CN113554595A - Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method - Google Patents
Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method Download PDFInfo
- Publication number
- CN113554595A CN113554595A CN202110687078.5A CN202110687078A CN113554595A CN 113554595 A CN113554595 A CN 113554595A CN 202110687078 A CN202110687078 A CN 202110687078A CN 113554595 A CN113554595 A CN 113554595A
- Authority
- CN
- China
- Prior art keywords
- tower
- point cloud
- tower head
- head
- actual measurement
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10044—Radar image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/44—Morphing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses an unmanned aerial vehicle laser radar point cloud tower head deformation detection device and a method based on standard CAD model matching, which comprises a tower head standard model point cloud module, a tower actual measurement point cloud segmentation module, a tower head point cloud model adaptation module, a tower head point cloud model matching module and a tower head deformation evaluation module, realizes the quick scanning of the geometric parameters of the tower head of the tower and the intelligent evaluation of the inclined parameters by automatically identifying and extracting the tower head of the tower, obviously improves the efficiency and the accuracy of the deformation inspection of the tower head of the line tower by utilizing the advanced three-dimensional laser radar remote sensing technology, reduces the inspection cost and the workload and the risk of workers, realizes the monitoring and early warning of the deformation of the tower head and the risk monitoring of the tower head in the engineering practice sense, improves the operation reliability of a power transmission line, and reduces the major safe and stable operation risk of a power grid, the method has wide application prospect in the fields of line maintenance, public management, disaster prevention and early warning and the like.
Description
Technical Field
The invention relates to the field of laser point cloud monitoring and early warning and risk prevention and control, in particular to a device and a method for detecting deformation of a tower head of a laser radar point cloud tower of an unmanned aerial vehicle.
Background
The inspection of the power transmission line is an important work which needs to be carried out by a power grid operation maintenance management department, and in order to ensure the operation safety of the power line, the line is generally required to be inspected regularly so as to find and eliminate potential safety hazards in time. With the development of the modernization and the urbanization of our country, the structure evolution of the power transmission tower is more and more huge and complex. The transmission line pole tower is used for a long time in a severe outdoor environment, and the pole tower components are subjected to external environments such as strong wind, ice coating,
Deformation, corrosion, bolt loosening and the like are generated to different degrees under the influence of geological disasters and the like. Due to the self gravity, wind load (horizontal load) and hanging wire load at the cross arm, the tower head part of the tower is easy to deform, the tower body can be inclined or even collapsed seriously, and potential safety hazards are brought to power transmission. The existing tower head deformation monitoring means still has a plurality of defects and short plates in practical application, and often has the problems that the manual inspection efficiency is low, the initial tiny change of the tower head is difficult to observe by naked eyes, the long repeated inspection period cannot be monitored in all weather, the intelligent early warning and statistical analysis functions are lacked, and the like. With the increasing number of high-voltage, high-power and long-distance power transmission lines, the geographic environment traversed by the line corridor is more and more complex, and the operation and maintenance of the line corridor are increasingly difficult if the line corridor passes through a large area of reservoir, lake and great mountains. Therefore, a practical technology capable of rapidly monitoring and early warning the tower head deformation state of the tower in a geological unstable area in a large range is urgently needed.
In recent years, unmanned aerial vehicle inspection becomes a line inspection mode which is widely popularized in national network systems, and line inspection efficiency and accuracy are obviously improved. On the basis, the application function of the existing inspection unmanned aerial vehicle operation platform is fully excavated and expanded, the monitoring and early warning of the tower head deformation three-dimensional space state of the tower is quickly and efficiently carried out by matching the unmanned aerial vehicle laser three-dimensional point cloud with the tower head standard CAD model, the technical feasibility is realized, the remarkable economic benefit is brought, and the important popularization value is realized.
Disclosure of Invention
The invention aims to realize the segmentation of tower head point cloud of a tower, the matching of actually measured point cloud of the tower head and a tower head model and the intelligent evaluation of deformation parameters of a tower head of a line tower by using a laser three-dimensional point cloud of an unmanned aerial vehicle and a standard CAD model of the tower head.
In order to achieve the purpose, the invention provides an unmanned aerial vehicle laser radar point cloud tower head deformation detection device, which comprises a tower head standard model point cloud module, a tower actual measurement point cloud segmentation module, a tower head point cloud model adaptation module, a tower head point cloud model matching module and a tower head deformation evaluation module, wherein the tower head standard model point cloud module is used for converting a tower head standard CAD model into tower head model point clouds according to a fixed-distance equal division method; the pole tower actual measurement point cloud segmentation module is used for constructing a unified pole tower conceptual model according to different types of pole towers, and carrying out pole tower actual measurement point cloud segmentation on pole tower actual measurement point cloud obtained by the unmanned aerial vehicle LiDAR to obtain pole tower actual measurement tower head point cloud and pole tower actual measurement tower body point cloud; the tower head point cloud model adaptation module is used for extracting mode characteristics from tower head point cloud actually measured by a tower, judging the type of a tower head of the tower through a mode identification method, and re-determining the height of a tower shoulder according to the type of the tower head so as to realize accurate segmentation of the tower head point cloud of the tower; the tower head point cloud model matching module automatically and initially matches the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and performs ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set; and the tower head deformation evaluation module evaluates the tower head deformation state of the tower by using the rigid body transformation parameters.
A method for detecting deformation of a tower head of a point cloud tower of a laser radar of an unmanned aerial vehicle comprises the following steps:
step 1, setting a point cloud sampling point distance according to the point cloud density of the laser radar of the unmanned aerial vehicle, and converting a tower head standard CAD model into tower head model point clouds of the tower according to a method of equally dividing the tower head standard CAD model by a fixed distance;
step 2, constructing a unified tower conceptual model according to different types of towers, and carrying out tower actual measurement point cloud segmentation on tower actual measurement point cloud obtained by the LiDAR of the unmanned aerial vehicle to obtain tower actual measurement tower head point cloud and tower actual measurement tower body point cloud;
step 3, extracting mode characteristics from the actually measured tower head point cloud of the tower, judging the tower head type of the tower through a mode identification method, re-determining the tower shoulder height according to the tower head type, and realizing accurate segmentation of the tower head point cloud of the tower;
step 4, carrying out automatic initial matching on the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and carrying out ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of the initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set;
and 5, evaluating the deformation state of the tower head of the tower by using the rigid body transformation parameters.
The invention has the beneficial effects that:
1) the invention provides a tower head deformation detection method and a technical process based on matching of unmanned aerial vehicle laser radar point cloud and a standard CAD model, which realize automatic identification and extraction of a tower head, and realize quick scanning of geometric parameters and intelligent evaluation of tilt parameters of a tower head of a line tower.
2) The invention obviously improves the efficiency and the accuracy of the inspection of the tower head deformation of the line tower by utilizing the advanced three-dimensional laser radar remote sensing technology, and reduces the inspection cost and the workload and the risk of workers.
3) The invention provides a technology for detecting the tower head deformation of the line tower, which has higher maneuverability, rapidness and flexibility, and can obtain more detailed and accurate tower head deformation parameters.
Drawings
FIG. 1 is a schematic diagram of the apparatus of the present invention;
FIG. 2 is an elevation histogram of a tower point cloud;
FIG. 3 is a schematic view of a point cloud projection based on a three-view feature extraction method;
FIG. 4 is a connection diagram of one-dimensional projection line segments based on a three-view feature extraction method;
FIG. 5 is a flow chart of the method of the present invention.
Detailed Description
The invention is described in further detail below with reference to the following figures and specific examples:
the invention relates to an unmanned aerial vehicle laser radar point cloud tower head deformation detection device, which comprises a tower head standard model point cloud module 1, a tower actual measurement point cloud segmentation module 2, a tower head point cloud model adaptation module 3, a tower head point cloud model matching module 4 and a tower head deformation evaluation module 5, wherein the tower head standard model point cloud module 1 is used for converting a tower head standard CAD model into tower head model point clouds according to a fixed-distance equal division method; the pole tower actual measurement point cloud segmentation module 2 is used for constructing a unified pole tower conceptual model according to different types of pole towers, and carrying out pole tower actual measurement point cloud segmentation on pole tower actual measurement point cloud obtained by the unmanned aerial vehicle LiDAR to obtain pole tower actual measurement tower head point cloud and pole tower actual measurement tower body point cloud; the tower head point cloud model adaptation module 3 is used for extracting mode characteristics from tower head point cloud actually measured by a tower, judging the type of a tower head of the tower by a mode identification method, and re-determining the height of a tower shoulder according to the type of the tower head to realize accurate segmentation of the tower head point cloud of the tower; the tower head point cloud model matching module 4 automatically and initially matches the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and performs ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set; and the tower head deformation evaluation module 5 evaluates the deformation state of the tower head by using the rigid body transformation parameters.
A method for detecting the deformation of a tower head of a point cloud tower of a laser radar of an unmanned aerial vehicle is shown in figure 5 and comprises the following steps:
step 1, setting a point cloud sampling point distance according to the point cloud density of the laser radar of the unmanned aerial vehicle, and converting a tower head standard CAD model into tower head model point clouds of the tower according to a method of equally dividing the tower head standard CAD model by a fixed distance;
step 2, constructing a unified tower conceptual model according to different types of towers, and carrying out tower actual measurement point cloud segmentation on tower actual measurement point cloud obtained by the LiDAR of the unmanned aerial vehicle to obtain tower actual measurement tower head point cloud and tower actual measurement tower body point cloud;
step 3, extracting mode characteristics from the actually measured tower head point cloud of the tower, judging the tower head type of the tower through a mode identification method, re-determining the tower shoulder height according to the tower head type, and realizing accurate segmentation of the tower head point cloud of the tower;
step 4, carrying out automatic initial matching on the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and carrying out ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of the initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set;
and 5, evaluating the deformation state of the tower head of the tower by using the rigid body transformation parameters.
In the above technical solution, the specific implementation method of step 2 is:
step 2.1, because the types of the towers are numerous, in order to facilitate subsequent three-dimensional reconstruction, the four main structural elevations of the tower foot elevation, the tower body elevation, the tower shoulder elevation and the tower head elevation of the towers of different types are utilized to divide the tower point cloud into three parts of the tower foot, the tower body and the tower head, and the towers of different types are constructed into a unified tower concept model;
step 2.2, generating an elevation histogram of the tower actual measurement point cloud according to the structural characteristics that a characteristic plane exists at intervals of elevations of a tower body, extracting a local extreme value of the elevation histogram of the tower actual measurement point cloud as shown in fig. 2, and eliminating an extreme value of a non-characteristic elevation through a constraint condition that the interval between characteristic elevations is greater than a certain threshold value to obtain the characteristic elevation and the corresponding characteristic plane in the tower actual measurement point cloud;
and 2.3, extracting the tower shoulder elevation at the joint of the tower head and the tower body by judging the number of points corresponding to each characteristic elevation of the tower actual measurement point cloud, wherein the characteristic elevation with the largest number of points corresponds to the tower shoulder elevation at the joint of the tower head and the tower body, so as to realize the tower actual measurement point cloud segmentation and obtain the tower actual measurement tower head point cloud and the tower actual measurement tower body point cloud.
In the above technical solution, the specific implementation method of step 3 is:
3.1, extracting the characteristic elevation corresponding to the local extreme value in the actually measured point cloud histogram of the tower by using a point cloud elevation histogram-based characteristic extraction method, and obtaining 3 types of model characteristics by using the difference of the characteristic elevations as mode characteristics, wherein the tower structures of the same type are the same and the histogram also shows some similar characteristics;
step 3.2, in the front view of the tower, the tower type information is abundant: on the characteristic elevation of the tower head range of the tower, a plurality of horizontal straight line segments exist, and the category information of the tower head can be judged by utilizing the lengths and the elevation difference of the straight line segments; two characteristic lengths of the tower are defined as a range length and an actual length, wherein the range length represents the span of the tower steel frame, and the actual length represents the actual length of the tower steel frame in the horizontal direction. Therefore, the tower three-view-based feature extraction method rotates the tower actual measurement point cloud according to the main direction to obtain a front view of the tower, and obtains 6 types of model features by taking the feature length and the difference between elevations corresponding to the feature length as mode features;
3.3, extracting features of point cloud samples of different types of tower models, training by using an SVM algorithm to obtain an initial SVM classifier, and training and updating the SVM classifier by using tower data of known types, so that the classifier is more stable and complete;
step 3.4, matching the towers to be matched by using the SVM classifier updated and trained in the step 3.3, and determining the type of the tower head; because the tower heads of the same type are completely consistent in specification and size theoretically, the tower shoulder height is determined again according to the type of the tower heads, and accurate segmentation of the tower head of the tower actual measurement point cloud is achieved.
In the above technical solution, the specific method for obtaining the 3 types of model features in step 3.1 is as follows: extracting the characteristic elevation corresponding to the local extreme value in the tower measured point cloud histogram by adopting a characteristic extraction method based on the tower point cloud elevation histogram, including the characteristic elevation H corresponding to the maximum value in the elevation histograms、HsFirst characteristic elevation above H1、HsThe feature height H with the largest number of points above2Characteristic elevation H nearest to the top of the tower3Using the difference between the feature elevations as the mode feature, 3 types of model features [ H ] can be obtained1-Hs,H2-Hs,H3-Hs]。
In the above technical solution, the specific method for obtaining the 6-type model features in step 3.1 is as follows: firstly, the actually measured point cloud of the pole tower is pressed according to the main partRotating to obtain a front view of the tower model point cloud, then performing one-dimensional projection and seed point growth on the tower actual measurement point cloud on each characteristic elevation H to obtain 6 types of model characteristic lengths, and calculating the longest range length Ls1And corresponding elevation
The next longest range length Ls2And corresponding elevation
And calculating the elevation difference between the twos，
Calculating the longest characteristic length Le1And corresponding elevation
Characteristic length L of next longeste2And corresponding elevation
And calculating the elevation difference between the twoe，
Finally obtaining 6 types of model characteristics Ls1,Ls2,Le1,Le2,ΔHs,ΔHe]。
In the above technical scheme, the specific implementation method for performing one-dimensional projection and seed point growth on the tower actual measurement point cloud at each characteristic elevation H comprises:
step 3.1.1, performing one-dimensional projection on the point clouds with the characteristic elevation H within a certain elevation range Z belonging to [ H-delta H, H + delta H ], performing grid division on one-dimensional projection X values of all the point clouds, setting the width of a projection grid, and taking 1 as the value of the projection grid if point cloud pixel points exist in the projection grid, or taking 0 as the value of the projection grid, as shown in FIG. 3;
step 3.1.2, searching a leftmost non-zero projection grid with a projection grid value of 1 from an initial position of the one-dimensional projection grid, defining the leftmost non-zero projection grid as an initial seed point, wherein the initial seed point is a segment starting point of a discrete segment, searching an adjacent projection grid from the initial seed point to the right side, if the adjacent projection grid value is 1, defining the adjacent projection grid as a new seed point, continuing to search to the right side of the new seed point until the adjacent projection grid value on the right side of the new seed point is 0, defining the current new seed point as a segment end point of the discrete segment, classifying all point cloud pixel points between the segment starting point and the segment end point into one class, and connecting the point cloud pixel points into the discrete segment;
3.1.3, taking the adjacent projection grid on the right side of the line segment end point as the initial position of the one-dimensional projection grid, repeating the step 3.1.2 until all point clouds are classified, and generating corresponding discrete line segments;
and 3.1.4, acquiring the adjacent distances of all the discrete line segments, and combining the line segments with the adjacent distances smaller than a set threshold into one line segment, as shown in fig. 4.
In the technical scheme, because the measurement coordinate system is inconsistent with the CAD model design coordinate system, the difference between the position and the direction of the tower head and the direction needs to be corrected before the deformation state of the tower head is evaluated.
In the above technical solution, the objective of the point cloud main direction pasting method is to reduce rotation and translation dislocation between point clouds, so that the exact matching is inconsistent and tends to wrong direction, and the specific method thereof is as follows:
the method comprises the steps of obtaining a main direction of the tower head actual measurement point cloud and two secondary directions perpendicular to the main direction by calculating feature vectors of all points in the tower head actual measurement point cloud, establishing a reference coordinate system with the center of gravity of the tower head actual measurement point cloud as an original point and the main direction and the secondary directions of the point cloud as coordinate axes, obtaining the main direction of the tower head model point cloud and the two secondary directions perpendicular to the main direction by calculating feature vectors of all points in the tower head model point cloud, establishing a reference coordinate system with the center of gravity of the tower head model point cloud as the original point and the main direction and the secondary directions of the point cloud as coordinate axes, and then adjusting the two reference coordinate systems to be consistent.
In the above technical solution, the ICP matching is performed to precisely match the measured point cloud of the tower head and the model point cloud of the tower head in the best fit state on the basis of the initial matching, and the specific method is as follows:
step 4.1, searching a point set closest corresponding point of a tower head model point cloud and a tower head actual measurement point cloud to be evaluated by a step-by-step iteration method;
step 4.2, calculating the optimal rigid body transformation parameters between the two point sets to enable the optimal rigid body transformation parameters to enable the sum of squares of distances of all corresponding point pairs to be the minimum, wherein the rigid body transformation parameters comprise rotation matrixes and translation parameters;
and 4.3, acting the optimal rigid body transformation parameters obtained in the step 4.2 on the tower head actual measurement point cloud after rigid body transformation, continuously iterating the step 4.1 and the step 4.2 until the error measure meets the given convergence precision or reaches the maximum iteration times, and finally obtaining the rigid body transformation parameters between the two point sets.
In the above technical solution, the specific implementation method of step 5 is:
step 5.1, obtaining the inclination angle of the tower head of the tower to be evaluated through a rotation matrix in the rigid body transformation parameters, and if the inclination angle is smaller than a certain threshold value, judging that the tower head of the tower to be evaluated does not have inclination within an allowable range; if the inclination angle is larger than the threshold value, determining that the tower head of the tower to be evaluated is inclined;
step 5.2, calculating the distance between the actual measurement point cloud of the tower head of the matched tower and the corresponding matching point in the tower head model of the tower, and if the distances are smaller than a certain threshold value, judging that no deformation exists in the tower head of the tower to be evaluated within an allowable range; and if the distance between the partial actual measurement point and the theoretical point is larger than the threshold value, judging that the tower head of the tower to be evaluated has deformation, wherein the deformation occurrence part is the part with the distance larger than the threshold value.
The invention adopts the three-dimensional laser radar remote sensing technology, remarkably improves the rapid and intelligent mass sensing level of the geometric state of the tower head of the transmission line tower, realizes the deformation monitoring and early warning and risk monitoring of the tower head of the transmission line tower in the engineering practice sense, improves the operation reliability of the transmission line, reduces the major safe and stable operation risk of a power grid, and has wide application prospect in the fields of line maintenance, public management, disaster prevention and early warning and the like.
Details not described in this specification are within the skill of the art that are well known to those skilled in the art.
Claims (9)
1. Unmanned aerial vehicle laser radar point cloud shaft tower head deformation detection device, its characterized in that: the system comprises a tower head standard model point cloud module (1), a tower actual measurement point cloud segmentation module (2), a tower head point cloud model adaptation module (3), a tower head point cloud model matching module (4) and a tower head deformation evaluation module (5), wherein the tower head standard model point cloud module (1) is used for converting a tower head standard CAD model into tower head model point clouds according to a fixed-distance equal division method; the pole tower actual measurement point cloud segmentation module (2) is used for constructing a unified pole tower conceptual model according to different types of pole towers, and carrying out pole tower actual measurement point cloud segmentation on pole tower actual measurement point cloud obtained by the LiDAR of the unmanned aerial vehicle to obtain pole tower actual measurement tower head point cloud and pole tower actual measurement tower body point cloud; the tower head point cloud model adaptation module (3) is used for extracting mode characteristics from tower head point cloud actually measured by a tower, judging the type of a tower head of the tower through a mode identification method, and re-determining the height of a tower shoulder according to the type of the tower head to realize accurate segmentation of the tower head point cloud of the tower; the tower head point cloud model matching module (4) automatically and initially matches the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and performs ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set; and the tower head deformation evaluation module (5) evaluates the tower head deformation state of the tower by using the rigid body transformation parameters.
2. A method for detecting deformation of a point cloud tower head of an unmanned aerial vehicle laser radar is characterized by comprising the following steps: it comprises the following steps:
step 1, setting a point cloud sampling point distance according to the point cloud density of the laser radar of the unmanned aerial vehicle, and converting a tower head standard CAD model into tower head point cloud of a tower according to a method of equally dividing the tower head standard CAD model by a fixed distance;
step 2, constructing a unified tower conceptual model according to different types of towers, and carrying out tower actual measurement point cloud segmentation on tower actual measurement point cloud obtained by the LiDAR of the unmanned aerial vehicle to obtain tower actual measurement tower head point cloud and tower actual measurement tower body point cloud;
step 3, extracting mode characteristics from the actually measured tower head point cloud of the tower, judging the tower head type of the tower through a mode identification method, re-determining the tower shoulder height according to the tower head type, and realizing accurate segmentation of the tower head point cloud of the tower;
step 4, carrying out automatic initial matching on the tower head actual measurement point cloud and the tower head model point cloud by using a point cloud main direction pasting method, and carrying out ICP (inductively coupled plasma) matching on the tower head actual measurement point cloud to be evaluated by using the tower head model point cloud on the basis of the initial registration to obtain rigid body transformation parameters between the tower head actual measurement point cloud and the tower head model point cloud set;
and 5, evaluating the deformation state of the tower head of the tower by using the rigid body transformation parameters.
3. The unmanned aerial vehicle laser radar point cloud tower head deformation detection device of claim 2, characterized in that: the specific implementation method of the step 2 comprises the following steps:
step 2.1, dividing the tower point cloud into a tower foot part, a tower body part and a tower head part by utilizing four main structure elevations of tower foot elevations, tower body elevations, tower shoulder elevations and tower head elevations of different types of towers, and constructing the different types of towers into a unified tower concept model;
step 2.2, generating an elevation histogram of the tower actual measurement point cloud, extracting a local extreme value of the elevation histogram of the tower actual measurement point cloud, and eliminating an extreme value of a non-characteristic elevation through a constraint condition that the interval between characteristic elevations is larger than a certain threshold value to obtain the characteristic elevation and a corresponding characteristic plane in the tower actual measurement point cloud;
and 2.3, extracting the tower shoulder elevation at the joint of the tower head and the tower body by judging the number of points corresponding to each characteristic elevation of the tower actual measurement point cloud, wherein the characteristic elevation with the largest number of points corresponds to the tower shoulder elevation at the joint of the tower head and the tower body, so as to realize the tower actual measurement point cloud segmentation and obtain the tower actual measurement tower head point cloud and the tower actual measurement tower body point cloud.
4. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 2, characterized in that: the specific implementation method of the step 3 is as follows:
step 3.1, extracting the characteristic elevation corresponding to the local extreme value in the tower measured point cloud histogram based on the point cloud elevation histogram feature extraction method, and obtaining 3 types of model features by taking the difference of the characteristic elevation as a mode feature;
step 3.2, rotating the measured point cloud of the tower according to the main direction based on a three-view feature extraction method to obtain a front view of the tower, and obtaining 6 types of model features by taking the feature length and the difference between elevations corresponding to the feature length as mode features;
3.3, extracting features of point cloud samples of different types of tower models, training by using an SVM algorithm to obtain an initial SVM classifier, and training and updating the SVM classifier by using tower data of known types;
step 3.4, matching the towers to be matched by using the SVM classifier updated and trained in the step 3.3, and determining the type of the tower head; and re-determining the tower shoulder height according to the type of the tower head, and realizing accurate segmentation of the tower head of the tower actual measurement point cloud.
5. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 4, characterized in that:
the specific method for acquiring the 3 types of model features in the step 3.1 comprises the following steps: extracting the characteristic elevation corresponding to the local extreme value in the tower measured point cloud histogram by adopting a characteristic extraction method based on the tower point cloud elevation histogram, including the characteristic elevation H corresponding to the maximum value in the elevation histograms、HsFirst characteristic elevation above H1、HsThe feature height H with the largest number of points above2Characteristic elevation H nearest to the top of the tower3Using the difference between the feature elevations as the mode feature, 3 types of model features [ H ] can be obtained1-Hs,H2-Hs,H3-Hs]。
6. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 4, characterized in that:
the specific method for acquiring the 6 types of model characteristics in the step 3.1 comprises the following steps: firstly, rotating the tower actual measurement point cloud according to the main direction to obtain a front view of the tower actual measurement point cloud, then performing one-dimensional projection and seed point growth on the tower actual measurement point cloud on each characteristic elevation H to obtain 6 types of model characteristic lengths, and calculating the longest range length Ls1And corresponding elevation
The next longest range length Ls2And corresponding elevation
And calculating the elevation difference between the twos，
Calculating the longest characteristic length Le1And corresponding elevation
Characteristic length L of next longeste2And corresponding elevation
And calculating the elevation difference between the twoe，
Finally obtaining 6 types of model characteristics Ls1,Ls2,Le1,Le2,ΔHs,ΔHe]；
The specific implementation method for performing one-dimensional projection and seed point growth on the tower measured point cloud at each characteristic elevation H comprises the following steps:
step 3.1.1, performing one-dimensional projection on the point clouds with the characteristic elevation H within a certain elevation range Z belonging to [ H-delta H, H + delta H ], performing grid division on one-dimensional projection X values of all the point clouds, setting the width of a projection grid, and if point cloud pixel points exist in the projection grid, taking 1 as the projection grid value, otherwise, taking 0 as the projection grid value;
step 3.1.2, searching a leftmost non-zero projection grid with a projection grid value of 1 from an initial position of the one-dimensional projection grid, defining the leftmost non-zero projection grid as an initial seed point, wherein the initial seed point is a segment starting point of a discrete segment, searching an adjacent projection grid from the initial seed point to the right side, if the adjacent projection grid value is 1, defining the adjacent projection grid as a new seed point, continuing to search to the right side of the new seed point until the adjacent projection grid value on the right side of the new seed point is 0, defining the current new seed point as a segment end point of the discrete segment, classifying all point cloud pixel points between the segment starting point and the segment end point into one class, and connecting the point cloud pixel points into the discrete segment;
3.1.3, taking the adjacent projection grid on the right side of the line segment end point as the initial position of the one-dimensional projection grid, repeating the step 3.1.2 until all point clouds are classified, and generating corresponding discrete line segments;
and 3.1.4, acquiring the adjacent distances of all the discrete line segments, and combining the line segments with the adjacent distances smaller than a set threshold into one line segment.
7. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 2, characterized in that: the specific method of the point cloud main direction pasting method comprises the following steps:
the method comprises the steps of obtaining a main direction of the tower head actual measurement point cloud and two secondary directions perpendicular to the main direction by calculating feature vectors of all points in the tower head actual measurement point cloud, establishing a reference coordinate system with the center of gravity of the tower head actual measurement point cloud as an original point and the main direction and the secondary directions of the point cloud as coordinate axes, obtaining the main direction of the tower head model point cloud and the two secondary directions perpendicular to the main direction by calculating feature vectors of all points in the tower head model point cloud, establishing a reference coordinate system with the center of gravity of the tower head model point cloud as the original point and the main direction and the secondary directions of the point cloud as coordinate axes, and then adjusting the two reference coordinate systems to be consistent.
8. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 2, characterized in that: the specific method for ICP matching comprises the following steps:
step 4.1, searching a point set closest corresponding point of a tower head model point cloud and a tower head actual measurement point cloud to be evaluated by a step-by-step iteration method;
step 4.2, calculating the optimal rigid body transformation parameters between the two point sets to enable the optimal rigid body transformation parameters to enable the sum of squares of distances of all corresponding point pairs to be the minimum, wherein the rigid body transformation parameters comprise rotation matrixes and translation parameters;
and 4.3, acting the optimal rigid body transformation parameters obtained in the step 4.2 on the tower head actual measurement point cloud after rigid body transformation, and continuously iterating the step 4.1 and the step 4.2 until the sum of the squares of the distances of all corresponding point pairs between the two point sets meets the given convergence precision or reaches the maximum iteration times, and finally obtaining the rigid body transformation parameters between the two point sets.
9. The unmanned aerial vehicle laser radar point cloud tower head deformation detection method of claim 2, characterized in that: the specific implementation method of the step 5 is as follows:
step 5.1, obtaining the inclination angle of the tower head of the tower to be evaluated through a rotation matrix in the rigid body transformation parameters, and if the inclination angle is smaller than a certain threshold value, judging that the tower head of the tower to be evaluated does not have inclination within an allowable range; if the inclination angle is larger than the threshold value, determining that the tower head of the tower to be evaluated is inclined;
step 5.2, calculating the distance between the actual measurement point cloud of the tower head of the matched tower and the corresponding matching point in the tower head model of the tower, and if the distances are smaller than a certain threshold value, judging that no deformation exists in the tower head of the tower to be evaluated within an allowable range; and if the distance between the partial actual measurement point and the theoretical point is larger than the threshold value, judging that the tower head of the tower to be evaluated has deformation, wherein the deformation occurrence part is the part with the distance larger than the threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110687078.5A CN113554595B (en) | 2021-06-21 | 2021-06-21 | Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110687078.5A CN113554595B (en) | 2021-06-21 | 2021-06-21 | Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113554595A true CN113554595A (en) | 2021-10-26 |
| CN113554595B CN113554595B (en) | 2022-11-25 |
Family
ID=78130756
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110687078.5A Active CN113554595B (en) | 2021-06-21 | 2021-06-21 | Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113554595B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113534188A (en) * | 2021-09-16 | 2021-10-22 | 天津市普迅电力信息技术有限公司 | Tower deformation defect detection method based on unmanned aerial vehicle laser point cloud modeling |
| CN114384937A (en) * | 2022-01-20 | 2022-04-22 | 成都奥伦达科技有限公司 | Tower and automatic marking method of key points thereof |
| CN114626131A (en) * | 2022-03-23 | 2022-06-14 | 武汉北曦盛科技有限公司 | Power grid power infrastructure security assessment analysis method, system and storage medium |
| CN116597616A (en) * | 2023-05-23 | 2023-08-15 | 中国建筑材料工业地质勘查中心四川总队 | Intelligent monitoring and early warning system for geological disasters in mining area |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104636603A (en) * | 2015-01-20 | 2015-05-20 | 华北电力大学（保定） | Extra-high-voltage single pole bracing wire tower torsion frequency two-freedom-degree calculation method |
| CN105333861A (en) * | 2015-12-02 | 2016-02-17 | 中国测绘科学研究院 | Pole and tower skew detection method and device based on laser-point cloud |
| CN107610223A (en) * | 2017-09-20 | 2018-01-19 | 广东电网有限责任公司机巡作业中心 | Power tower three-dimensional rebuilding method based on LiDAR point cloud |
| CN107644433A (en) * | 2017-08-30 | 2018-01-30 | 北京控制工程研究所 | Improved closest approach iteration point cloud registration method |
| CN108830933A (en) * | 2018-06-26 | 2018-11-16 | 广东电网有限责任公司 | A kind of electric force pole tower tower body method for reconstructing, system, medium and equipment |
| CN109872352A (en) * | 2018-12-29 | 2019-06-11 | 中国科学院遥感与数字地球研究所 | Power-line patrolling LiDAR data autoegistration method based on shaft tower characteristic point |
| CN110060256A (en) * | 2019-03-08 | 2019-07-26 | 广东工业大学 | A kind of shaft tower extractive technique based on airborne LiDAR point cloud |
| CN110136260A (en) * | 2019-04-24 | 2019-08-16 | 广州供电局有限公司 | Electric power line pole tower and its implementation based on laser point cloud full feature model library |
| US10565787B1 (en) * | 2017-01-27 | 2020-02-18 | NHIAE Group, LLC | Systems and methods for enhanced 3D modeling of a complex object |
| CN111830528A (en) * | 2020-06-29 | 2020-10-27 | 西安交通大学 | Tower characteristic point automatic identification and inclination parameter automatic measurement method based on laser point cloud |
| CN112669444A (en) * | 2020-12-17 | 2021-04-16 | 国网山西省电力公司运城供电公司 | Rapid construction method and system for typical scene of distribution network line unmanned aerial vehicle routing inspection |
| CN112698303A (en) * | 2020-12-23 | 2021-04-23 | 国网电力科学研究院武汉南瑞有限责任公司 | Method and system for measuring point cloud tower inclination parameters based on unmanned aerial vehicle laser radar |
| CN112711987A (en) * | 2020-12-11 | 2021-04-27 | 国网电力科学研究院武汉南瑞有限责任公司 | Double-laser-radar electric power tower three-dimensional point cloud enhancement system and method |
-
2021
- 2021-06-21 CN CN202110687078.5A patent/CN113554595B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104636603A (en) * | 2015-01-20 | 2015-05-20 | 华北电力大学（保定） | Extra-high-voltage single pole bracing wire tower torsion frequency two-freedom-degree calculation method |
| CN105333861A (en) * | 2015-12-02 | 2016-02-17 | 中国测绘科学研究院 | Pole and tower skew detection method and device based on laser-point cloud |
| US10565787B1 (en) * | 2017-01-27 | 2020-02-18 | NHIAE Group, LLC | Systems and methods for enhanced 3D modeling of a complex object |
| CN107644433A (en) * | 2017-08-30 | 2018-01-30 | 北京控制工程研究所 | Improved closest approach iteration point cloud registration method |
| CN107610223A (en) * | 2017-09-20 | 2018-01-19 | 广东电网有限责任公司机巡作业中心 | Power tower three-dimensional rebuilding method based on LiDAR point cloud |
| CN108830933A (en) * | 2018-06-26 | 2018-11-16 | 广东电网有限责任公司 | A kind of electric force pole tower tower body method for reconstructing, system, medium and equipment |
| CN109872352A (en) * | 2018-12-29 | 2019-06-11 | 中国科学院遥感与数字地球研究所 | Power-line patrolling LiDAR data autoegistration method based on shaft tower characteristic point |
| CN110060256A (en) * | 2019-03-08 | 2019-07-26 | 广东工业大学 | A kind of shaft tower extractive technique based on airborne LiDAR point cloud |
| CN110136260A (en) * | 2019-04-24 | 2019-08-16 | 广州供电局有限公司 | Electric power line pole tower and its implementation based on laser point cloud full feature model library |
| CN111830528A (en) * | 2020-06-29 | 2020-10-27 | 西安交通大学 | Tower characteristic point automatic identification and inclination parameter automatic measurement method based on laser point cloud |
| CN112711987A (en) * | 2020-12-11 | 2021-04-27 | 国网电力科学研究院武汉南瑞有限责任公司 | Double-laser-radar electric power tower three-dimensional point cloud enhancement system and method |
| CN112669444A (en) * | 2020-12-17 | 2021-04-16 | 国网山西省电力公司运城供电公司 | Rapid construction method and system for typical scene of distribution network line unmanned aerial vehicle routing inspection |
| CN112698303A (en) * | 2020-12-23 | 2021-04-23 | 国网电力科学研究院武汉南瑞有限责任公司 | Method and system for measuring point cloud tower inclination parameters based on unmanned aerial vehicle laser radar |
Non-Patent Citations (4)
| Title |
|---|
| TAO LI等: "Ultra High Voltage Power Tower SAR Interferometry and Icing Tower Testing Results", 《2019 6TH ASIA-PACIFIC CONFERENCE ON SYNTHETIC APERTURE RADAR (APSAR)》 * |
| 胡庆武等: "基于LIDAR点云数据的电力线提取和拟合方法研究", 《测绘与空间地理信息》 * |
| 虢韬等: "机载LiDAR快速定位高压电塔方法研究", 《遥感技术与应用》 * |
| 郑晓光等: "基于全要素组件模型库的输电线路杆塔三维高效高精度建模", 《广东电力》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113534188A (en) * | 2021-09-16 | 2021-10-22 | 天津市普迅电力信息技术有限公司 | Tower deformation defect detection method based on unmanned aerial vehicle laser point cloud modeling |
| CN113534188B (en) * | 2021-09-16 | 2021-12-17 | 天津市普迅电力信息技术有限公司 | Tower deformation defect detection method based on unmanned aerial vehicle laser point cloud modeling |
| CN114384937A (en) * | 2022-01-20 | 2022-04-22 | 成都奥伦达科技有限公司 | Tower and automatic marking method of key points thereof |
| CN114626131A (en) * | 2022-03-23 | 2022-06-14 | 武汉北曦盛科技有限公司 | Power grid power infrastructure security assessment analysis method, system and storage medium |
| CN114626131B (en) * | 2022-03-23 | 2023-12-19 | 新风光电力科技(北京)有限公司 | Power grid power infrastructure safety evaluation analysis method, system and storage medium |
| CN116597616A (en) * | 2023-05-23 | 2023-08-15 | 中国建筑材料工业地质勘查中心四川总队 | Intelligent monitoring and early warning system for geological disasters in mining area |
| CN116597616B (en) * | 2023-05-23 | 2023-11-28 | 中国建筑材料工业地质勘查中心四川总队 | Intelligent monitoring and early warning system for geological disasters in mining area |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113554595B (en) | 2022-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113554595B (en) | Unmanned aerial vehicle laser radar point cloud tower head deformation detection device and method | |
| CN111830528A (en) | Tower characteristic point automatic identification and inclination parameter automatic measurement method based on laser point cloud | |
| CN107392247B (en) | Real-time detection method for ground object safety distance below power line | |
| CN110794413B (en) | Method and system for detecting power line of point cloud data of laser radar segmented by linear voxels | |
| CN111537515A (en) | Iron tower bolt defect display method and system based on three-dimensional live-action model | |
| CN112698303A (en) | Method and system for measuring point cloud tower inclination parameters based on unmanned aerial vehicle laser radar | |
| CN102768022A (en) | Tunnel surrounding rock deformation detection method adopting digital camera technique | |
| CN110910440B (en) | Power transmission line length determination method and system based on power image data | |
| CN114998338A (en) | Mining quantity calculation method based on laser radar point cloud | |
| CN112150412A (en) | Insulator self-explosion defect detection method based on projection curve analysis | |
| CN116486289A (en) | Gas pipeline high-consequence area identification method driven by multi-source data and knowledge | |
| CN117387603B (en) | Power inspection map navigation method and device, medium and electronic equipment | |
| CN112711987B (en) | Double-laser-radar electric power tower three-dimensional point cloud enhancement system and method | |
| CN112187174A (en) | Solar photovoltaic bracket abnormity detection method and system based on artificial intelligence | |
| CN112365543A (en) | Geological structure surface extraction method and device based on optical image | |
| CN112862921B (en) | Power grid distribution image drawing method | |
| CN113534188B (en) | Tower deformation defect detection method based on unmanned aerial vehicle laser point cloud modeling | |
| CN113487601A (en) | Method for calculating lightning protection angle based on high-precision three-dimensional point cloud data | |
| CN114509777A (en) | Electric transmission project entity acceptance method based on Cesium platform | |
| Su et al. | Automatic multi-source data fusion technique of powerline corridor using UAV Lidar | |
| Lu et al. | A tilt angle measurement method of power tower using UAV LiDAR point cloud | |
| CN115828400B (en) | Satellite cloud image-based power transmission and transformation project acceptance rechecking method and system | |
| CN111598036A (en) | Urban group geographic environment knowledge base construction method and system of distributed architecture | |
| CN112053365A (en) | Method for extracting laser point cloud data of power transmission line conductor based on linear characteristics | |
| CN115685222B (en) | Automatic power line tower detection method based on laser point cloud data |
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 |