deep learning model according to the extracted target features, wherein a + b + c =1 is satisfied;
step S202, constructing a kdtree for the currently input point cloud, wherein the kdtree is used for performing neighbor neighborhood retrieval on the feature vector of the current point cloud and quickly indexing other point clouds approximate to the feature vector of the current point cloud in the current space, and taking other point cloud sets of the current point cloud and the feature vector as target point clouds;
step S203, presetting a target point cloud comprehensive probability threshold value p, carrying out probability judgment on the current point cloud, and if the current point cloud is a point which meets the condition that the probability a that the current point cloud belongs to a power transmission line point and the probability b that the current point cloud belongs to a tower point are greater than p, judging that the current point cloud is a target point of a required power line point and the tower point;
step S204, presetting a lowest standard probability threshold value of a power line point as Kb and a lowest standard probability threshold value of a pole tower point as Ka; carrying out probability judgment on the current point cloud, and judging the current point cloud to be a power line point if the current point cloud meets the condition that the probability a of the current point cloud belonging to the power line point is greater than Kb; if the current point cloud meets the condition that the probability b of the current point cloud belonging to the tower point is greater than Ka, judging that the current point cloud is the tower point;
step S205, aiming at the point clouds which do not meet the condition of being more than p, the point clouds which do not meet the condition of being more than Kb and the point clouds which do not meet the condition of being more than Ka in the current point clouds as uncertain point clouds, an uncertain point cloud set is constructed;
in the execution process of the step S200, further identifying the current point cloud in the uncertain point cloud set;
step S206, the kdtree in the step S204 is called, and all point clouds with the current point cloud neighborhood radius within d in the uncertain point cloud set are obtained;
step S207, calculating the point cloud density in the point cloud with the approximate feature vectors of all the feature vectors with the current point cloud neighborhood radius of d, namely the point cloud density of the target point cloud;
step S208, obtaining a power line point density standard threshold pl, comparing and judging the power line point density standard threshold pl according to the point cloud density of the target point cloud with the current point cloud neighborhood radius within d, and if the difference between the point cloud density of the target point cloud with the current point cloud neighborhood radius within d and the pl is judged to be within the standard range, judging that the target point cloud in the current target point cloud set is the power line point.
2. The method for extracting and processing the features of the power line and the tower according to claim 1, wherein after the step S207, after the current point cloud is identified, the method further comprises storing the identification result of the current point cloud.
3. The method for extracting and processing the features of the power line and the tower as claimed in claim 2, wherein the step of storing the recognition result of the current point cloud comprises the following steps:
step S209, storing the classified identification result of the current point cloud into the kdtree, and outputting the identification result of the current next frame point cloud when the next frame point cloud is obtained; the kdtree is used for the classified identification result of the current point cloud;
step S210, when the identification result of the current next frame of point cloud is output, identifying the classification result of the point cloud retrieved from the adjacent neighborhood with approximate vector characteristics, and if the point cloud in the adjacent neighborhood is identified as a tower point, identifying the current next frame of point cloud as the tower point; and if the point cloud in the neighborhood is identified as the background point, the current next frame of point cloud is identified as the background point.
4. A computer-readable storage medium, comprising: which stores a computer program which, when executed, implements a method for performing feature extraction processing on a power line and tower according to any one of claims 1 to 3.
CN202210475297.1A 2022-04-29 2022-04-29 Method and system for extracting and processing characteristics of power line and tower and storage medium Active CN114821327B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210475297.1A CN114821327B (en) 2022-04-29 2022-04-29 Method and system for extracting and processing characteristics of power line and tower and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210475297.1A CN114821327B (en) 2022-04-29 2022-04-29 Method and system for extracting and processing characteristics of power line and tower and storage medium

Publications (2)

Publication Number Publication Date
CN114821327A CN114821327A (en) 2022-07-29
CN114821327B true CN114821327B (en) 2022-12-20

Family

ID=82510519

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210475297.1A Active CN114821327B (en) 2022-04-29 2022-04-29 Method and system for extracting and processing characteristics of power line and tower and storage medium

Country Status (1)

Country Link
CN (1) CN114821327B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109063789A (en) * 2018-09-07 2018-12-21 北京数字绿土科技有限公司 The classification method of shaft tower and power line, device and system
CN111145159A (en) * 2019-12-27 2020-05-12 北京数字绿土科技有限公司 Method and device for extracting routing inspection key component points
CN113538264A (en) * 2021-06-30 2021-10-22 深圳大学 Denoising method and device for point cloud data and storage medium

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110414577A (en) * 2019-07-16 2019-11-05 电子科技大学 A kind of laser radar point cloud multiple target Objects recognition method based on deep learning

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109063789A (en) * 2018-09-07 2018-12-21 北京数字绿土科技有限公司 The classification method of shaft tower and power line, device and system
CN111145159A (en) * 2019-12-27 2020-05-12 北京数字绿土科技有限公司 Method and device for extracting routing inspection key component points
CN113538264A (en) * 2021-06-30 2021-10-22 深圳大学 Denoising method and device for point cloud data and storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
输电通道机载LiDAR点云分类方法研究;彭淑雯;《中国优秀硕士学位论文全文数据库基础科学辑》;20220215(第2期);第3章 *

Also Published As

Publication number Publication date
CN114821327A (en) 2022-07-29

Similar Documents

Publication Publication Date Title
CN107276805B (en) Sample prediction method and device based on intrusion detection model and electronic equipment
CN112311780B (en) Method for generating multi-dimensional attack path and attack graph
CN109872162B (en) Wind control classification and identification method and system for processing user complaint information
CN110874550A (en) Data processing method, device, equipment and system
CN110263824B (en) Model training method, device, computing equipment and computer readable storage medium
CN109995611B (en) Traffic classification model establishing and traffic classification method, device, equipment and server
CN110992307A (en) Insulator positioning and identifying method and device based on YOLO
CN112118551A (en) Equipment risk identification method and related equipment
CN114553591B (en) Training method of random forest model, abnormal flow detection method and device
CN111159481B (en) Edge prediction method and device for graph data and terminal equipment
CN111428511A (en) Event detection method and device
CN109993391B (en) Method, device, equipment and medium for dispatching network operation and maintenance task work order
CN114090601B (en) Data screening method, device, equipment and storage medium
CN110739692A (en) probability graph model-based power distribution network structure identification method
CN113223011B (en) Small sample image segmentation method based on guide network and full-connection conditional random field
CN114821327B (en) Method and system for extracting and processing characteristics of power line and tower and storage medium
CN115482436B (en) Training method and device for image screening model and image screening method
CN115225373A (en) Network space security situation expression method and device under incomplete information condition
CN116821087A (en) Power transmission line fault database construction method, device, terminal and storage medium
CN115579069A (en) Construction method and device of scRNA-Seq cell type annotation database and electronic equipment
CN112115996B (en) Image data processing method, device, equipment and storage medium
CN112818972B (en) Method and device for detecting interest point image, electronic equipment and storage medium
CN114330542A (en) Sample mining method and device based on target detection and storage medium
CN112183622B (en) Mobile application bots installation cheating detection method, device, equipment and medium
CN114860903A (en) Event extraction, classification and fusion method oriented to network security field

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant